Gluten Intake Interferes With the Humoral Immune Response to Recombinant Hepatitis B Vaccine in Patients With Celiac Disease
OBJECTIVE. Patients with celiac disease, who often carry human leukocyte antigen-DR3;DQ2, are prone to inadequate response to hepatitis B immunization. We evaluated vaccine response in relation to disease activity and whether previous treatment with a gluten-free diet influences the achievement of protective antibody titers.
PATIENTS AND METHODS. We studied 128 children and adolescents with celiac disease and 113 age-matched control subjects. Twenty-two patients with celiac disease were prospectively immunized after diagnosis during dietary treatment (group 1). A total of 106 (group 2) and the control subjects received vaccination by mass immunization in schools at 14 years of age regardless of diet status and when celiac disease was still undiagnosed in 27 of these children. Diet compliance and celiac disease activity were monitored by measurement of antibodies against transglutaminase and endomysium. Vaccine response was determined by measuring antihepatitis B antibodies from serum.
RESULTS. The seroconversion after hepatitis B vaccination was 95.5% in group 1. All of these patients carried human leukocyte antigen DQ2. The response rate in group 2 was 50.9% and correlated with gluten intake (untreated patients: 25.9%, non-strict diet: 44.4%, strict diet: 61.4%). Treated and compliant patients did not significantly differ from control subjects (75.2%). Thirty-seven antihepatitis B–negative patients with celiac disease received a booster during a controlled gluten-free diet, and 36 (97.3%) seroconverted, irrespective of the presence of human leukocyte antigen DQ2.
CONCLUSIONS. Nonresponse to recombinant hepatitis B surface antigen may be a sign of undiagnosed celiac disease. However, there is a good vaccine response in adequately treated patients. Human leukocyte antigen DQ alleles do not seem to have a primary role. Revaccination is recommended during a controlled gluten-free diet.
- celiac disease
- gluten-free diet
- hepatitis B immunization
- vaccine nonresponse
- endomysium antibody
- transglutaminase antibody
- disease activity
Chronic hepatitis B virus (HBV) infection is a major global health problem that affects >350 million people worldwide, causing acute and chronic liver disease, cirrhosis, and hepatocellular carcinoma.1 The introduction of highly effective and safe vaccines against HBV is considered the main strategy for control of the infection and viral transmission.2 Implementation of routine HBV vaccination of the population reduced the incidence of infection, but vaccine nonresponsiveness is an important potential challenge. The mass immunization of populations has been recommended by the World Health Organization since 1991.3 Hungary introduced the routine HBV vaccination of all 14-year-old schoolchildren in 1999.
Since the introduction of HBV vaccination in 1982, many epidemiologic studies have been conducted to determine the efficacy of the vaccine, and specific antibody levels ≥10 IU/L are commonly considered as protective against HBV infection.4–6 Approximately 4% to 10% of healthy, immunocompetent individuals fail to elicit protective levels of antibodies to recombinant hepatitis B surface antigen (HBsAg) after completing the standard HBV vaccination schedule.7,8 The exact percentage depends on the definition of nonresponse or hyporesponse. Because nonresponders to the vaccine are at risk for HBV infection, mechanisms underlying an inadequate immune response to recombinant HBsAg are important to clarify.
