Advertisement

Effects of Coffee Consumption, Smoking, and Hormones on Risk for Primary Sclerosing Cholangitis

  • Ina Marie Andersen
    Affiliations
    Norwegian PSC Research Center, Department of Transplantation Medicine, Division of Cancer Medicine, Surgery and Transplantation, Oslo University Hospital, Rikshospitalet, Oslo
    Search for articles by this author
  • Guro Tengesdal
    Affiliations
    Norwegian PSC Research Center, Department of Transplantation Medicine, Division of Cancer Medicine, Surgery and Transplantation, Oslo University Hospital, Rikshospitalet, Oslo
    Search for articles by this author
  • Benedicte Alexandra Lie
    Affiliations
    Department of Immunology, Oslo University Hospital, Rikshospitalet, Oslo

    Institute of Clinical Medicine, Faculty of Medicine, University of Oslo, Oslo
    Search for articles by this author
  • Kirsten Muri Boberg
    Affiliations
    Norwegian PSC Research Center, Department of Transplantation Medicine, Division of Cancer Medicine, Surgery and Transplantation, Oslo University Hospital, Rikshospitalet, Oslo

    Institute of Clinical Medicine, Faculty of Medicine, University of Oslo, Oslo

    Section for Gastroenterology, Department of Transplantation Medicine, Division of Cancer Medicine, Surgery and Transplantation, Oslo University Hospital, Rikshospitalet, Oslo
    Search for articles by this author
  • Tom Hemming Karlsen
    Affiliations
    Norwegian PSC Research Center, Department of Transplantation Medicine, Division of Cancer Medicine, Surgery and Transplantation, Oslo University Hospital, Rikshospitalet, Oslo

    Section for Gastroenterology, Department of Transplantation Medicine, Division of Cancer Medicine, Surgery and Transplantation, Oslo University Hospital, Rikshospitalet, Oslo

    K. G. Jebsen Inflammation Research Centre, Research Institute of Internal Medicine, Division of Cancer Medicine, Surgery and Transplantation, Oslo University Hospital, Rikshospitalet, Oslo

    Division of Gastroenterology, Department of Clinical Medicine, University of Bergen, Bergen, Norway
    Search for articles by this author
  • Johannes Roksund Hov
    Correspondence
    Reprint requests Address requests for reprints to: Johannes Roksund Hov, MD, PhD, Norwegian PSC Research Center, Department of Transplantation Medicine, Division of Cancer Medicine, Surgery and Transplantation, Oslo University Hospital, Rikshospitalet, Pb 4950 Nydalen, N-0424 Oslo, Norway. fax: +47 23 07 39 28.
    Affiliations
    Norwegian PSC Research Center, Department of Transplantation Medicine, Division of Cancer Medicine, Surgery and Transplantation, Oslo University Hospital, Rikshospitalet, Oslo

    Institute of Clinical Medicine, Faculty of Medicine, University of Oslo, Oslo

    Section for Gastroenterology, Department of Transplantation Medicine, Division of Cancer Medicine, Surgery and Transplantation, Oslo University Hospital, Rikshospitalet, Oslo

    K. G. Jebsen Inflammation Research Centre, Research Institute of Internal Medicine, Division of Cancer Medicine, Surgery and Transplantation, Oslo University Hospital, Rikshospitalet, Oslo
    Search for articles by this author
Published:September 27, 2013DOI:https://doi.org/10.1016/j.cgh.2013.09.024

      Background & Aims

      Little is known about nongenetic risk factors for primary sclerosing cholangitis (PSC), except a possible protective effect of smoking. We investigated the relationship between environmental risk factors and susceptibility to PSC.

      Methods

      A questionnaire was distributed to patients with PSC, recruited from Oslo University Hospital Rikshospitalet in Norway through 2011, and randomly chosen individuals from the Norwegian Bone Marrow Donor Registry (control subjects). Data were analyzed from 240 patients with PSC and 245 control subjects, matched for gender and age.

