The data analyzed in this study were collected as part of the Second Bavarian Food Consumption Survey (BVS II), a mono-centric study conducted between 2002 and 2003. This cross-sectional study assessed dietary intake and lifestyle factors among a random sample of 1,050 individuals aged 13 to 80 years from the Bavarian population in Germany. We included free-living individuals in the age range of 13–80 years (18–80 years for blood collection) identified by random sampling and willing to participate in the study and to provide informed consent. Exclusion criteria were language barriers and other reasons for not understanding the study information, thus unable to give informed consent.

Data collection was performed using standardized computer-assisted personal interviews and included information on socio-economic status, lifestyle behaviors, health status, diagnosed diseases, and medication use.

Adult participants (aged ≥18 years) who had completed at least one 24-hour dietary recall and provided information on physical activity (n=879) were invited for blood sampling and further anthropometric assessments. Of these, samples from 568 participants were obtained.

All participants provided written informed consent. The BVS II study was approved by the Ethics Committee of the Bavarian Medical Association (Bayerische Landesärztekammer) on June 19, 2002 (No. 02111). The BVS II study was conducted in accordance with the principles defined in the Declaration of Helsinki.

Venous blood samples were taken from 568 participants, chilled at 4°C and centrifuged, aliquoted, and stored at -80°C (serum) or -20°C (red blood cells, RBC) until analysis. Pre-processing of the blood samples (centrifugation, aliquoting) was performed in the local health offices, and the samples were transported to the central freezers on the same day according to standard specifications. The influence of pre-analytical variables on sample quality can therefore be regarded as low. Continuous tracking of temperature enabled the safe and efficient storage of samples, which is essential for the long-term stability of biospecimens. Measurements of specific IgE and AABs were conducted in serum, while fatty acid profiles were measured in RBC membranes. The analysis of RBC membrane fatty acids was conducted no later than 6 months after blood drawing. Serum samples used for IgE determination were stored for a maximum of 1.25 years, and AAB analysis was performed about 17 years after sample collection. Due to the requirement to use previously unthawed serum samples, a set of 331 samples was selected for analysis of AABs; this selection is based on sample availability, and availability is very likely at random. 17 years after blood collection (in 568 persons), no full set of unthawed plasma or serum samples was available anymore. The characteristics of the participants in the present study and in the full group (with blood collection) are given in the Supplementary Table S1.

After the exclusion of eight individuals with a diagnosis of cancer, data from 323 participants remained for the present analysis.

Information about allergic sensitization was derived by specific serum IgE measurements using the CAPSX1 in-vitro screening test (Pharmacia Upjohn, Uppsala, Sweden). This ELISA-based test, also known as Phadiatop test, detects the presence of specific serum IgE against common allergens (timothy, rye, birch, mugwort pollen, house dust mite (Dermatophagoides pteronyssinus), cat and dog epithelia, as well as Cladosporium herbarum mould). All serum samples were analyzed in a blinded manner according to the manufacturer’s recommendations at the Federal State Health Office Baden-Württemberg, and the results were categorized in seven CAP classes. A result of CAP class 2 or higher, i.e., at least one specific IgE ≥700 U/l, was considered as allergic sensitization. Other researchers also used CAP classes ≥1 to define allergic sensitization. According to Fall and colleagues CAP class 1 represents low IgE concentrations, and CAP class 2 moderately increased IgE concentrations (11). We added additional analysis comparing the results of both definitions.

During the face-to-face interview, study participants were asked about ever diagnosis (by a physician) of allergic diseases, including the IgE-mediated allergies allergic rhinitis, atopic dermatitis/neurodermatitis, and food allergy; we also included asthma, though we could not separate allergic asthma. We defined “at least one diagnosis of allergic disease” as an exposure variable, comprising information on asthma, allergic rhinitis, atopic dermatitis/neurodermatitis, and food allergy.

