Celiac Disease: Risk of Hepatitis B Infection

Abstract

Objective:

The aim of our study was to assess the response to hepatitis B virus (HBV) vaccination and risk of HBV infection in patients with celiac disease (CD).

Patients and Methods:

We performed a cross-sectional study using the National Health and Nutrition Examination Survey (NHANES) database (2009–2014) to assess the rate of HBV vaccination, immune response, and HBV infection risk in patients with and without CD. We also determined the rate of HBV infection via retrospective analysis of 2 cohorts: 1) patients seen at Mayo Clinic (1998–2021), and 2) a stable longitudinally observed cohort, the Rochester Epidemiology Project (REP; 2010–2020).

Results:

Based on the NHANES data, the rate of HBV infection in the US was 0.33% (95% CI, 0.25%−0.41%). Of 93 patients with CD, 46 (49%) were vaccinated for HBV and of the remaining 19,422 without CD, 10,228 (53%) were vaccinated. Twenty-two (48%) vaccinated patients with CD had HBV immunity and 4,405 (43.07%) vaccinated patients without CD had HBV immunity, which was not statistically different. In NHANES data there were no cases of HBV infection in patients with CD. During the study period, 3,568 patients with CD were seen at Mayo Clinic and 3,918 patients with CD were identified using the REP database. Of those patients with CD, only 4 (0.11%) at Mayo Clinic and 9 (0.23%) of the REP patients had HBV infection.

Conclusion:

The rate of HBV vaccination and immunity were similar in individuals with and without CD. Predictably, no increased risk of HBV infection was detected in CD patients. These results do not support screening and revaccination practice for HBV immunity in patients with CD within the US.

INTRODUCTION

Celiac disease (CD) is a common immune-mediated enteropathy that affects up to 1% of the general population.¹ Most patients are diagnosed based on positive sensitive serologic markers like anti-tissue transglutaminase immunoglobulin A followed by a confirmatory small bowel biopsy.^2,3

Newly diagnosed patients with CD usually undergo screening for nutritional deficiencies and anemia along with screening for common comorbidities like thyroid disease and liver disease. It has been suggested that patients with CD have lower immune response to hepatitis B virus (HBV) vaccination.^4–12 Current recommendations call for assessing serologic evidence for HBV vaccine-induced immunity at the time of new CD diagnosis.^4–13 Not all authorities, however, identify patients with CD as having a particular problem with HBV vaccination; the Advisory Committee on Immunization Practices accepts documentation of the receipt of the vaccine series as superior to serologic evidence as proof of immunity in CD patients as it does in most recipients.¹⁴

The presumed lower rate of HBV vaccine response in CD patients has been attributed to many factors, including genetic polymorphism in major histocompatibility complex, active gluten consumption, and the inflammatory state associated with active CD.^9,10 Individuals who are homozygous for certain human leukocyte antigen (HLA) haplotypes, including HLA-DQ2 (which is disproportionately higher in patients with CD), are more likely to be hyporesponders or nonresponders to HBV vaccination.⁹ Also, consumption of gluten at the time of vaccination may blunt the ability of the immune system to present antigenic components of the vaccine effectively due to competition with gluten on the HLA -DQ2 binding site.¹⁵ Based on these observations, CD experts have advocated for screening and revaccinating patients with CD following initiation of gluten-free diet when their disease is in remission, as reflected by negative celiac serology titers.^6,7,9

According to estimates by the World Health Organization, HBV caused 257,000 chronic infections (3.5% of the global population) and 887,000 deaths in 2015.¹⁶ In the United States, there are 2 million people with chronic HBV infection. Vertical transmission is responsible for most cases worldwide.¹⁷ Since the introduction of universal HBV vaccination in children, the prevalence of chronic HBV infection has been reduced dramatically.¹⁶

According to current literature, there is no clear evidence of increased risk of HBV infection in patients with CD despite the presumed attenuated response to HBV vaccine, as measured by hepatitis B surface antigen antibody (anti-HBs) titers. Based on that paucity of evidence, we aimed to assess if there is an increased risk of HBV infection in patients with CD using the National Health and Nutrition Examination Survey (NHANES) data, the Mayo Clinic electronic health record (EHR), and the Rochester Epidemiology Project (REP) data.

PATIENTS AND METHODS

We performed a cross-sectional study using the NHANES database between 2009 and 2014 to assess HBV vaccination rate, immunity response to vaccination, and the risk of HBV infection in patients with CD. We also performed a retrospective study using Mayo Clinic EHRs between 1998 and 2021 and the REP data between 2010 and 2020 to assess the rate of HBV infection in CD patients.