Many nongenetic factors, including age, obesity, smoking, drug abuse, alcoholism, infections, immunosuppression, and route of vaccination, are known to be associated with nonresponsiveness.9–13 Human leukocyte antigen (HLA) phenotype is considered the most important genetic marker of nonresponse. The immune response to HBV vaccine is largely determined by presentation of the immunogenic peptides via HLA-DR and DQ molecules,14,15 and the DR3;DQ2 or DR7;DQ2 haplotypes predispose to a lower response rate.16–19
Celiac disease (CD) is an HLA-associated disease where an abnormal immune response to wheat gluten drives pathology in the small intestine characterized by villous flattening and malabsorption.20 The intestinal damage is caused by interactions between specific deamidated glutamine residues of gliadin and HLA-DQ2 (DQA1*05/DQB1*02) or DQ8 (DQA1*03/DQB1*0302) molecules. This binding induces the proliferation of intestinal T lymphocytes and the production of autoantibodies against type 2 transglutaminase (TG2) found in endomysial structures. Malabsorption symptoms are not always present, and the diagnosis might be delayed until adulthood. The abnormal immune response stops after the exclusion of gluten from diet, TG2 autoantibodies disappear, and the intestinal mucosa heals. Patients with CD are clinically normal as long as they follow a strict gluten-free diet. HLA-DQ2 is present in 90% to 95% of patients with CD.20,21 Because HLA-DQ2 is a marker of the HBV vaccine nonresponder state, CD may be associated with nonresponsiveness or hyporesponsiveness to the HBV vaccine. Patients with CD with an earlier history of HBV vaccination were found to have low rates of protective antibodies,22,23 but it was never investigated whether failure to respond to HBV vaccination was related to the active disease or exposure to gluten.
In this study we examined the specific antibody production to HBsAg (antihepatitis B surface antibody [anti-HBs]) in patients with CD after prospective HBV vaccination after treatment and the possible role of dietary gluten in the subnormal protective immune response.
PATIENTS AND METHODS
A total of 128 patients with small-intestinal biopsy-proven CD and 113 age-matched nonceliac subjects who received HBV vaccination during the same time period were included in the study. All of the nonceliac control subjects were negative for blood autoantibody markers of CD. The patients were investigated and followed-up at the Department of Pediatrics, Medical and Health Science Center, University of Debrecen; the Department of Gastroenterology-Nephrology, Heim Pál Children's Hospital; the Department of Pediatrics, Markusovszky Hospital; and the Department of Infectology, Hetényi Géza County Hospital. The diagnosis of CD was based on the presence of severe villous atrophy in a small-bowel biopsy sample and the clinical and/or histologic response to a gluten-free diet.
Study Group 1: Vaccination and Sample Collection
From April 2004 to March 2006, 22 patients with CD (11 girls and 11 boys) were prospectively vaccinated with a recombinant HBV vaccine (Engerix B, GlaxoSmithKline, Rixensart, Belgium) after the diagnosis of CD during dietary treatment. Informed consent was obtained from all of the parents. The age of the subjects ranged from 4.0 years to 12.5 years (median: 8.8 years), and they were at the start of the vaccination on a gluten-free diet for 5.0 months to 4.5 years (median: 12.5 months). Blood samples were taken from all of the patients to check for HBV markers (HBsAg, hepatitis B envelope [HBe], anti-HBe, anti-HBs, and antihepatitis B core antibodies) and celiac autoantibodies before the HBV vaccination. The vaccination schedule was 0, 1, and 6 months with 10 μg of HBV vaccine, intramuscularly. Blood samples for anti-HBs and celiac autoantibody determinations were collected 4 weeks after the second and third HBV vaccinations and stored at −20°C until tested.
Study Group 2: Vaccination and Sample Collection
Another group of 106 patients with CD (67 girls and 39 boys) received a recombinant HBV vaccine (Engerix B, GlaxoSmithKline, or H-B-VAX II, Merck Sharp & Dohme, Whitehouse Station, NJ) between September 1999 and March 2006, regardless of diagnosis and diet status, as part of the national vaccination program at the age of 14 years in their schools. Of these children, 79 had been diagnosed with CD before their vaccinations and were advised to maintain a gluten-free diet. In the remaining 27 children (25.5%), CD diagnosis was established only after the vaccinations; thus, these children consumed unrestricted amounts of gluten at the time of vaccination.
Three doses of either 10 μg of Engerix B or 5 μg of H-B-VAX II or 2 doses of either 20 μg of Engerix B or 10 μg of H-B-VAX II were provided by the National Public Health Service for vaccinations at 0, 1, and 6 months or at 0 and 6 months, respectively. These vaccination schedules were established on the basis of previously published efficacy studies with the respective vaccines as observed in other populations. At the time of this study, no data were available on the response rate of healthy Hungarian children in the investigated age group. Blood samples for anti-HBs and celiac autoantibody determinations were collected at a median age of 16.7 years, with variable time intervals after the vaccinations (mean time: 28 months [range: 2–75 months]).
