Variable Food-Specific IgG Antibody Levels in Healthy and Symptomatic Chinese Adults

Background The presence of food-specific IgG antibodies in human serum may be useful for diagnosis of adverse food reactions. However, the clinical utility of tesing for such antibodies remains very controversial. The aim of this study was to evaluate the serum levels and population distribution of food-specific IgGs and their association with chronic symptoms in a large-scale Chinese population. Methodology/Principal Findings A total of 21305 adult participants from different regions of China had 14 type of food-specific serum IgG antibodies that were measured by enzyme-linked immunosorbent assay. Amongthese, 5,394 participants were randomly chosen to complete follow-up questionnaire surveys on their dietary characteristics and chronic symptoms. The concentrations of food-specific IgGs against 14 foods ranged from a median (interquartile range) of 7.3 (3.8, 12.6) U/mL of pork-specfic IgG to 42.3 (28.8, 60.2) U/mL of crab-specific IgG. The concentration of food-specific IgGs was closely related to gender; after adjustment for region and age, women had higher concentrations of food-specific IgGs against all of the 14 foods except chicken (regression coefficient (95% CI): 0.01 (−0.003, 0.023); P = 0.129) and corn (0.002 (−0.013, 0.016); P = 0.825). Similar results were also found in the relationship of geographic region to the food-specific IgG concentrations for the 14 foods. Chronic symptoms were negatively associated with the concentrations of a few food-specific IgGs, and were positively associated with the concentrations of other food-specific IgGs. Conclusions The levels of food-specific IgGs were variable both in healthy and in symptomatic Chinese adults. These findings raise awareness that demographic factors, the type of food and specific chronic symptoms should be considered before food elimination treatment based on IgG testing in patients with chronic symptoms is used in clinical practice.


Introduction
The prevalence of migraine, asthma, allergic dermatitis and irritable bowel syndrome has been continuously increasing. Etiological studies suggest that these diseases may be related to adverse food reactions (food hypersensitivity) [1,2,3,4,5]. According to a recent report of an expert panel sponsored by the U.S. National Institute of Allergy and Infectious Diseases [6], adverse food reactions include food allergies and food intolerances, and the gold standard for the diagnosis of adverse food reactions remains the double-blind, placebo-controlled food challenge (DBPCFC). However, DBPCFC is difficult to use in large population studies due to the tedious, inconvenient, and costly procedures involved [7]. Other diagnostic tests such as skin prick tests, serum food-specific IgE tests, and atopy patch tests lack specificity and may lead to the unnecessary exclusion of foods from the patient's diet [6].
Previous studies have found that the levels of food-specific IgGs and IgG subclasses in blood serum were significantly higher in individuals with food hypersensitivity [8,9,10], and that IgGmediated immunologic responses may play an important role in the pathogenesis of adverse food reactions [11,12,13]. However, other studies indicate that the serum level of food-specific IgGs may also be elevated in healthy individuals, suggesting that the elevation of food-specific IgGs and IgG subclasses may merely reflect the presence of immunological tolerance to food-specific antigens, i.e., that it may only reflect a history of exposure to a certain food [14,15]. Therefore, it was determined that food-specific IgGs and IgG subclasses lack diagnostic value for food allergy/intolerance [15,16,17].
Although the diagnostic utility of the presence of food-specific IgGs in patients with adverse food reactions remain controversial, serum food-specific IgG tests have been used in patients with chronic symptoms because of their convenience [18,19]. In addition, selective food elimination based on the results of foodspecific IgG testing has been suggested to be effective in the treatment of irritable bowel syndrome and migraine [4,5]. However, the distribution of food-specific IgG concentrations, and the relationship of food-specific IgGs to chronic symptoms has not been reported in any large-scale population studies [18]. The aim of this study was to investigate the factors related to the presence of food-specific IgGs and their association with diet characteristics and chronic symptoms in a large adult population from various regions of China.

