Environmental risk factors are associated with autoimmune hepatitis

Abstract Background Failure of immunologic homeostasis and resultant hepatocyte destruction in autoimmune hepatitis (AIH) is likely the result of environmental triggers within a permissive genetic architecture. Aims We aimed to identify risk factors associated with AIH in a well‐phenotyped AIH cohort. Methods We prospectively collected environmental questionnaires from 358 AIH cases and 563 healthy controls. Response frequencies were compared using logistic regression, adjusting for age at recruitment, sex and education. Results AIH cases were more likely to ever have a urinary tract infection (UTI) (53.6% vs 33.9%, P < .001) and recurrent UTI (more than 1 per year) (23.5% vs 15.9%, P = .002) compared to controls. Female cases more frequently had ever used oral contraceptives (83.0% vs 73.7%, P = .006), fewer pregnancies (median = 1 vs 3, P < .001) and less often used hormone replacement therapy compared to controls (28.5% vs 60.1%, P < .001). Current smoking was more prevalent in cases (18.9% vs 7.4%, P = .022), yet no difference according to historical smoking behaviours was observed. Finally, cases were less likely to have history of mumps (32.4% vs 53.1%, P = .011) and rheumatic fever (1.1% vs 4.4%, P = .028), but reported higher vaccination frequency to chicken pox (38% vs 28.1%), measles (66.5% vs 39.3%), mumps (58.7% vs 34.6%), rubella (55.3% vs 32.7%), pertussis (59.8% vs 40.1%) and pneumococcus (47.2% VS 39.4%) (P < .002). Conclusions Environmental factors are important in AIH pathogenesis. Replication of these findings and prospective examination may provide new insight into AIH onset and outcomes.


| INTRODUC TI ON
Broken immune tolerance and failure of immunologic homeostasis underlying autoimmune hepatitis (AIH) pathogenesis are likely the result of environmental triggers in the background of a permissive genetic architecture. Similar to other autoimmune diseases, the Human Leucocyte Antigen (HLA) locus 1 at 6p21 has been associated with AIH in prior candidate gene studies and confirmed by the only genome-wide association study. 2 A single causal variant does not exist, and further complicating the HLA-AIH association is the presence of alleles conferring disease risk and protection. 3 In addition to genetic factors, environmental triggers such as viral infections and drug exposures have been temporally associated with AIH onset. 4 For other autoimmune diseases, liver diseases, primary sclerosing cholangitis (PSC) and primary biliary cholangitis (PBC), environmental risk assessments have provided a more complete understanding of disease risk, onset and possible modifiable factors. [5][6][7] To date, such assessments in AIH are incomplete. Case series and small retrospective studies have linked a variety of drugs, specifically nitrofurantoin and minocycline, 8 and viruses such as Epstein-Barr virus, varicella zoster virus, viral hepatitis A, B and C, to AIH. 4 Unfortunately, these potentially important triggers are seldom considered systematically in clinical practice. Further evidence supports an environmental role in AIH pathogenesis, as factors such as microbiome diversity 9 and psychosocial stress 10 have also been associated with disease.
In this study, we aimed to identify associations between AIH and lifetime environmental exposures as well as lifestyles, personal and family histories using the Genetic Repository of Autoimmune Liver Diseases and Contributing Exposures (GRACE) study.

| AIH cases and controls
The GRACE study contains consecutively recruited AIH patients (cases) from Indiana University. 1,11 Details regarding the establishment of this registry have been published previously. The GRACE study was developed in 2014 to collect well-phenotyped AIH cases to pursue investigation into genetic and environmental contribution to disease development. The diagnosis of AIH among cases was based on clinically accepted criteria as per the recent clinical guidelines. 12 Controls were serially recruited from the Mayo Clinic Division of General Internal Medicine during annual visits for preventive medical examination. Control exclusion criteria included evidence of prior or current liver disease (known alcohol, fatty liver or viral hepatitis (per clinical notes, imaging or serology)) or any historical abnormal liver enzyme levels. All participants in the study had informed consent collected by study staff. This study included cohorts that adhere to the ethical guidelines of the 1975 Declaration of Helsinki, and was approved by the Institutional Review Board of Indiana University and Mayo Clinic.

| Questionnaire establishment and collection
The study questionnaire was developed by the Mayo Clinic Survey Centre and data acquired via this instrument among other autoimmune liver disease cohorts have been published previously. 5,6 The survey contains questions regarding demographics, anthropometric features, education, lifestyle and environmental exposures as well as extensive personal and familial medical history. In total, the survey includes 70 primary and 300 secondary questions covering 37 pages with responses provided using scantron technology (Table S1).
This study instrument was either given directly to or mailed to enrolled participants along with a prepaid return envelope. In the case of study participants (cases and controls) who did not return the first questionnaire, a second questionnaire was mailed within 2 months from the initial contact. At the time of study completion, the GRACE study included 512 AIH patients and 358 (70% response rate) had returned the study questionnaire. The control group included 563 participants (87% response rate).

| Statistical analysis
Information collected by the questionnaires was double entered into a SAS database. Each question was summarized by descriptive statistics for both AIH cases and controls. The P-values were calculated using logistic regression models, adjusting for age, gender and education. Statistical analysis was conducted in R version 3.6.2 (R Foundation for Statistical Computing) and a P-value of .05 was considered significant.