Study Group 3: Vaccination and Sample Collection
A total of 113 age-matched subjects (70 girls and 43 boys) were enrolled as control subjects. They received their routine HBV vaccines in the same years and in an identical fashion as study group 2. All of these subjects were negative for CD-specific autoantibodies. Blood samples for anti-HBs determination were collected at variable time intervals after the routine HBV vaccination (median age: 16.1 years). The mean duration for testing for anti-HBs was 23 months (range: 4–52 months).
Study Group 4: Vaccination and Sample Collection
An additional vaccine dose was offered to subjects without protective anti-HBs antibodies. The booster dose was administered during a gluten-free diet to 37 participants of study group 2 with CD. The booster dose was 20 μg of recombinant HBsAg (Engerix B, GlaxoSmithKline) administered intramuscularly. Blood samples for anti-HBs determination were obtained 4 weeks after the booster dose.
Determination of Serum Antibodies Against Recombinant HBsAg
Anti-HBs antibodies were measured by enzyme-linked immunosorbent assay by using commercially available assay (Hepanostica anti-HBs, bioMérieux, Boxtel, Netherlands and titers were reported in international units per liter. A seroconversion rate was defined by anti-HBs ≥10 IU/L. Patients with an anti-HBs titer between 10 and 100 IU/L were defined as low responders, and those with titers >100 IU/L were defined as high responders.
Determination of CD-Specific Autoantibodies
All of the participants were screened for the presence of immunoglobulin A (IgA) class serum antibodies against endomysium antibody (EMA) by indirect immunofluorescent assay,24 and total serum IgA was determined. All of the control subjects had normal serum total IgA levels. IgA class antibodies against TG2 were measured by enzyme-linked immunosorbent assay with human recombinant TG2 expressed in Escherichia coli25 after the incubation protocol described in ref 24. The cutoff value for anti-TG2 positivity was 5 U/mL. In subjects with total serum IgA <0.2 g/L, IgG class EMA and TG2 antibodies were investigated and interpreted instead of IgA class antibodies.26 EMA and TG2 antibodies were serially determined in patients with CD at least at yearly intervals after the diagnosis as noninvasive markers for diet compliance and disease activity. EMA results were available from 1988 onward, whereas TG2 antibody results were available only from 2002. Concordance for EMA and TG2 antibody positivity was 99.2% in our laboratory.21
All of the subjects in study group 1 and those with available whole blood samples from group 2 were typed for HLA-DQ2 and HLA-DQ8 alleles. HLA-DQA1 and HLA-DQB1 were typed by polymerase chain reaction-based techniques with sequence-specific primers (Genovision, Oslo, Norway).
The χ2 test was used to compare whether there was significant difference in proportions of protective anti-HBs titers between different groups of subjects. Results were regarded as statistically significant when the P value was <.05. The confidence intervals (CIs) for proportions were constructed by using the Wilson procedure with a confidence level of 95%. To examine whether there is an association between anti-HBs concentration and time to testing after the HBV vaccination, we conducted a statistical test that estimates Pearson's product moment correlation coefficient and computes a test of the value being 0. The null hypothesis is that the correlation coefficient is 0.
All of the patients in the prospectively immunized study group 1 were negative for HBV markers (HBsAg, HBe antigen, anti-HBe, anti-HBs, and antihepatitis B core antibodies) and showed a decline or negativity for celiac autoantibody markers (EMA and anti-TG2) before the start of the HBV vaccinations (Fig 1). There were no adverse effects because of the vaccinations during the study. The seroconversion rate for anti-HBs was 95.5% (95% CI: 78.2%–99.2%) after vaccination. Among the 21 responders, 19 were high responders, and 2 were low responders. Anti-HBs was ≥10 IU/L in 13 (59.1% [95% CI: 38.7%–76.8%]) of the subjects after the second vaccine dose (Fig 2). All of the subjects were either homozygous or heterozygous for DQ2, and 20 (90.1%) achieved negative celiac autoantibody status at the completion of the vaccination (Fig 1). One patient with CD in the prospective group did not respond to basic and booster vaccinations and also remained persistently high positive for celiac autoantibodies because of dietary noncompliance (Fig 3). He was homozygous for DQ2 and also had type 1 diabetes mellitus and several nonorgan-specific autoantibodies (anti-DNA, antinuclear antibody, antineutrophil cytoplasmic autoantibody, and cytoskeleton autoantibodies). After repeated and intensified dietary educations, he finally achieved celiac autoantibody-negative status 2 years after initial diagnosis, and after that he also responded to the additional HBV vaccination.