Study Population
The study population was recruited as patients presented to the clinic and underwent routine physical examination in general hospitals in 23 provinces in China ( Figure S1), beginning in August 2008. Participation in the study was voluntary. Individuals aged 18-75 years were randomly selected using a single-blind method;, patients who were pregnancy or had a history of anaphylactic shock, secondary chronic symptoms caused by other known cardiovascular diseases, chronic kidney diseases or organic intestinal diseases were excluded. The concentrations of serum IgGs that reacted with 14 foods that are commonly consumed in China were measured in these individuals. By October 2010, a total of 21305 participants were included after excluding those with incomplete data. From this population, 6000 subjects were randomly selected and contacted for a questionnaire survey. Informed written consent was obtained from all participants, and the study protocol was approved by the Institution Review Board of the Chinese PLA General Hospital, Beijing, China.

Questionnaire Design
A questionnaire was designed with questions adapted to Chinese adults. The questions included consumption frequencies of 14 common foods: beef, chicken, codfish, corn, crab, eggs, mushrooms, cow's milk, pork, rice, shrimp, soybeans, tomatoes and wheat. Regular consumption referred to a consumption frequency of $3 times/week, and rare consumption referred to a consumption frequency # one time/3 months; moderate consumption was defined as a consumption frequency that fell between these two extremes. The questionnaire also included information about perceived connections between ingestion of specific foods and chronic symptoms, such as gastrointestinal symptoms (abdominal pain, diarrhea or nausea), eczema, urticaria, rhinitis, asthma, arthritis, migraine headaches or other chronic symptoms (chest pain, or heart discomfort).

Food-specific Serum IgG Assays
Participants' sera were collected and sent, with only numerical identifiers, to authorized hospital laboratories, where enzymelinked immunosorbent assays (ELISA) were performed to measure the concentrations of 14 different food-specific IgGs using a commercially available kit (Biomerica, Inc. USA). This assay has been described in a previous study [20]. Briefly, the serum samples were diluted 100-fold, and added to reaction wells coated with one of 14 food-specific allergens. After the plate was sealed, incubated at room temperature for 1 hour, washed and patted dry, anti-human IgG antibody conjugated to horseradish peroxidase was added to the wells. After another incubation at room temperature for 30 minutes, washing and patting dry, 100 ml of a substrate mixture consisting of equal proportions of 3,39,5,59-tetramethylbenzidine and hydrogenperoxide was added. The reaction was stopped after 10 minutes, and absorbance was measured at 450 nm. The concentrations of food-specific IgGs (U/mL; 1 U = 1.47 ng) were calculated using the standard curves of 14 types of food-specific IgGs provided by the manufacture. The detection limit was 0.01 U/mL. All authorized laboratories were asked to satisfy several standards of quality assurance for the determinations of these 14 types of food-specific IgGs. The quality assurance standards were as follows: the optical density of the negative controls should be ,0.2, the concentrations of IgG in the positive quality controls should be .100 U/mL and the optical density in 50 U/mL, 100 U/mL, 200 U/mL and 400 U/mL standard wells should be 1.2 times more than in negative 50 U/ mL, 100 U/mL and 200 U/mL standard wells, respectively. For the great majority of clinically relevant allergens, no standardized reference allergens are available for comparisons between methods. To ensure the reproducibility and accuracy of the ELISA kit assays, each serum sample was tested in duplicate parallel tests, all ELISA kits were required to be used within 10 days of their delivery from the factory, and all testing was performed using an automated spotting system that was supervised by well-trained medical laboratory technologists who were blinded to the clinical data. On average, the intra-and inter-assay coefficients of variant were 2.7% and 5.3%, respectively, in the Chinese PLA General Hospital laboratory, and 8.5% and 12.1% in other authorized laboratories. Before the study, all participating laboratories calibrated food-specific IgG measurements according to the quality control results distributed from the Chinese PLA General Hospital laboratory.