Serum samples were analyzed by indirect immunofluorescence tests (IIFTs), ELISAs, and line blots purchased from EUROIMMUN Inc. (Lübeck, Germany) and evaluated by EUROIMMUN devices (Analyzer, Euroblotone, and Sprinter XL) and fluorescence microscope (IIFTs). Overall, 44 humoral AABs were measured, and seven screening tests were conducted as reported elsewhere (5). In the present study, we analyzed the results of the five most frequently detected AABs (rheumatoid factor (RF), ß2-glycoprotein (IgM), cardiolipin (IgM), cardiolipin (IgG), and anti-dsDNA) as well as all screening tests which refer to systemic rheumatic diseases (ANA nuclear, ANA cytoplasmic, ANA mitotic, ANCA, cANCA, pANCA, ENA) (see Supplementary Table S2). Among the excluded AABs were anti-CCP and ß2GPI IgG. Positivity of AABs was defined according to cutoffs recommended by the manufacturer for each type of blood analysis (IIFT, ELISA, or line blot) and the clinical laboratory of the University Hospital Augsburg (Supplementary Table S2). An exception was RF and ß2-g-glycoprotein IgM (ß2GPI). Based on the EULAR guidance (12), RF was classified as “high-positive” with values higher than three times the upper limit of normal (ULN), and ß2GPI results ≥40 U/l were defined as “moderate-positive” (13). This was carried out to increase specificity, given the high frequency of RF positivity in the sample when using the ULN value as the cut-off value. In the main analysis, the participants with (high-)positive results were compared to all other persons, called “non-positive”. In a sensitivity analysis, non-normal findings were compared to “normal” findings, where the cut-offs were lowered to 14 U/l for RF and 20 U/l for ß2GPI, and all borderline results were captured in the group “non-normal” (Supplementary Table S2).

The test results were not used for diagnostic purposes. Positive test results were not validated by means of other test methods.

Body weight and height were measured on occasion of blood collection, and body mass index (BMI, kg/m²) was calculated. Sports activity was assessed during the baseline interview and repeated 24-h recalls (14), respectively. The concentration of plasma C-reactive protein (CRP) was analyzed in the clinical laboratory. Fatty acid analysis of RBC membranes was conducted using gas chromatography in combination with a flame ionization detector as described elsewhere (15). Fatty acids were expressed as the percentage of total fatty acid methyl esters (% FAME). Serum β-carotene concentrations (µmol/l) were estimated by high-performance liquid chromatography–ultraviolet/visible spectrophotometry (9). Eicosapentaenoic acid (EPA) and ß-carotene were shown to be inversely associated with specific IgE and/or allergic rhinitis in the BVS II study (9, 15), and thus were considered as possible confounders. The variables “smoking” (never, ex, current), “hormone replacement use” (yes/no), and a proxy variable for premenopausal vs. peri-/postmenopausal status, i.e., age ≤ 50 years and age >50 years (in women) were tested as possible confounders; as we saw no impact of all three variables on the model results, these were not included in the final models.

We did not collect information on current use of immunomodulatory drugs or corticosteroids, known autoimmune diseases, or infections in our study; thus, we are unable to consider these variables as possible confounders or for stratified analyses.

Characteristics of individuals with at least one positive AAB vs. non-positive persons were described as n (%) for categorical variables and as median (Q1-Q3) and arithmetic mean (standard deviation) for continuous variables. Group differences were tested with Fisher’s exact test for categorical and Mann-Whitney-U test for non-normally distributed continuous data, both within females and males.

Logistic regression models, all stratified by sex, were employed to assess the association between AAB positivity (outcome) and the two binary exposures, allergic sensitization (no/yes) and “at least one diagnosis of allergic disease” (no/yes). Only groups with acceptable numbers of persons with AAB positivity were evaluated; thus, analyses were conducted for RF, ANA nuclear, ANCA, and “at least one positive AAB test”. Also for this approach, the power was limited, and chance findings cannot be excluded.

The initial model included adjustment only for age, while the multivariable-adjusted models included BMI, sports activity (yes/no), CRP, EPA in RBC, and serum β-carotene concentration. All continuous covariables (age, BMI, CRP, EPA, β-carotene) were tested for multicollinearity with the variance inflation factor and for correlation with a correlation matrix. Statistical significance was set at p<0.05. Analyses were primarily conducted using Python with pandas (16), NumPy (17), statsmodels (18), matplotlib (19), and tableone (20) packages. In addition, R was used for validation.

Systemic connective tissue diseases are a classic example of AIDs that develop as a result of immune tolerance disorders. The presence of AABs in blood and joint fluid is a characteristic feature of rheumatoid arthritis (RA) that distinguishes this disease from other inflammatory joint disorders. The hallmark antibodies of RA are RF and anticitrullinated protein antibodies (ACPA), which are detectable in 60–70% of RA patients already in the earliest stages of the disease (25). The RF is a representative AAB against the crystallizable fragment (Fc) of denatured IgG that is primarily detected in patients with RA (26). RF is not limited to RA but is also found in patients with other AIDs like Sjögren’s syndrome, or infectious diseases like hepatitis and tuberculosis, as well as in non-symptomatic healthy individuals. In 2010, the American College of Rheumatology/European League Against Rheumatism developed new classification criteria for RA, including serologic markers. They suggested defining high-positive RF values, i.e., values exceeding ULN by more than three times (12). We followed this suggestion and ended with 17.6% of our population classified as high-positive. However, when applying the definition of values above normal (RF≥14 U/l), 35.9% of our population had non-normal RF values, meaning that 1/3 of the population has detectable RF. In our study, RF positivity in women was associated with allergic sensitization and reached statistical significance when using the non-normality definition (OR = 2.16, p<0.030). However, a relationship with allergic diseases was only observed in women when using the classification of high-positive values (versus other values) with an OR of 2.50 (p<0.026).

