Introduction

Celiac disease (CD) is associated with increased susceptibility to certain viral and bacterial infections [1,2,3,4,5]. Herpes zoster (HZ), colloquially known as shingles, is caused by the reactivation of the varicella-zoster virus and affects roughly 1 million people in the United States (US) each year, with a lifetime risk of 30% [6]. It classically manifests as a cluster of painful vesicular skin lesions in a dermatomal distribution [7]. While HZ pain lasts 2 to 4 weeks, Postherpetic neuralgia (PHN), a common HZ consequence, is characterized by debilitating persistent pain that lasts 3 months or more following an outbreak of shingles [7, 8]. This condition is linked with enduring and often unresponsive neuropathic pain, and patients may encounter multiple types of discomfort such as continuous deep, throbbing, or burning sensations [8]. While tricyclic antidepressants or gabapentinoids can be employed to manage neuropathic pain in PHN patients, these drugs come with significant side effect profiles that might impede long-term patient care [8]. Moreover, as current antiviral medications do not prevent the emergence of PHN, individuals at risk of developing HZ should be immunized to prevent PNH.

HZ is a vaccine-preventable disease (VPD), and the most prevalent risk factors include age ≥ 50 years, emotional distress, immunocompromised state, and immunosuppression [7, 9]. Currently, the US Centers for Disease Control and Prevention (CDC) recommends HZ vaccination of healthy individuals ≥ 50 years or individuals ≥ 19 years who are immunodeficient or immunosuppressed because of disease or therapy [6, 10]. Current CDC guidelines do not recommend vaccination in patients with CD younger than 50 years old (e.g., early vaccination).

Clinical experience suggests an increased risk of HZ in patients with CD, but robust studies are limited. A large study from Sweden suggests a 1.62-fold increased risk of HZ in adults with CD [9]. Lack of mucosal healing after treatment with a gluten-free diet was not associated with increased risk of HZ infection and the absolute frequency of HZ infection was low [11]. While Swedish research found a greater relative risk of HZ infection in CD patients, the absolute risk was modest. It is uncertain if CD patients < 50 years are more likely to develop HZ [9]. This gap of knowledge is crucial to inform about the possible benefit of routine early vaccination in patients with CD. Our hypothesis is that increased risk of HZ infection in CD is limited to patients ≥ 50 years. Therefore, our aim was to determine the risk of developing HZ infection in individuals with CD who were < 50 or ≥ 50 years old.

Methodology

Data Source

All clinical data is compiled from the TriNetX research network (Cambridge, MA, USA). This globally federated health research network and provides clinical information from 72 heath care organizations (HCOs), and over 60 million patients located within the United States. All clinical data is de-identified and aggregated directly from electronic medical records (EMR) of participating HCOs. All information relating to participating HCOs are kept anonymous. Participating HCOs include a mix of inpatient, outpatient, and specialty care services, with a typical participating HCO being a large academic health center consisting of main locations, satellite hospitals, and outpatient clinics.

Selection of Patients

The presented data were queried on 8/14/2023, following the criteria provided by TriNetX to perform a retrospective cohort study on adults (aged ≥ 18 years). CD was defined as patients with the ICD-10 code K90.0 [Celiac disease] and confirmed positive Celiac serology (elevated tissue transglutaminase IgA or deaminated gliadin IgA or endomysium IgA). Every patient included had an additional visit at six months and the K.90 ICD-10 diagnosis code was mentioned at least twice. Patients with confirmed negative serology and lacking CD ICD-10 codes served as controls. HZ infection was defined as patients with the ICD-10 code B02 [Herpes zoster]. Patients who had HZ before CD were not eligible. “Zero time” was defined as CD diagnosis in cases and negative CD testing in controls. We formed two sub-cohorts, those aged < 50 (cohort 1) and aged ≥ 50 years (cohort 2), and examined a 10-year follow-up period for the development of HZ infection.

Statistical Analyses

To account for any confounding variables, we used a greedy nearest-neighbor matching technique to accomplish 1:1 propensity score matching (PSM) for demographics and other common immune deficiency diseases on 14 characteristic(s). In the demographics category patients were matched on age at index, sex (male or female), ethnicity (not Hispanic or Latino/Hispanic or Latino), and race (Caucasian, African American, American Indian or Alaska Native, Asian, Native Hawaiian or Other Pacific Islander) characteristic(s). The patients were also matched on the following co-morbidities: Type 1 diabetes mellitus (T1DM), HIV/AIDS, ulcerative colitis, Crohn’s disease, common variable immunodeficiency (CVID), IgA deficiency, rheumatoid arthritis (RA), asthma, chronic obstructive pulmonary disease (COPD), and autoimmune hepatitis (AIH). The above conditions were defined based on ICD-10 diagnosis codes. For continuous data, we performed independent t-tests. For categorical data (presented as frequencies and percentages), we performed chi-square tests. Risk ratios (RR) with 95% confidence intervals (CI) were reported. All tests were two-tailed with an alpha level of 0.05.

Results

We included 12,217 patients with CD and 273,606 controls without CD. For cohort 1 (< 50 years), we identified 6837 patients with CD and 156,911 controls without CD. After PSM, both groups had 6826 patients, mean age at index event being 25 years. Patients were more likely to be female, and non-Hispanic Caucasians in both groups. There were no other noteworthy differences between the CD and control groups, and both groups had similar number of patients with immunocompromised status or on immunosuppression (T1DM, HIV/AIDS, IBD, CVID, RA, asthma, COPD, and AIH) Table 1.