Statistical Analysis
All data were entered into an Epidata 3.0 database through double-blinded entry, followed by automatic check and verification. Continuous variables are reported as median values with the corresponding 25th and 75th percentiles. Discrete variables are summarized in terms of frequencies and percentages. Because the data demonstrated a skewed distribution, serum concentrations of food-specific IgGs were logarithmically transformed; undetectable specific IgG concentrations (,0.01 U/mL) were considered 0.005 U/mL. Maximum likelihood estimation was performed to evaluate the means and standard deviations of food-specific IgG concentrations using R 2.15.2. To visualize the distribution of food-specific IgG concentrations in all subjects and demographic subgroups (sex, geographic region and age group), histograms or line graphs were generated, and one-way analysis of variance (ANOVA) and post-hoc Tukey tests were used for comparison of the resultings curves. P-values for trends were calculated when age group-related differences in food-specific IgG concentrations were examined. Pearson correlation was performed to analyze the relationship of food-specific IgG concentrations among the 14 foods. Differences in food consumption frequencies in demographic subgroups were tested using the Pearson chi-square test and P values for trends were calculated when differences in the food consumption frequencies of different age group were examined. Differences in food-specific IgG concentrations between subjects with chronic symptoms and subjects without chronic symptoms were also examined with ANOVA. Multivariate linear regression models were applied to evaluate the association of food-specific IgG concentrations of each of the 14 foods with demographic factors and food consumption frequen-cies. Logistic regression models were performed to analyze the association of chronic symptoms with food-specific IgG concentrations of the 14 foods, and odd ratios (OR) and 95% CI were calculated. All statistical analyses were performed with SPSS 10.0 (SPSS Inc., Chicago, IL, USA). Two-tailed P values ,0.05 were considered significant.

Study Population Characteristics
Of the 21305 subjects (mean age: 46.63610.52 years; 13,426 men and 7,879 women), most (74.0%) were northern Chinese. Questionnaire replies were received from 5394 individuals (89.9% of the randomly selected subjects); of the 5394 individuals who responded to the questionnaire, 32.2% were women and 77.2% were northern Chinese. In the 2353 subjects who reported having chronic symptoms, the most common chronic symptoms were gastrointestinal symptoms (23.6%), rhinitis (11.8%) and migraine (9.5%) ( Table 1).

Distribution of Food-specific IgG Concentrations
The distributions of the serum concentrations of food-specific IgG against the 14 foods tested varied significantly ( Figure 1). The log-distributed curves for food-specific IgG concentrations of crab, codfish, egg and shrimp are shifed slightly to the right; the medians (interquartile range) of the IgG concentrations against these foods were 42.3 (28.8, 60.2) U/mL, 27.5 (16.9, 44.2) U/mL, 26.6 (11.4, 72.5) U/mL and 22.8 (14.9, 35.2) U/mL, respectively. For porkspecific IgG, the log-distributed curve is shifted slightly to the left; the median and interquartile range of the concentration of porkspecific IgG is 7.3 and 3.8 to 12.6 U/mL. The means and standard deviations of log-transformed food-specific IgG concentrations that were calculated using maximum likelihood estimation are shown in Table S1.
The distribution of food-specific IgG concentration against egg varied more than for other foods. For example, a total of 61 subjects had levels of egg-specific IgG above 1000 U/mL, and the proportion of subjects who had egg-specific IgG levels above 200 U/mL was 8.4%, the highest od the 14 foods. The proportion of subjects who had the egg-specific IgG levels less than 10 U/mL was 21.7%; this level was higher than for other foods against which subjects had similar IgG levels, such as codfish (only 8.8% of the subjects had codfish-specific IgG levels that were less than 10 U/ mL). Interestingly, the food-specific IgG levels against the 14 foods tested were all correlated with each other. The strongest relationship between food-specific IgG levels were found for beef and chicken (correlation coefficient, 0.691; P,0.01); the next strongest was for codfish and crab (correlation coefficient, 0.561; P,0.01) ( Table S2).

Relationship between Consumption Frequency and Food-specific IgG
Of the 5394 subjects who responded to the questionnaire, the majority regularly consumed ($3 times/week) rice and pork (83.8% and 70.1%, respectively); a minority of subjects regularly consumed shrimp and crab (15.4% and 6.9%, respectively). Men showed higher diet frequencies of codfish and meat, while more women regularly consumed cow's milk and non-meat foods. The proportions of subjects who regularly consumed rice, crab, shrimp and codfish were higher in South than in North China, while the proportion of subjects who regularly consumed wheat was higher in North China population (69.7% vs. 35.4%, respectively; P,0.001).Shrimp, chicken, beef and pork were consumed most frequently by the 18-34 yearold age group, and the consumption of egg, cow's milk, soybeans, corn, tomatoes and wheat increased with age (P for trend ,0.001) (Table S3).