This study was approved by the Mayo Clinic and Olmsted Medical Center Institutional Review Boards.

NHANES Data

The NHANES database is an ongoing cross-sectional study that is conducted every 2 years by the National Center for Health Statistics of the Centers for Disease Control and Prevention. It is a complex, stratified, multistage, probability-based sampling design that appropriately represents the noninstitutionalized US civilian population. Using the NHANES data, we evaluated the rate of HBV vaccination, HBV immunity, and the rate of HBV infection in individuals (≥ 6-year-old) with and without CD. The study was conducted utilizing the pooled NHANES data between 2009 and 2014. During the study time, 22,277 of 25,547 respondents, whose serum samples were tested for CD, were used as the CD cohort. The analytic sample was restricted to 19,515 participants who responded to the question of HBV vaccination and tested for HBV infection. The detailed methods for testing CD were described elsewhere.¹⁸

Vaccination history of HBV was obtained from the NHANES immunization questionnaire (IMQ020): “Have you ever received the 3-dose series of the hepatitis B vaccine?” The response for this question was classified into 2 groups: 1) yes, or 2) no. To define the HBV immunity by serology, the NHANES laboratory tested for HBsAg and anti-HBs. The hepatitis B core antibody (anti-HBc) test is performed on all examined participants aged ≥ 6-year-old while the HBsAg is tested only when the anti-HBc is positive. A combination of markers from serologic testing was used to determine the status of HBV vaccination, immunity, and infection. Susceptibility to HBV infection was considered if the patient had negative HBsAg, anti-HBs, and anti-HBc. Immunity due to HBV vaccination was considered if patients had positive Anti-HBs along with negative HBsAg and anti-HBc, while immunity due to previous infection was considered when patients had negative HBsAg along with positive anti-HBs and anti-HBc. Finally, patients were considered infected with HBV if they had negative anti-HBs along with positive HBsAg and HBcAg.¹⁹

Mayo Clinic EHR Data

We retrospectively searched Mayo Clinic EHRs between 1998 and 2021 for all patients (children and adults) with a diagnosis of CD. Then we identified those with comorbid HBV infection using International Classification of Diseases, 9th revision (ICD-9) and ICD-10 codes. Patient demographics (age, sex, and ethnicity) were abstracted.

REP Data

The REP is a unique medical record-linkage system that provides longitudinal epidemiological data about the residents of counties of southeastern/south-central Minnesota and west-central Wisconsin. As of 2010, the REP was expanded to include a total of 27 contiguous counties in Minnesota and Wisconsin. The REP has electronic data that include demographic information, diagnostic and procedural codes and reports, and anthropometric measures including weight, height, and body mass index. The data from the REP have been shown to represent residents of the Minnesota-Wisconsin region and a significant portion of the US population. Based on the findings from previous studies, results of studies using the REP can be generalized to the US population.^20–22 We retrospectively searched the REP data between 2010 and 2020 for all patients (children and adults) residing in the 27 counties covered by REP with the diagnosis of CD, HBV infection, and both CD and HBV infection using ICD-9 and ICD-10 codes.

RESULTS

NHANES Findings

Among NHANES participants (aged ≥ 6 years), CD prevalence was 0.71 % (95% CI, 0.53%−0.90%), while HBV infection prevalence was 0.33% (95% CI, 0.25%−0.41%). Using the NHANES data, 19,515 were included in the study and of those, 93 carried the diagnosis of CD, leaving 19,422 without CD. There were 46 (19%) individuals with CD who were vaccinated for HBV and 10,228 (53%) in individuals without CD vaccinated. When comparing the immune response to HBV between vaccinated individuals with CD and those vaccinated without CD, we found that 22 (48%) individuals with CD and 4,445 (43%) of individuals without CD had developed immunity reflected by anti-HBs titers >10 mIU/mL, which was not statistically significant (relative risk ratio= 1.01; 95% CI, 0.53–1.92; P=0.98). Interestingly, no cases of HBV infection were found in individuals with CD. (Tables 1 and 2)

Table 1.