      Results

      A lower proportion of patients with PSC were daily coffee drinkers than control subjects, both currently (76% vs 86%; odds ratio [OR], 0.52; 95% confidence interval [CI], 0.32–0.82; P = .006) and at the age of 18 years (35% vs 49%; OR, 0.58; 95% CI, 0.40–0.83; P = .003). The associations were mainly attributed to differences observed in men. Twenty percent of the patients were ever (current or former) daily smokers compared with 43% of control subjects (OR, 0.33; 95% CI, 0.22–0.50; P < .001). Ever daily smoking before PSC diagnosis was associated with older age at diagnosis (42 years vs 32 years; P < .001). Ever daily smoking (P < .001) and being a coffee drinker at the age of 18 years (P = .048) were independently and negatively associated with PSC. Fewer female patients with PSC than control subjects reported ever use of hormonal contraception (51% vs 85%; P < .001). Among female patients, there was a strong correlation between increasing number of children before the diagnosis of PSC and increasing age at diagnosis (r = 0.63; P < .001).

      Conclusions

      Coffee consumption and smoking might protect against development of PSC. In women, the disease might be influenced by hormonal factors.

      Keywords

      Abbreviations used in this paper:

      CI (confidence interval), IBD (inflammatory bowel disease), OR (odds ratio), PBC (primary biliary cirrhosis), PSC (primary sclerosing cholangitis), UC (ulcerative colitis)
      Primary sclerosing cholangitis (PSC) is a chronic inflammatory disease of the bile ducts predominately affecting men (2:1 male/female ratio).
      • Karlsen T.H.
      • Schrumpf E.
      • Boberg K.M.
      Update on primary sclerosing cholangitis.
      There is currently no effective medical treatment, and PSC is a common indication for liver transplantation. A close relationship with inflammatory bowel disease (IBD), most often ulcerative colitis (UC), and an increased frequency of typical autoimmune diseases are seen in PSC.
      • Saarinen S.
      • Olerup O.
      • Broome U.
      Increased frequency of autoimmune diseases in patients with primary sclerosing cholangitis.
      • Lamberts L.E.
      • Janse M.
      • Haagsma E.B.
      • et al.
      Immune-mediated diseases in primary sclerosing cholangitis.
      Up to 90% of patients with PSC of Northern European origin have coexistent IBD,
      • Saich R.
      • Chapman R.
      Primary sclerosing cholangitis, autoimmune hepatitis and overlap syndromes in inflammatory bowel disease.
      whereas lower frequencies have been reported in single studies from Southern Europe (54%) and Asia (21%).
      • Okolicsanyi L.
      • Fabris L.
      • Viaggi S.
      • et al.
      Primary sclerosing cholangitis: clinical presentation, natural history and prognostic variables: an Italian multicentre study. The Italian PSC Study Group.
      • Takikawa H.
      • Takamori Y.
      • Tanaka A.
      • et al.
      Analysis of 388 cases of primary sclerosing cholangitis in Japan. Presence of a subgroup without pancreatic involvement in older patients.
      The incidence of PSC is highest in populations of Northern European descent, with incidence reports from United States, Wales, Great Britain, and Norway of 0.9–1.3 per 100,000 and prevalence of 8.5–14.2 per 100,000,
      • Boberg K.M.
      • Aadland E.
      • Jahnsen J.
      • et al.
      Incidence and prevalence of primary biliary cirrhosis, primary sclerosing cholangitis, and autoimmune hepatitis in a Norwegian population.
      • Kingham J.G.
      • Kochar N.
      • Gravenor M.B.
      Incidence, clinical patterns, and outcomes of primary sclerosing cholangitis in South Wales, United Kingdom.
      • Bambha K.
      • Kim W.R.
      • Talwalkar J.
      • et al.
      Incidence, clinical spectrum, and outcomes of primary sclerosing cholangitis in a United States community.
      • Kaplan G.G.
      • Laupland K.B.
      • Butzner D.
      • et al.
      The burden of large and small duct primary sclerosing cholangitis in adults and children: a population-based analysis.
      whereas the incidence rates are likely to be much lower in Asia, Southern Europe, and Alaska.
      • Takikawa H.
      • Manabe T.
      Primary sclerosing cholangitis in Japan: analysis of 192 cases.
      • Escorsell A.
      • Pares A.
      • Rodes J.
      • et al.
      Epidemiology of primary sclerosing cholangitis in Spain. Spanish Association for the Study of the Liver.
      • Hurlburt K.J.
      • McMahon B.J.
      • Deubner H.
      • et al.
      Prevalence of autoimmune liver disease in Alaska Natives.
      The causes of PSC and the variation of incidence and comorbidities are currently not known. A 9- to 39-fold increased risk of PSC among first-degree relatives strongly suggests a genetic contribution.
      • Bergquist A.
      • Montgomery S.M.
      • Bahmanyar S.
      • et al.
      Increased risk of primary sclerosing cholangitis and ulcerative colitis in first-degree relatives of patients with primary sclerosing cholangitis.
      Recent studies have identified multiple robust genetic risk factors in PSC.
      • Folseraas T.
      • Melum E.
      • Rausch P.
      • et al.
      Extended analysis of a genome-wide association study in primary sclerosing cholangitis detects multiple novel risk loci.
      • Liu J.Z.
      • Hov J.R.
      • Folseraas T.
      • et al.
      Dense genotyping of immune-related disease regions identifies nine new risk loci for primary sclerosing cholangitis.
      • Melum E.
      • Franke A.
      • Schramm C.
      • et al.
      Genome-wide association analysis in primary sclerosing cholangitis identifies two non-HLA susceptibility loci.