In study group 2, the HBV vaccinations were performed in the school unrelated to CD diagnosis and dietary compliance. Protective anti-HBs concentrations were found in 54 of the 106 patients with CD (50.9% [95% CI: 41.6%–60.3%]) and in 85 of the 113 control subjects (75.2% [95% CI: 66.5%–82.3%]); the difference was statistically significant (P < .001). Thirty patients with CD were high responders, and 24 were low responders. In the control group, 52 were high responders, and 33 were low responders. The response rate among the 27 undiagnosed and, thus, untreated patients with CD was only 25.9% (95% CI: 13.2%–44.7%), which was significantly lower than in control subjects (P < .001; Fig 4). From the 79 known patients with CD, 70 followed a strict gluten-free diet and had negative EMA-TG2 antibody status before vaccination or achieved seronegativity during the vaccination. The response rate among these treated and compliant patients with CD was 61.4% (95% CI: 49.7%–71.9%), which was not significantly lower than in control subjects (P = .102). The 9 patients with CD with frequent dietary transgressions and constantly positive celiac autoantibodies during vaccination had a 44.4% (95% CI: 18.9%–73.3%) response rate, which was intermediate between untreated and well-treated patients. Celiac autoantibody titers in these patients ranged between 3% and 50% (mean: 25%) of their initial antibody titers measured at the time of diagnosis.
The response rate in celiac male subjects was 56.4% (95% CI: 41.0%–70.7%) compared with 47.8% (95% CI: 36.3%–59.5%) among celiac female subjects, the difference not being statistically significant (P = .685). There was no significant correlation between the anti-HBs concentration and the time between vaccination and testing in control subjects; the correlation coefficient was −0.0902 (P = .348). In the case of patients with CD, the value of the coefficient was −0.2261 (P = .019). This result implies that there was a slight correlation between the anti-HBs concentration and the time until testing after the HBV vaccination. However, anti-HBs titers did not show significant reductions in individual patients with CD from whom several serum samples were available after vaccination during a time period of 2 to 3 years.
Thirty seven of the 52 anti-HBs-negative patients with CD (71.2%) who got vaccinations at school agreed to receive an additional vaccine dose, prospectively, during a controlled gluten-free diet (group 4), and 36 of them (97.3% [95% CI: 86.2%–99.5%]) seroconverted 4 weeks after revaccination (Fig 5).
HLA-DQ typing results were available from 53 patients (50%) from study group 2. DQ8 was present in 2 subjects, who both responded to initial school vaccination. All of the other patients with CD carried DQ2. The response rates after school HBV vaccination among patients with CD who were homozygous or heterozygous for DQ2 were 64.3% (95% CI: 38.8%–83.7%) and 51.4% (95% CI: 35.9%–66.6%), respectively. There was no statistically significant difference between these 2 groups (P = .407).
We observed an impaired humoral immune response to recombinant HBV vaccine in adolescents with untreated CD. However, patients with CD who were immunized prospectively during a treatment with a gluten-free diet developed protective immunity with similar success as healthy people. Because CD occurs in 1% of the population in Europe and North America,20 failure of celiac subjects to respond to HBV vaccination may also have relevance for public health policies. CD is a genetically determined lifelong intolerance to gluten and starts in infancy after regular ingestion of cereals. Despite the ongoing gluten-triggered autoimmune process, clinical symptoms may be vague or nonspecific for a long time, and 90% of patients remain undetected during childhood.27 Current HBV mass immunization campaigns recommend the vaccination of young children or adolescents; thus CD subjects in most countries will receive HBV vaccination when they are still undiagnosed and exposed to gluten. Even in the United States, where HBV vaccination is recommended in infancy, the study by Park et al23 showed that the immunization of children was completed on average only by the age of 3 years, thus after a quite long gluten exposure.