This finding is supported by several systematic reviews and meta-analyses (SR&MA) that examined the association between allergic diseases and RA. One SR&MA showed a significantly higher risk of incident RA in atopic dermatitis patients (27). Subgroup analysis also revealed a significantly higher risk of RA in cohort studies (pooled OR, 1.37; 95% CI, 1.25–1.50). In another SR&MA, a significantly higher risk of RA among patients with allergic rhinitis than individuals without allergic rhinitis was observed when only studies with acceptable quality were included (28). A third SR&MA described a significant association between asthma and a higher risk of incident RA but found evidence of publication bias (29). Patients with asthma were also at higher risk of systemic lupus erythematosus (30). A large population-based study in the U.S. (NHANES) with 20,050 participants found that physician diagnosis of allergic disorders was associated with an increased risk of physician-diagnosed AIDs (adjusted odds ratio, 1.67; 95% CI, 1.35-2.07; p<0.001) (31). However, studies investigating direct associations between RF and allergic diseases or RF and allergic sensitization could not be identified (32).

The confirmation of ANA positivity in a healthy individual is usually of unknown significance and, in most cases, benign. From a clinical point of view, the classic screening test for SLE is the presence in serum of ANAs, and ANA positivity is required to make a diagnosis of lupus since more than 99% of patients with SLE have significant levels of this ABB detected at some time during the course of the disease. However, since the prevalence of SLE is low, most individuals presenting to a physician with ANA positivity do not have SLE and are not at high risk for developing this disease (2, 3).

The percentage of the population with ANAs was reported to be approximately 25% by using the standard analysis methods (indirect immunofluorescence assay performed on HEp-2 cells (IIFA on HEp-2 or HEp-2000)), and 2.5% have distinctly elevated ANA levels (2, 33). According to other reports using IIFA, ANA in low counts may appear in up to 40% of healthy people (3, 34), which is in good agreement with the here observed ANA prevalence of 42% with non-normal values (Supplementary Table S4); also, the prevalence of 17% of individuals with positive results (Table 2) in our study fits well with the prevalence of 14% measured in the U.S. population (1, 35) or 15.8% obtained in a Swedish population (36). Among the NHANES participants who had ANA, nuclear staining was seen in 85%. ANA positivity is more frequently detected in women, and thus female gender is a risk factor for ANA positivity (3, 35). Both the high percentage of nuclear ANA and the female dominance were confirmed in our study (Table 2).

The relationship between ANA occurrence and allergic diseases is poorly documented. However, the mechanism of allergic sensitization and AIDs has a common thread. An increased production of IgE antibodies and the presence of ANA in selected disease entities were observed. Both autoimmune and allergic diseases seem to show overlapping pathogenic processes and often occur in genetically predisposed individuals (3). The activation of basophils secreting proinflammatory factors and affecting the differentiation of TH17 lymphocytes occurs in both conditions. The presence of ANAs was confirmed in many systemic connective tissue diseases and some allergic diseases, such as atopic dermatitis, non-allergic asthma, and pollen allergy (3). In the present study, we found a strong positive association between nuclear ANA positivity and allergic sensitization (OR = 2.85, p<0.023) only in women, but no association with allergic diseases, arguing for a relationship between joint biologic processes rather than between diseases.

ANCA-associated vasculitis (AAV) is a disease entity characterized by systemic vasculitis positive for ANCA; it often leads to severe organ damage, such as diffuse bronchoalveolar hemorrhage and rapidly progressive glomerulonephritis (37). ANCA targets lysosomal proteins, such as proteinase 3, myeloperoxidase, and lactoferrin. The presence of ANCA was also reported in other diseases such as inflammatory bowel disease (38) and SLE (39). It is more frequently found in males than females (39), which was confirmed in the present study. The literature describes several situations and diseases related to ANCA positivity (AAV) where an association with allergies has been identified. For example, in eosinophilic granulomatosis with polyangiitis (EGPA), nearly all patients have an allergic background (asthma and/or allergic sinusitis) (40). However, we could not identify a population-based study that explored the relationship between ANCA positivity (without clinical diagnoses) and allergic sensitization or allergic disease. The increased OR observed for the association between ANCA positivity and allergic sensitization in men did not reach statistical significance, and no association was seen with allergic diseases.