      Half of these seem to overlap with susceptibility loci in typical autoimmune diseases (eg, type 1 diabetes), whereas the other half overlap with genetic risk factors in IBD,
      • Liu J.Z.
      • Hov J.R.
      • Folseraas T.
      • et al.
      Dense genotyping of immune-related disease regions identifies nine new risk loci for primary sclerosing cholangitis.
      strongly suggesting that PSC is an immune-mediated disease, linked to but different from IBD without PSC. Of interest are the observations of some genetic heterogeneity between PSC in Northern and Southern Europe,
      • Liu J.Z.
      • Hov J.R.
      • Folseraas T.
      • et al.
      Dense genotyping of immune-related disease regions identifies nine new risk loci for primary sclerosing cholangitis.
      • Melum E.
      • Franke A.
      • Schramm C.
      • et al.
      Genome-wide association analysis in primary sclerosing cholangitis identifies two non-HLA susceptibility loci.
      • Hov J.R.
      • Lleo A.
      • Selmi C.
      • et al.
      Genetic associations in Italian primary sclerosing cholangitis: heterogeneity across Europe defines a critical role for HLA-C.
      which could contribute to differences in the incidence of PSC and of concurrent IBD. However, environmental and lifestyle factors are also highly likely to influence the development and hence the epidemiology of PSC.
      Few studies have investigated potential environmental risk factors in PSC. Three reports have suggested a protective effect of smoking
      • Mitchell S.A.
      • Thyssen M.
      • Orchard T.R.
      • et al.
      Cigarette smoking, appendectomy, and tonsillectomy as risk factors for the development of primary sclerosing cholangitis: a case control study.
      • Loftus Jr., E.V.
      • Sandborn W.J.
      • Tremaine W.J.
      • et al.
      Primary sclerosing cholangitis is associated with nonsmoking: a case-control study.
      • van Erpecum K.J.
      • Smits S.J.
      • van de Meeberg P.C.
      • et al.
      Risk of primary sclerosing cholangitis is associated with nonsmoking behavior.
      and one study found tonsillectomy to be protective.
      • Mitchell S.A.
      • Thyssen M.
      • Orchard T.R.
      • et al.
      Cigarette smoking, appendectomy, and tonsillectomy as risk factors for the development of primary sclerosing cholangitis: a case control study.
      Appendectomy exerts a protective effect in UC,
      • Rutgeerts P.
      • D'Haens G.
      • Hiele M.
      • et al.
      Appendectomy protects against ulcerative colitis.
      • Russel M.G.
      • Dorant E.
      • Brummer R.J.
      • et al.
      Appendectomy and the risk of developing ulcerative colitis or Crohn's disease: results of a large case-control study. South Limburg Inflammatory Bowel Disease Study Group.
      but a similar protective effect has not been seen in PSC.
      • Mitchell S.A.
      • Thyssen M.
      • Orchard T.R.
      • et al.
      Cigarette smoking, appendectomy, and tonsillectomy as risk factors for the development of primary sclerosing cholangitis: a case control study.
      • van Erpecum K.J.
      • Smits S.J.
      • van de Meeberg P.C.
      • et al.
      Risk of primary sclerosing cholangitis is associated with nonsmoking behavior.
      None of these findings can be specifically linked to geographic differences; however, in other inflammatory diseases (eg, multiple sclerosis) a strong North-South gradient in prevalence has been attributed to a protective effect of UV exposure and high levels of 25-hydroxyvitamin D.
      • Munger K.L.
      • Levin L.I.
      • Hollis B.W.
      • et al.
      Serum 25-hydroxyvitamin D levels and risk of multiple sclerosis.
      Increasing prevalence of IBD in previous low-incidence areas, such as Asia and Northern Africa, suggests that westernization of lifestyle and industrialization may be other environmental factors affecting disease susceptibility.
      • Ng S.C.
      • Bernstein C.N.
      • Vatn M.H.
      • et al.
      Geographical variability and environmental risk factors in inflammatory bowel disease.
      Identification of such nongenetic risk factors can give important information regarding the pathogenesis of a disease and possibly guide future research and therapeutic or preventive interventions. In autoimmune diseases, examples include the increased risk of rheumatoid arthritis in smokers
      • Sugiyama D.
      • Nishimura K.
      • Tamaki K.
      • et al.
      Impact of smoking as a risk factor for developing rheumatoid arthritis: a meta-analysis of observational studies.
      and urinary tract infections and nail polish use in primary biliary cirrhosis (PBC),
      • Gershwin M.E.
      • Selmi C.
      • Worman H.J.
      • et al.
      Risk factors and comorbidities in primary biliary cirrhosis: a controlled interview-based study of 1032 patients.
      which are preventable exposures that may possibly be linked to pathogenetic mechanisms. For common diseases, large cohort studies may be valuable to detect risk factors, whereas in rare liver diseases, such as PBC or PSC, case-control studies remain the only option.
      • Gershwin M.E.
      • Selmi C.
      • Worman H.J.
      • et al.
      Risk factors and comorbidities in primary biliary cirrhosis: a controlled interview-based study of 1032 patients.
      In PSC, there is an urgent need to identify new environmental factors that could shed light on the pathogenesis and that could point to preventive measures (eg, in a setting of IBD). Hence, we aimed to investigate the relationship between selected nongenetic risk factors and disease susceptibility in PSC by means of a questionnaire.