The RRR for Immunity Against HBV Vaccination or Infection According to CD Status: NHANES Database (2009–2014)

	No CD N=19,422	CD N=93	RRR (95% CI), P value
Vaccination status, No. (%)	10,228(53%) (10228/19422)	46 (49%) (46/93)
Susceptiblea to HBV, No. (%)	5,788(56.6%) (n=10,228) (5788/10228)	24 (52.2%) (n=46) (24/46)	1.0 (Ref)
Immunity by vaccination or infectionb, No. (%)	4,405 (43.1%) (n=10,228) (4405/10,228)	22 (47.9 %) (n=46) (22/46)	1.01 (0.53–1.92), P=0.98
HBV infectionc, No. (%)	35 (0.34 %) (35/10,228)	0	N/A

Abbreviations: anti-HBc, hepatitis B core antibody; anti-HBs, Hepatitis B surface antibody; CD, celiac disease; HBsAg Hepatitis B surface antigen; HBV, hepatitis B virus; N/A, not applicable; NHANES, National Health and Nutrition Examination Survey; RRR, relative risk ratio.

^aNegative HBsAg and negative anti-HBs and negative anti-HBc.

^bIf negative HBsAg and positive anti-HBs and negative anti-HBc OR if negative HBsAg and positive anti-HBs and positive anti-HBc.

^cNegative anti-HBs and positive anti-HBc.

Table 2:

Patients With CD and Comorbid HBV Infection Irrespective of Immunization Status: Mayo Clinic, REP, and NHANES Data

Age range	Mayo Clinica (CD patients, n=3,568)		REPb (CD patients, n=3,918)		NHANESc (CD patients, n=93)
	Male	Female	Male	Female	Male	Female
0–18	174	342	252	377	5	20
19–65	675	1,655	697	1,832	22	31
65+	330	392	279	481	4	11
HBV Infectiond	4 (0.11%) (4/3568)		9 (0.23%) (9/3918)		0

Abbreviations: anti-HBc, hepatitis B core antibody; anti-HBs, Hepatitis B surface antibody; CD, celiac disease; HBV, hepatitis B virus; NHANES, National Health and Nutrition Examination Survey; REP, Rochester Epidemiology Project.

^aMayo Clinic electronic health record (1998–2021).

^bREP data, 2010–2020. A total of 27 counties provide data to the REP, 19 in Minnesota (Blue Earth, Brown, Dodge, Faribault, Fillmore, Freeborn, Goodhue, Houston, Le Sueur, Martin, Mower, Nicollet, Olmsted, Rice, Steele, Wabasha, Waseca, Watonwan, Winona) and 8 in Wisconsin (Barron, Buffalo, Chippewa, Dunn, Eau Claire, La Crosse, Pepin, Trempealeau).

^cNHANES database, 2009–2014.

^dNegative anti-HBs and positive anti-HBc.

Mayo Clinic EHR Findings

By searching EHRs between 1998 and 2021, we identified 3,568 patients with confirmed CD diagnosis who were seen at Mayo Clinic. Of those, 2,389 (67%) were women and 3,294 (92%) were White. Of the identified patients with CD, 516(14%) were aged less than 18 years, 2,330 (65%) were aged 19 to 65 years, and 722 (21%) were 65 years or older. During the study period only 4 patients out of 3,568 (0.11%) with CD disease were found to have comorbid HBV infection. In all 4 cases with both diagnoses the HBV infection predated the diagnosis of CD and was a result of clear risk factors for the acquisition of HBV infection, including adoption from or residence in areas with high endemicity (eg, Thailand), intravenous drug use, and blood transfusion prior to 1991 (Table 2).

REP Findings

In the period between 2010 and 2020, 749,375 individuals resided in the 27 counties included in the REP database. We identified 3,918 (0.5%) with CD and 2,020 (0.27%) with HBV infection. Of the 3,918 individuals with CD, 2,690 (69%) were women, 629 (16%) were aged less than 18 years, 2,529 (65%) were aged 19 to 65 years, and 760 (19%) were older than 65 years. Of the 3,918 individuals with CD, only 9 (0.23%) were found to have comorbid HBV infection (Table 2).