| Medical history: autoimmune and other conditions
AIH cases were more likely to have any another autoimmune disease in addition to AIH compared to controls (22.3% vs 10.1%, P < .001) ( Table 2). The most common autoimmune disease found in cases in addition to AIH was autoimmune thyroid disease (17.4%), followed by rheumatoid arthritis (RA) (11.8%), Sjogren's syndrome (5.6%), Raynaud's syndrome (5.6%) and celiac disease (4.1%). Cases were more likely to have any autoimmune disease assessed compared to controls except for autoimmune thyroid disease, Raynaud's syndrome and type 1 diabetes mellitus (T1DM).
Frequencies of other assessed medical conditions did not differ the two groups (Table 2).

| Gynaecologic and reproductive history
Reproductive history was completed by 284 AIH cases and 414 controls (Table 3). Age at study completion was lower in cases compared to controls (55 yrs (44, 63) vs 61.2 yrs (54.3, 67.4)). There was no difference in the age of first menarche between the study groups, nor when age when menstruation stopped. AIH cases were similar to cases in regards to ever being pregnant and total live births, but had less lifetime pregnancies (1 (1, 2) vs 3 (3,4), P < .001) compared to controls (Table 3).
AIH cases were more likely to have ever used oral contraceptives (OC) (83% vs 73.7%, P = .006) compared to controls, but the duration of OC use between groups was similar ( Table 3). Among those that used OC, the age of OC start was significantly younger in AIH cases compared to controls (19 yrs (18, 22) vs 21 yrs (21, 25), P < .001). Among cases with a history of OC use, 93.4% were exposed to OC prior to AIH diagnosis. Ever use of hormone replacement therapy (HRT) was lower in cases compared to controls (28.5% vs 60.1%, P < .001). Among those that used HRT, AIH cases tended to be younger at HRT start compared to controls (45.5 yrs (37.2, 50) vs 49 yrs (44, 52), P = .001). Among cases with a history of HRT use, and 91.1% were exposed to HRT prior to AIH diagnosis.

| Health and lifestyle exposures
AIH cases were not any more likely to have every regularly smoked cigarette as controls (38.4% vs 41.2%, P = .888) ( Table 4). Among those that had smoked, the age started smoking and years smoked regularly were no different between cases and controls. AIH cases were less likely to have been exposed to second-hand smoke at work (35.2% vs 51.6%, P < .001) but not at home (62.1% vs 69.2%, P = .748). However, AIH cases were more likely than controls to be current regular smokers (18.9% vs 7.4%, P = .022). AIH cases were less likely than controls to be current drinkers (

| Family history
Detailed history of autoimmune diseases, gastrointestinal and liver conditions among both first-degree and second-degree relatives (FDRs and SDRs) was collected in cases and in controls (

| Environmental risk discrepancy according to gender
Prior disparities among genders in AIH studies have revealed that male patients have higher frequencies of DRB1 * 03:01, worse overall survival and increased development of hepatocellular cancer. [13][14][15] Thus, stratification of reported medical histories and exposures were completed according to gender. Observed differences between medical histories in cases and controls were most often associated with female gender (Table S3). Smoking has not been assessed previously in AIH patients, yet it has been well documented in other autoimmune liver diseases PSC and PBC. In PSC, smoking most recently has been shown to only associate with disease among patients that also had concurrent IBD. 5 On the other hand, PBC has routinely been associated with smoking. 6,[22][23][24] In our study, we observed no difference in AIH cases compared to controls among frequency of "ever regularly smoked", age of smoking start and years of regularly smoked. However, we did observe that AIH cases were more likely to currently smoke (18.9%  These observed risks may highlight fundamental disease differences in pathogenesis between males and females including other genderspecific risks and interactions (increased HLA-DRB1 * 03:01 in males 13,14 and hormonal therapy or pregnancy in females).

| D ISCUSS I ON
We admit that given the observational approach to our study, there are limitations worth considering. Despite being one of the first systematic approaches to examine environmental risk in wellphenotyped AIH case development, the size of our study remains moderate. We were dependent on patient-reported data for many outcomes of interest which may introduce recall bias. Furthermore, this case-control study does not include a disease control group for direct comparison, yet two other liver disease populations (PBC and PSC) have utilized this questionnaire previously. 5,6 We were careful to compare our findings to these prior studies, yet given the inherent disease differences, we must consider our observations in the context of AIH. This methodology is necessary until well-cultivated prospective studies of AIH are established and completed. We believe our cohort is an excellent representation of AIH patients and well poised to address these questions, as other key findings support prior observations of similar demographics, concurrent autoimmune disease and enrichment of FDR (and few SDR) with autoimmune diseases (Table S2).
In summary, we report novel associations between a variety of exogenous risk factors and AIH using an established AIH cohort (GRACE study) and healthy controls. History of UTI, recurrent UTI, ever and earlier start of OC, current smoking and a variety of immunizations were associated with the development of AIH. On the other hand, history of mumps, HRT use, history of working with a smoker were protective against AIH. Despite plausible biologic mechanisms, these associations do not indicate disease causation.
Yet, counselling on smoking cessation, screening for other autoimmune illnesses at diagnosis and during follow-up and careful antibiotic stewardship remain important and are currently actionable by clinicians. We advocate for multicentre prospective studies of AIH to better understand modifiable risk factors in the setting of a permissive genetic framework.

E TH I C S A PPROVA L S TATEM ENT
This study included cohorts that adhere to the ethical guidelines of the 1975 Declaration of Helsinki, and was approved by the Institutional Review Board of Indiana University and Mayo Clinic.

CO N FLI C T O F I NTE R E S T
All included authors declare no outside interests that are directly or significantly related to this paper. Conception of study and design, collection and assembly of data, drafting and approval of final manuscript version.

PATI ENT CO N S ENT S TATEM ENT
All participants in the study had informed consent collected by study staff.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.