Due to their presence in the general population as well as in multiple AIDs, the presence of an AAB alone does not make a diagnosis; the result has to be interpreted along with clinical findings. Similarly, the absence of AABs does not exclude a disease (41). The common AABs used in clinical practice include RF, anti-CCP antibodies, ANAs, ANCA, and antiphospholipid antibodies. These and other AABs were included in the variable “at least one positive AAB test”. Thus, this variable reflects the presence of biologic autoimmunity in individuals. Testing its association with allergic sensitization means searching for an association with biologic allergenicity. Overall, we obtained evidence for an association between these two biological phenotypes. Accordingly, this association was strongest (OR = 3.17) and statistically significant (p<0.007) after including all non-normal findings (Table 4), and no association with allergic disease was observed.

Both allergic and autoimmune diseases are marked by a dysregulation of the immune system, reacting to otherwise harmless environmental substances or autoantigens, which leads to inflammation and tissue damage (42). Allergies are primarily driven by type 2 inflammation, which can intensify autoimmune responses in susceptible individuals by promoting tissue damage or autoantibody production. The type 2 immune response, mainly triggered by the activation of type 2 helper T cells (Th2 cells), plays a central role in the overlap between allergies and certain autoimmune conditions due to shared immune pathways, the release of specific cytokines (signaling molecules), and the activation of immune cells that drive processes like eosinophil activation, antibody production, and tissue repair (43). However, other factors, including genetic predisposition, environmental triggers, and additional immune mechanisms (e.g., Th1/Th17 responses or autoantibodies), also play significant roles (42). The interplay is intricate and varies depending on the specific combination of diseases.

The association between allergies and autoimmune diseases, observed to be stronger in women in this study, likely arises from a combination of hormonal, genetic, and environmental factors that differentially regulate immune responses by sex (44). In women, estrogen amplifies Th2 (allergy-related) and Th17 (autoimmunity-related) signaling pathways (45, 46). Conversely, in men, testosterone suppresses Th2 and Th17 responses, promotes regulatory T cells (Tregs), reducing inflammation and the overlap between these conditions (47).

The odds ratio for ANA nuclear and allergic sensitization was statistically significant when comparing high-positive vs. non-positive, but no association existed when comparing non-normal vs. normal. The opposite was observed with RF. As this resulted only from the change in classification, these results must be interpreted as its direct consequence. When including non-normal values in the evaluation, the associations became slightly stronger for RF and for the summary variable “at least one non-normal AAB test”. The results of the summary variable seem to be driven by RF, leading to the conclusion that the results of this study support the idea of an association between AAB prevalence and allergic sensitization. However, ANA nuclear behaved differently. For RF, continuous values were measured, while for ANA nuclear different titers were used. Including the titer =1:100 as non-normal titers for ANA nuclear may have added more “noise” than clear information and thus including this titer may have not been useful as discussed by others (48).

The present study investigated a comprehensive panel of systemic AABs and screening tests in the general population for their association with allergic sensitization and physician-diagnosed allergic diseases. However, many more known systemic AABs could not be included in our study. With the current sample size, only the most frequent AABs and a summary variable (at least one positive AAB test) could be evaluated in multivariable adjusted regression models. Only a very limited number of individuals were positive for more than two AAB, thus precluding the use of a continuous variable “number of AAB”. There is an increased risk of chance findings due to the limited number of positive subjects for some individual AABs, and thus underpowered analyses. Due to the requirement to use unthawed samples for AAB analysis, we could include only 331 subjects (out of a study with 568 subjects).

Also, the exposure variables (specific IgE, diagnosis of allergy) are either not comprehensive or low in numbers. The Phadiatop test (CAPSX1) analyzed antibodies (specific IgE) against the most common antigens, while many more antigens are known. Among allergic diseases, we could not assess all atopic diseases or all type-1 allergies.

Due to the cross-sectional design of our study, our findings cannot be used to infer causality. In addition, the specific features of the Bavarian population sample preclude a direct transfer of the findings to other ethnicities and age groups. We did not collect information on current use of immunomodulatory drugs or corticosteroids, known autoimmune diseases, or infections in our study; thus, unmeasured confounding cannot be excluded. In addition, analytic results may vary depending on the manufacturer’s test design and test performance, which could explain some differences in frequencies observed for certain AABs in different studies. Notably, all test kits used in our analyses were purchased from the same provider and analyzed in one laboratory.