      Materials and Methods

       Study Cohort

      In 2011, all alive patients in the Norwegian PSC patient registry (n = 336) at Oslo University Hospital Rikshospitalet, a tertiary referral center, were invited to participate in the study. All patients had been recruited at hospital admission over the last 20 years. After 1 reminder 245 patients answered the questionnaire (response rate, 73%). Before the final analysis, only patients with typical PSC according to accepted criteria
      • Chapman R.W.
      • Arborgh B.A.
      • Rhodes J.M.
      • et al.
      Primary sclerosing cholangitis: a review of its clinical features, cholangiography, and hepatic histology.
      with cholangiographic large duct affection and no known IgG4 disease were included (n = 240). The questionnaire was also distributed to 1096 randomly chosen individuals in the Norwegian Bone Marrow Donor registry (response rate, 85% after 2 reminders), who 6 years before this study gave consent to participate in health research. The bone marrow donor registry is a nationwide registry of blood donors who have consented to serve as a screening source for stem cell donors. At original inclusion these individuals were screened for major disease and drug use according to blood donor regulations, but for research purposes there has been no additional screening. Out of the responding control subjects (control pool), 245 individuals were selected as control subjects using the Fuzzy command in the Python extension in SPSS (IBM, Chicago, IL), requiring absolute gender match. To get a best possible age matching despite a wider age distribution in the patient group than the control pool, the selection of control subjects was performed in 3 stages, accepting an increasing age mismatch; with ±5 years age mismatch, 77% of the patients were assigned a control individual; with ±10 years, an additional 15% were matched, whereas the remaining patients were assigned a control individual without age constraints. All participants gave written informed consent. The study conforms to the ethical guidelines of the 1975 Declaration of Helsinki and was approved by the Regional Committee for Medical and Health Research Ethics South East.