Relationship between Food-specific IgGs and Chronic Symptoms
The association of chronic symptoms with serum levels of specific IgGs against the 14 foods tested is shown in

Discussion
Although commercial laboratories worldwide are currently offering broad-scale = testings for serum levels of food-specific IgGs to the public, to our knowledge, the study described here is the first representative cross-sectional study to address the distribution of food-specific IgGs and its relationship to chronic symptoms in a large population. Previous studies have either involved \ research on the potential involvement of various IgG subclasses in food allergy/intolerance [9,21] or have addressed food elimination based on IgG testing [4,5].
Although testing for the presence of food-specific IgGs has been regarded as a potential tool for the diagnosis of food allergy/ intolerance, it's the accuracy and clinical utility of such testing remain unclear. In our study, serum concentrations of foodspecific IgGs against the 14 foods tested varied widely. Different foods have different proteins and components, and this may induce different immune reactions and lead to the presence of variable amounts of food-specific IgG. The completeness of extraction of different food proteins and other components of the 14 foods tested and the relative coupling efficiency of different extracted food components/antigens to the ELISA plate would also be expected to influence the accuracy of measurement of serum concentrations of food-specific IgGs. Although the levels of  Regression coefficients and P values were derived from multivariate linear regression models in which the dietary frequencies of 14 foods (rare, 0; moderate, 1; regular, 2) were used as independent variables and the corresponding and logarithmically transformed serum food-specific IgG concentrations for 14 foods were the dependent variables. Demographic factorsincluding sex, region and age were adjusted for in these models. doi:10.1371/journal.pone.0053612.g003 food-specific IgG against the 14 tested foods were significantly different, the concentrations of all 14 food-specific IgGs, especially IgGs directed against beef and chicken, were correlated with one another. Similar results were reported by Kanagawa et al [22] in an analysis of food antigens in 6549 subjects; the results of that study suggested that the appearance of food antigens clusters may be a result of immune cross-reaction between similar food antigens [6].
In our study, the distribution of food-specific IgG levels varied as a function of demographic subgroups (sex, region and age). Women had higher concentrations offood-specific IgGs against almost all of the tested foods than men. It has been consistently reported that the prevalence of food allergy/intolerance is significantly higher in women than in men [23,24,25]. In childhood, boys suffer from asthma, urticaria, anaphylaxis and food allergies to eggplant and peanuts more frequently than girls, while the sex difference is reversed after puberty [26,27,28,29]. Studies suggest that some female sex steroids are pro-inflammatory and increase the susceptibility of atopy, while testosterone is a powerful inhibitor of histamine and known to inhibit mast cell degranulation [26,30]. Moreover, for the first time, our study demonstrated that the distribution of food-specific IgGs against almost all of the 14 foods tested were significantly different between Southern and Northern Chinese. In the present study, this difference is not related to the subject's gender, age or dietary habits; however, it may be due to the different genetic backgrounds of subjects of different ethnicities [31,32]. Our study also showed that the serum concentrations of all food-specific IgGs except egg-specific IgG were positively associated with age. Previous studies have suggested that older people may be more likely than younger people to experience allergic reactions to foods [33]. It is not clear whether the higher levels of foods-specific IgGs in older persons indicates a higher probability of food allergies with increasing age. It is also unclear why egg-specific IgG concentrations were highest in the 20-34-year-old subjects and lowest in the 35-44-yearold subjects.
Previous studies have shown that abdominal discomforts selfattributed to food hypersensitivity is not related to eating habits or meal patterns [34]. Our study also showed that although the dietary frequencies of the 14 foods differed significantly in the demographic subgroups, only the dietary frequencies of chicken, egg, cow's milk, rice and tomato were associated with the Only the foods for which food-specific IgG levels were significantly associated with certain chronic symptoms (P,0.05) are listed. Odds ratios and P values were derived from logistic regression models that set logarithmically transformed serum food-specific IgG concentrations for the 14 foods as the independent variables and the presence of various chronic symptoms (No, 0; Yes, 1) as the dependent variables. Demographic factors including sex, region and age were adjusted for in these models. doi:10.1371/journal.pone.0053612.g004 corresponding food-specific IgG serum levels after adjustment for demographic factors. The immunological mechanism responsible for the association of chicken, egg, cow's milk, rice and tomato eating patterns with food-specific serum IgG concentrations is unclear, and further studies of this phenomenon are needed.