Development of the protective immune response to HBsAg is T-cell dependent and is associated with the production of specific neutralizing antibodies. Previous studies in nonresponsive but otherwise healthy people did not find defects in antigen uptake or processing by antigen-presenting cells.28,29 However, the polymorphism in the HLA region, which encodes different cell surface glycoproteins responsible for presenting protein antigens to CD4+ T cells, largely contributes to the human antibody response to HBV vaccine. Nonresponding patients may have a failure of major histocompatibility complex class II molecules in the interaction with processed protein antigen, in the stimulation of T-helper cells, or in both.30,31 Godkin et al19 investigated the binding affinities of envelope and core peptides of HBV to particular HLA glycoproteins, and their data supported the direct involvement of HLA-DR3 in HBV vaccine nonresponsiveness. The lack of response to HBsAg has been also attributed to defective or insufficient HBsAg-specific T-helper cells,32,33 inadequate T-helper 1 and T-helper 2 cytokine production,34–36 or diminished expression of cell-cell contact signals between activated T and B cells, like CD40L.37
The interactions between specific deamidated glutamine residues of gliadin and the HLA-DQ2 or HLA-DQ8 molecule in CD induce proliferation of T lymphocytes.20,38 In our study, the seroconversion rate after HBV immunization was significantly lower among untreated versus treated and compliant subjects with CD or control subjects. Both HBsAg protein fragments and gliadin peptides bind to HLA-DQ2 molecules, and their competition may result in a defective antibody response against the recombinant HBsAg vaccine in active CD.
We demonstrated that the rate of primary nonresponse to the standard regimen of recombinant HBV vaccination was surprisingly high (74.1%) in undiagnosed and untreated celiac adolescents. Noh et al22 found that the nonresponse rate to the HBV vaccine was similarly high in adult patients with CD (68%), and Park et al23 also observed a 54% nonresponse rate in celiac children. However, these retrospective studies did not analyze untreated and treated patients with CD separately. In our study, a normal responder rate (95.5%) was achieved by prospective immunization on a gluten-free diet, and the nonresponse rate was only 49.1% when HBV immunization was performed unrelated to diagnosis and diet status. Success with repeated vaccinations after controlled diet and correlation of nonresponse with celiac autoantibody positivity and diet transgressions suggest that disease activity may play a primary role in vaccination failure. Therefore, the antibody response to HBsAg should be determined in newly diagnosed, previously vaccinated patients with CD. Depending on the specific antibody titers, booster doses may be required to achieve and maintain seroprotection in these patients.
Our data suggest that HLA-DQ2 alone is not a good marker for the insufficient immune response to HBV vaccination and that the HBV nonresponder status in patients with CD can be caused by other celiac-related autoimmune problems. Furthermore, we did not find a weaker response in DQ2 homozygous than in DQ2 heterozygous patients, although it has been suggested that homozygous DR3;DQ2 patients would have less antigen presentation because of the lower binding affinity of HBsAg peptides.19 The patient in the prospectively immunized group who did not respond to the primary and booster vaccinations was homozygous for DQ2, but he also had type 1 diabetes mellitus and a poor dietary compliance characterized by persistent EMA and TG2 antibody positivity, and he responded to further vaccination after the achievement of EMA-negative status.
