Children with celiac disease (CD) have an increased risk for influenza infection and secondary hospital admission (1). In addition, patients with CD seem to have weaker humoral responses to certain immunizations, such as hepatitis B vaccine (HBV) (2,3). To our knowledge, no data describe the response of patients with CD to influenza vaccination. The emergence of the pandemic influenza A/H1N1/09 strain presented an opportunity to investigate vaccine responses in immunologically naïve patients.
We conducted a prospective open-label field study between November and February 2009–2010, recruiting a total of 86 children, including 14 with proven CD (serology and biopsy) and 14 sex- and age-matched immunocompetent controls (4). Blood was collected at baseline (optional) and 4 to 7 weeks postvaccination. Vaccines administered were 2 monovalent, split-virion, squalene-adjuvanted vaccines containing H1N1 A/California/7/2009 X-179A: Pandemrix (GlaxoSmithKline, Rixensart, Belgium) or Focetria (Novartis, Sienna, Italy), according to official guidelines. Antibody (AB) titers were measured by hemagglutination inhibition (HAI) and microneutralization (MN) in the laboratory of the Center for Vaccinology-Neonatal Immunology, Geneva, Switzerland. An HAI titer of ≥40 was considered protective as previously described. Presently, no protective thresholds are defined for MN titers. Geometric mean titers (GMT) were calculated to compare average pre- and postvaccination titers. The present study was approved by the institutional ethics committee of the University Hospitals of Geneva and the Swiss Agency for Therapeutic Products (Swissmedic, Hallerstrasse, Switzerland); it was registered on www.clinicaltrials.gov before patient recruitment.
The mean age of 14 white children with CD was 12.9 years (interquartile range [IQR] 9.8–15.9 years), and that of 14 age- and sex-matched controls was 12.0 years (IQR 8.2–14.4 years). Their baseline characteristics are presented in Table 1. Average age at diagnosis of CD was 3.8 years (IQR 1.2–5.1 years), and average time since diagnosis was 9.1 years (IQR 5.3–13.5 years). The majority of patients were girls (N = 9, 64%). Four of the children with CD (29%) had a positive serum anti-tissue transglutaminase (tTG) immunoglobulin (Ig) A concentration, suggesting possible noncompliance with a gluten-free diet. Human leukocyte antigen (HLA) typing was available for all 14 children: 7 (50%) were heterozygous and 7 were homozygous for HLA-DQ2. At baseline, HAI-GMT (available for 6 children with CD) was 39 (95% confidence interval [CI] 8%–197%); 3 children (50%) showed protective titers (≥40), suggesting possible earlier exposure to the virus. Among controls, prevaccination blood samples were available for 11 children. HAI-GMT was 17.2 (95% CI 7%–44%), and 3 children (27%) showed protective AB titers.
The responses to influenza A/H1N1/09 vaccination are represented in Table 1. All 28 children reached protective AB titers postvaccination. Among children with CD, HAI-GMT was 530 (95% CI 333%–842%) and MN-GMT was 488 (95% CI 253%–939%). A good correlation between the 2 laboratory measurements (R 2 = 0.96, for postvaccination titers) confirmed reliable laboratory methods. Age-/sex-matched controls reached slightly higher but not significantly different GMT (HAI-GMT 573, 95% CI 369%–890%, P = 0.812; MN-GMT 625, 95% CI 339%–1153%, P = 0.592). Among children with CD, HAI titer was significantly dependent on dietary compliance (HAI-GMT of 4 children with anti-tTG IgA levels >20 IU around time of vaccination: 224, 95% CI 143%–352%; HAI-GMT of 10 children with anti-tTG IgA levels < 20 IU around time of vaccination: 747, 95% CI 461%–1209%, P = 0.028). HAI titer was not dependent on HLA status (HAI-GMT for HLA-DQ2 heterozygotes: 478, 95% CI 228%–1004%; HAI-GMT for HLA-DQ2 homozygotes: 587, 95% CI 320%–1078%; P = 0.798).