Previous studies have shown that chronic symptoms are closely related to food hypersensitivity [2,6]. In our study, we found the food-specific concentrations of a few IgGs to be associated with a variety of chronic symptoms; the levels of some food-specific IgG were inversely associated with chronic symptoms, while the levels of other food-specific IgG were positively associated with chronic symptoms. Chronic symptoms were positively related to the serum levels of certain food-specific IgG. This may be a result of chronic inflammatory responses induced by elevated IgG or IgG subclasses in highly sensitive individuals [5]. For instance, gliadins and glutenins in wheat may increase specific IgG levels and may lead to chronic symptoms [35]. Our study also revealed negative associations between the level of rice-specific IgG and gastrointestinal symptoms, rhinitis, migraine, urticaria and eczema; such associations may result from different (protective or negative) roles of different IgG antibodies. Similarly, Tay et al [36] found that strong IgG responses to ovalbumin may be a normal physiological response to protein ingestion frequently shown in infancy, whereas up-regulated IgG responses that occur in peanut allergy may be indicative of a dysregulated immune response to peanut allergens. The normal physiological response reflected by the presence of food-specific IgG subclasses may be immunological tolerance due to the secretion of IL-10 by the activation of regulatory T cells upon prolonged exposure to food antigen [37].
The present study has several limitations. First, because foodspecific IgG antibody testing is popular in China, we only tested for the presence of IgGs against commonly consumed foods, and did not measure levels of specific IgGs against pollen, fruits, nuts or food additives [1,24,25]. However, it should be kept in mind that self-reported chronic symptoms may result from the presence of IgGs specific for the latter foods and substances rather than from the presence of IgGs against the 14 tested food allergens. Second, in our study, the prevalence of chronic symptoms was determined by self-reporting of such symptoms on a questionnaire. The chronic symptoms reported by the patients in this study have not been confirmed by the adjudication of professional doctors. Thus, misclassification of chronic symptoms is possible, and the conclusions of the study regarding the associations of specific chronic symptoms with certain food-specific IgG levels should be regarded cautiously. Third, children and adolescents 0-17 years of age were not included in our study. It has been reported that the presence of food-specific IgGs in serum is more common in children and adolescents than in adults [15]. In children and adolescents, high levels of specific IgG are not significantly associated with clinical symptoms, and are likely to represent a protective mechanism that can result in immune tolerance to the corresponding foods later in life [38]. However, a similar protective role of food-specific IgG and IgG subclasses in adults has not been confirmed [15,21]. Finally, the study does not provide an answer to the question of whether food-specific IgGs play a protective role or an allergic role in adults. The immunological mechanisms by which food-specific IgGs are induced are also unclear. A complete understanding of the importance of the presence of food-specific IgGs requires further accumulation of epidemiologic and experimental data.
In summary, the present study demonstrates for the first time that food-specific IgG serum concentrations are variable in both healthy and symptomatic Chinese adults. These findings raise awareness that demographic factors, types of food and the presence or absence of specific chronic symptoms should be considered before food elimination treatment based on IgG testing in patients with chronic symptoms is used in clinical practice. Further elucidation of the role played by food-specific IgGs in food allergies and tolerance should be an objective of future studies. Table S1 Means and standard deviations of log-transformed food-specific IgG concentrations. *The means and standard deviations of log-transformed food-specific IgG concentrations were calculated after the undetectable food-specific IgG concentrations were set as the half-value of the limit of detection. {The means and standard deviations of log-transformed foodspecific IgG concentrations were calculated using maximum likelihood estimation. (DOC) Table S2 Correlation of food-specific IgG concentrations among 14 foods*. *The Pearson correlation was used to analyze possible correlations among the food-specific IgG concentrations for 14 foods. **Indicates that the correlation is significant at the 0.01 level (2-tailed). (DOC)