In our study, 75.2% of healthy control adolescents retained protective HBV antibody levels (≥10 IU/L) after primary vaccination. This value is similar to that observed after HBV vaccination in other age groups in Hungary (74% in 15-month-old healthy children and 68% in medical professionals),39 but seems to be lower compared with vaccine effectiveness studies performed at 1 month postvaccination.8 This can partly be explained by the longer time between vaccination and testing (on average: 2 years) but was not associated with the type of the vaccine or vaccination schedule (data not shown). Long-term follow-up studies demonstrated that persisting protective anti-HBs titers may decrease to 48% to 64% after 10 years in normal children after mass immunization.40,41 In adolescents with CD, we found a significantly lower rate of protective postvaccination antibody titers at similar time intervals even within groups that received the vaccination with an identical schedule. Longer time intervals between the vaccination and testing may also explain why treated and compliant patients with CD in group 2 had lower anti-HBs antibody positivity rates than the prospectively immunized group 1 evaluated at 1 month after the vaccination. We did not have anti-HBs antibody data at 1 month postvaccination in the mass immunization group; thus, we cannot rule out the possibility that untreated celiac subjects had just a faster decline of serum antibodies and may still have some immunologic memory, and that is why they responded well to the booster dose after the diet. However, it has been shown that, in pediatric liver transplant patients with vanished antibody response, the presence of memory T lymphocytes did not prevent de novo HBV infection.42
The nonresponder status to primary HBV vaccination is not permanent in CD and may improve after gluten exclusion. Revaccination is recommended for patients after treatment with a gluten-free diet.
This study was supported by the Hungarian Scientific Research Fund (OTKA K61868) and the Marie Curie Transglutaminases: Role in Pathogenesis, Diagnosis, and Therapy project.
- Accepted November 14, 2007.
- Address correspondence to Éva Nemes, MD, Department of Pediatrics, Medical and Health Science Center, University of Debrecen, Nagyerdei krt 98, H-4032 Debrecen, Hungary. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
What's Known on This Subject
Inadequate response to hepatitis B vaccine occurs in 4% to 10% of healthy people. Nonresponse is often associated with the HLA-DR3;DQ2 haplotype. Patients with CD, who often carry HLA-DR3;DQ2, are prone to inadequate response to hepatitis B immunization.
What This Study Adds
Nonresponse to recombinant HBsAg may be a sign of undiagnosed CD. HLA-DQ alleles do not seem to have a primary role. The nonresponder status is not permanent. Revaccination is recommended during a controlled gluten-free diet.
- Fisman DN, Agrawal D, Leder K. The effect of age on immunologic response to recombinant hepatitis B vaccine: a meta-analysis. Clin Infect Dis.2002;35 (11):1368– 1375
- ↵Belloni C, Avanzini MA, De Silvestri A, et al. No evidence of autoimmunity in 6-year-old children immunized at birth with recombinant hepatitis B vaccine. Pediatrics.2002;110 (1). Available at: www.pediatrics.org/cgi/content/full/110/1/e4
- ↵van Heel DA, West J. Recent advances in coeliac disease. Gut.2006;55 (7):1037– 1046
- ↵Korponay-Szabó IR, Dahlbom I, Laurila K, et al. Elevation of IgG antibodies against tissue transglutaminase as a diagnostic tool for coeliac disease in selective IgA deficiency. Gut.2003;52 (11):1567– 1571
- ↵Egea E, Iglesias A, Salazar M, et al. The cellular basis for lack of antibody response to hepatitis B vaccine in humans. J Exp Med.1991;173 (3):531– 538
- ↵Quarsten H, McAdam SN, Jensen T, et al. Staining of celiac disease-relevant T cells by peptide-DQ2 multimers. J Immunol.2001;167 (9):4861– 4868
- ↵Rusvai E, Brojnás J, Takács M, Berencsi G. Hepatitis B serologic markers in immunized babies born to hepatitis B surface antigen positive mothers. Acta Microbiol Immunol Hung.2006;53 (3):335
- ↵Ni YH, Ho MC, Lu CY, et al. Immune response to booster hepatitis B vaccines after liver transplantation in children who received primary immunoprophylaxis in infancy. Presented at: 40th annual meeting of the European Society for Pediatric Gastroenterology, Hepatology and Nutrition; May 9–12, 2007; Barcelona, Spain. Available at: www.colloquium.fr/ei/viewpdf.esp?id=210&file=e%3A%5Cevents%5Cserveur%5Ceventwin%5Cdocs%5Cpdf%5C07ESPGHAN%5Cdef%5CPH1%2D03%2Epdf. Accessed April 5, 2008
- Copyright © 2008 by the American Academy of Pediatrics