Children with CD can produce antibodies following influenza A/H1N1/09 vaccination, with GMT clearly above the protective threshold. This is important because data concerning AB responses of patients with CD to other vaccine types remain contradictory. Certain studies have reported diminished responses of patients with CD to HBV, and attributed this to HLA-DQ2, -DR3, and -DR7 serotypes (2,3). These serotypes (particularly DQ2) are strongly associated with CD, and have been independently linked to the nonresponse to HBV, especially if present in homozygous form (2,3). In our study, DQ2 homozygotes responded equally well as heterozygotes; therefore, we cannot confirm any hypothesis linking diminished immune responses to the DQ2 zygosity. Other studies have shown a good response of patients with CD to HBV if a strict gluten-free diet was followed and low disease-activity markers were present (eg, serum titers of anti-tissue transglutaminase antibodies) (5), suggesting that genetic factors play a subordinate role to environmental ones. In our study, the 4 children with high anti-tTG levels (therefore considered less compliant with gluten-free diet) did in fact reach significantly lower titers than children who were fully compliant. Competition between H1N1 protein fragments and gliadin peptides to bind to HLA-DQ2 molecules may result in a defective AB response in noncompliant children (6); however, all 3 children with possible pre-exposure to the H1N1 virus were part of the diet-compliant group, which may have led to the overestimation of vaccine effect in that group.
Regarding present evidence on other vaccine types, response rates to pneumococcal (7), rubella (2), and Haemophilus influenza type b (2) vaccines seem to be similar in patients with CD and controls. Evidence about tetanus vaccine is conflicting (2,8), and immune responses to oral polio vaccine may be stronger in patients with CD (9).
The limitations of our study include small sample size, mainly due to Swiss parents’ general resistance to H1N1 vaccination and the overlapping of vaccination with the peak of the epidemic due to delayed national vaccines’ availability leading to possible overestimation of the vaccine effect. Due to widespread parental apprehension, baseline AB sampling was not possible for all children, making accurate estimation of baseline titers difficult. Furthermore, 2 vaccines were used because of changing official recommendations; however, in our larger-scale study including more children, they were shown to be equally immunogenic (4).
In spite of its limitations, the key message of the present study remains strong and valid: pediatric patients with CD respond well to influenza vaccination and this can be used to help limit the particular risk of influenza complications in this population.
1. Marild K, Fredlund H, Ludvigsson JF. Increased risk of hospital admission for influenza
in patients with celiac disease
: a nationwide cohort study in Sweden. Am J Gastroenterol
2. Park SD, Markowitz J, Pettei M, et al. Failure to respond to hepatitis B vaccine in children
with celiac disease
. J Pediatr Gastroenterol Nutr
3. Zingone F, Morisco F, Zanetti A, et al. Long-term antibody persistence and immune memory to hepatitis B virus in adult celiac patients vaccinated as adolescents. Vaccine
4. Meier S, Bel M, L’huillier A, et al. Antibody responses to natural influenza
A/H1N1/09 disease or following immunization with adjuvanted vaccines, in immunocompetent and immunocompromised children
5. Ertem D, Gonen I, Tanidir C, et al. The response to hepatitis B vaccine: does it differ in celiac disease
? Eur J Gastroenterol Hepatol
6. Nemes E, Lefler E, Szegedi L, et al. Gluten intake interferes with the humoral immune response to recombinant hepatitis B vaccine in patients with celiac disease
7. McKinley M, Leibowitz S, Bronzo R, et al. Appropriate response to pneumococcal vaccine in celiac sprue. J Clin Gastroenterol
8. Pettingale KW. Immunoglobulin and specific antibody responses to antigenic stimulation in adult coeliac disease. Clin Sci
9. Mawhinney H, Love AH. The immunoglobulin class responses to oral poliovaccine in coeliac disease. Clin Exp Immunol