Table 1 Baseline characteristics
Full size table

For cohort 2 (≥ 50 years), we identified 5380 patients with CD and 116,695 controls without CD. Following PSM, both groups consisted of 5337 patients, mean age at index event being 61 years. In both groups, patients were more likely to be female and non-Hispanic Caucasians. There were no other noteworthy differences between the CD and control groups, and both groups had similar number of patients with immunocompromised status or on immunosuppression (T1DM, HIV/AIDS, IBD, CVID, RA, asthma, COPD, and AIH) Table 1.

Both cohorts were followed for 10 years after “zero time” for development of HZ. In cohort 1, 62 CD patients developed new HZ infection as compared to 57 controls, RR: 1.09 (CI: 0.76–1.56, p-value = 0.64). In cohort 2, 200 CD patients as compared to 159 controls developed HZ, RR: 1.2 (CI: 1.02–1.54, p-value = 0.03) Table 2.

Table 2 Clinical outcomes
Full size table

Discussion

The study's main findings were that CD patients < 50 years had no significant difference in the risk of developing HZ when compared to referents without CD, whereas CD patients ≥ 50 years had a slightly higher risk (RR 1.2). Our findings in a US population are consistent with a prior Swedish study in terms of the modestly increased risk of HZ in CD.

CD is a chronic immune-mediated condition associated with increased risk of sepsis, pneumococcal infection, influenza, and tuberculosis [2, 3, 12]. CD patients also have a reduced or absent immune response to some immunizations against infectious viruses, such as hepatitis B [13]. Costantino et al. showed that although the majority of CD patients had a positive attitude towards vaccination, a minority (~ 20%) reported getting vaccinated against VPD as potentially harmful to their chronic illness [14]. Current guidelines suggest that patients with CD should be vaccinated against pneumococcal disease to prevent infection [5]. CD patients faced a significantly higher hospitalization risk from COVID-19 infection compared to controls, but COVID vaccination was effective in lowering this risk [3]. Our findings support the CDC recommendations and showed that CD patients ≥ 50 years have a greater risk of HZ than those who do not have CD. We were unable to elucidate a significantly higher risk of developing HZ infection in CD patients < 50 years (RR: 0.97). This suggests that younger individuals may not be at increased risk of HZ infection and this group may not benefit from earlier vaccination, in the absence of other at-risk conditions.

While the specific cause of increased risk for encapsulated bacterial infections in CD is likely related to functional asplenia [1,2,3,4,5], the mechanism underlying the increased risk of developing HZ or other viral infections in CD patients is less certain, there are several theories. CD patients are more likely to be diagnosed with other autoimmune disorders, which might rise the likelihood of the varicella-zoster virus reactivating and causing HZ [1, 2, 7]. Micronutrient deficiencies such as (e.g., zinc and vitamin D) are common in CD patients and can persist even after being on a gluten-free diet for several years [15]; these deficiencies can lead to immune system suppression and virus reactivation [16, 17]. Nemes et al. showed that gluten intake in CD patients resulted in nonresponse to hepatitis B vaccination, but CD patients who were vaccinated while on a gluten-free diet generated protective immunity with the same success as healthy persons [18]. This suggests a maladaptive immune response in CD patients who have been exposed to gluten, and it calls into question the efficacy of some vaccines in CD patients who do not adhere to a rigorous gluten-free diet.

Our study has a number of strengths and limitations. We included a large sample size of 12,217 CD patients, 6837 of whom were < 50 years old and 5337 of whom were ≥ 50 years old. CD diagnosis was not limited to an ICD-code but also required a positive CD serology to mitigate the risk of misclassification. 1:1 Propensity score matching was performed to mitigate the influence of additional confounding which are at-risk conditions for HZ infection. These conditions include immunocompromised states or individuals necessitating extended immunosuppression, such as asthma, COPD, HIV/AIDS, IBD, CVID, RA, and AIH. Long follow-up, up to 10 years may decreased the risk of ascertainment bias related to potential of more frequent medical visits around the time of CD diagnosis. Limitations of our study include its retrospective nature. A further limitation is the inability to determine the number of participants who had previously been vaccinated against HZ. In the US, the single-dose varicella vaccine was introduced for children in 1995, leading to a significant reduction in varicella cases in the first decade of the program [19]. The vaccine has been linked to a reduced occurrence of HZ in those vaccinated [19], and this could be especially relevant for individuals aged 50 or younger who may have been vaccinated after birth. Estimations suggest that in 2018, approximately 34.5% of individuals aged 60 and above in the US received the HZ vaccine. The HZ vaccine has a 90% success rate in preventing shingles and PHN among adults ≥ 50 years, and this protection endured for a minimum of 7 years post-vaccination [20, 21]. Notably, among immunocompromised adults, the vaccine demonstrated efficacy ranging from 68 to 91% in preventing shingles [21]. In an effort to reduce bias related to disparities, we performed propensity score matching for demographics. The potential effect of gluten-free diet adherence after CD diagnosis on risk of HZ infection was not investigated.

In conclusion, we found that ≥ 50 years CD patients were at an increased risk of HZ. In the US, the CDC recommends routine HZ vaccination for adults ≥ 50 years. Thus, all CD patients ≥ 50 years should be educated on the necessity of HZ immunization during routine medical care because they are at a higher risk than the general population. The HZ vaccine should be recommended even if the patient had shingles in the past or received the varicella vaccine. CD patients < 50 years were at no increased risk, so early vaccination is probably not beneficial unless suffering from other immunosuppressive conditions.