DISCUSSION

It has been long assumed that patients with CD have lower immunity response to the HBV vaccination when compared to the general population. This lower immunity rate has been attributed to over representation of HLA-DQ2 in patients with CD, gluten consumption at the time of vaccination, and by extension, CD activity. According to the Evidence-Informed Expert Recommendations for the Management of Celiac Disease in Children, published in 2016, there are currently somewhere between 600,000 and 1,400,000 patients with CD in the US who are potentially nonimmune and represent a public health risk.^12,23,24 It is important to note that there is significant variability in clinical practice adopted by providers caring for children and adults with CD regarding the need for HBV immunity assessment, timing of the assessment, and number of/timing of booster doses given for revaccination.^8,25,26

This variability in practice may in part be due to the paucity of compelling scientific evidence to argue for or against the current recommendation. Our study addresses both the issue of HBV immunity and the risk of infection in CD patients. The relation between CD and HBV infection has been examined in the literature and 3 studies are worthy of discussion. One study found no increased risk of CD in 60 patients who acquired HBV infection during childhood.²⁷ On the other hand, studies from Brazil and Iran ^28,29 reported an increased risk of CD in patients with HBV infection. Both studies recommended screening patients with HBV infection for CD at the time of diagnosis.^28,29 However, we could not find any study that addresses the prevalence of HBV infection in CD cohorts.

The rate of HBV infection in patients with CD was lower than the national prevalence in all 3 studied data sets (0.33% national average v. 0%, 0.11%, and 0.23% among patients with CD in the studied cohorts).

Our study found similar rates of HBV vaccination and immune response to HBV vaccination between vaccinated individuals with and without CD, resulting in no increased risk of HBV infection rate in patients with CD. This finding can be attributed to a few reasons, including the lower rate of endemicity of HBV infection in the US (prevalence of <2% of the population) and poor performance of anti-HBs titer as a measure of long-lasting immunity after receiving the HBV vaccine series. In support of the effect of HBV infection endemicity, out of the 4 patients in the Mayo Clinic EHR identified with CD and chronic HBV infection, 2 resided and/or acquired the infection in areas with high endemicity (Southeast Asia).

Second, the presumed increased risk of HBV infection in patients with unprotective titers of anti-HBs might be inaccurate, as only 16% of HBV vaccine recipients have measurable protective levels (≥ 10 mIU/mL) by age 18 years and the rate of HBV infection in the US has remained stable.³⁰ Many studies doubted the long-term protection provided by the HBV vaccine, but more recent studies have debunked the claims of waning immunity.^31–33 In a study by Simons et al³⁴, 44 participants who received the HBV vaccine in 1981 were followed for 32 years. Of the 44 participants, 13 had an anti-HBs level of ≥10 mIU/mL (group 1) while 31 had levels <10 mIU/mL (group 2). The study concluded that both groups had persistent T-cell immunity against hepatitis B irrespective of serum anti-HBs levels. None of the participants had evidence of breakthrough infection despite the higher endemicity of HBV infection in Alaska compared to other regions in the United States.³⁴ The findings by Simons et al were corroborated by other studies.³⁵,³⁶

There are several disadvantages to the current practices, particularly checking anti-HBs titers at the time of diagnosis of CD and revaccinating those with unprotective levels (<10 mIU/mL). Disadvantages include, but are not limited to, the additional cost of HBV immunity assessment and the inconvenience of a blood draw and revaccination, particularly for children. In light of our findings and the lack of compelling evidence supporting such practices, screening patients with CD for HBV immunity and revaccinating should be individualized based on the presence of other risk factors (e.g., occupation, endemicity of hepatitis B in the community, intravenous drug use).

The limitations of our study include its retrospective nature and limited generalizability to other communities where HBV infection is more endemic. Also, we elected to keep the potential overlap between the Mayo Clinic and REP cohorts to emphasize the consistency of our findings in different clinical settings.

CONCLUSION

The rate of vaccination and immunity for HBV is similar in individuals with and without CD. There is no evidence of an increased risk of HBV infection in patients with CD in the US. In the small subset of patients with CD and HBV infection, other risk factors were identified. These results do not support screening in the US for hepatitis B immunity and revaccination in patients with CD.

The Scientific Publications staff at Mayo Clinic provided copyediting support.

What is known:

• Patients with celiac disease (CD) were thought to have a suboptimal response to the hepatitis B virus (HBV) vaccine. Due to concerns for weakened immunity, current guidelines recommend screening patients with CD for immunity against HBV at the time of diagnosis and revaccinating those deemed to be at risk for HBV infection.

What is new:

• In a national cross-sectional study and 2 retrospective analyses of a clinical cohort and community-based cohort, HBV vaccination and immunity rates were similar between patients with and without CD.

• CD does not appear to be a risk factor for hepatitis B infection and therefore is not an indication for HBV revaccination. Testing for HBV vaccine nonresponse appears unnecessary in patients with CD.