       Questionnaire

      The study participants received a questionnaire consisting of 41 questions, mostly selected or modified from questionnaires from multiple previous questionnaire studies.
      • Hedstrom A.K.
      • Baarnhielm M.
      • Olsson T.
      • et al.
      Tobacco smoking, but not Swedish snuff use, increases the risk of multiple sclerosis.
      • Lund E.
      • Dumeaux V.
      • Braaten T.
      • et al.
      Cohort profile: The Norwegian Women and Cancer Study–NOWAC–Kvinner og kreft.
      • Maniaol A.H.
      • Brunborg C.
      • Tallaksen C.M.
      Development and validation of a self-administered questionnaire for myasthenia gravis patients.
      • Krokstad S.
      • Langhammer A.
      • Hveem K.
      • et al.
      Cohort profile: The HUNT Study, Norway.
      • Pugliatti M.
      • Casetta I.
      • Drulovic J.
      • et al.
      A questionnaire for multinational case-control studies of environmental risk factors in multiple sclerosis (EnvIMS-Q).
      See Supplementary Methods for translation of the questionnaire and the source and modifications of the individual items. The items included lifestyle factors, somatic comorbidity, workplace exposure, and dietary habits. Females were also asked about menstrual status, number of children and duration of breast-feeding, contraceptive use (summary question without details on progesterone only or combinations of estrogen/progesterone), abortions, and infertility treatment. Coffee and tea use was recorded as number of cups per day, zero if no use. Smoking was recorded by multiple questions, based on which the patients were categorized as never, former, or current daily smokers, which was subsequently aggregated to ever or never daily smokers. The reported age at start of smoking was used to define smoking status at the time of PSC diagnosis.

       Statistics

      The unadjusted analyses of differences between cases and control subjects for potential risk factors were performed using Student t test or Mann-Whitney U test for continuous variables and the chi-square test for categorical variables. In multivariate analyses, logistic regressions were used, which included dichotomous categorical variables (coded 0 or 1) or continuous variables (age). All analyses were two-sided. P < .05 was considered statistically significant. Continuous variables were expressed as mean ± standard deviation or median (interquartile range). The date of PSC diagnosis was defined as first-ever diagnostic cholangiogram (ie, where relevant, examinations from the local hospital were obtained and reviewed). Patient-reported cases of osteoporosis and IBD were verified in patient records before inclusion in analyses. Agreement between questionnaires and patient records was assessed with Cohen kappa. Body mass index was calculated by using the formula mass (kg)/square of height(m2). Month-specific risk of PSC was calculated by comparing each month with the other 11 months. Only females were included in analyses of reproductive health and nail polish use. All statistical analyses were performed using SPSS (IBM) version 20.

      Results

      An overview of the study cohort is presented in Table 1. Seventy-one percent of the included patients and 70% of the control subjects were men. Median age of the patients with PSC was 48 years compared with 46 years in the control subjects (P = .085). The geographic distribution of the study population was nationwide and similar in patients and control subjects, as defined by current postal code. The nonresponding (ie, not included) patients were younger than the responding (40 vs 48 years; P < .001); there was no significant gender difference. Median age at diagnosis of PSC was 34 years (range, 13–73 years). The first available liver biochemistry after diagnosis is shown in Supplementary Table 1. Ninety-five (39%) of the patients were liver transplanted at the time of study. More than 97% of the study population was of Nordic descent. The distribution of birth month or season was similar in patients and control subjects. Forty-eight percent of the patients reported having finished college or university-level education compared with 65% of control subjects (P < .001).
      Table 1Demographic Characteristics of Patients and Control Subjects
      Control subjects (n = 245)Patients (n = 240)Unadjusted P value
      Male gender, n (%)172 (70)171 (71)NS
      Age all (y), median (IQR)46 (40–53)48 (38–57)NS (.085)
       Men46 (39–52)47 (38–56)NS
       Women47 (42–54)49 (39–60)NS
      College educated, n (%)156 (65)114 (48)<.001
      Nordic country of birth for mother, n (%)236 (97)227 (97)NS
      Nordic country of birth for father, n (%)237 (97)228 (97)NS
      Height all (cm), median (IQR)179 (172–184)179 (171–184)NS
      Weight all (kg), median (IQR)84 (75–95)77 (69–86)<.001
      Weight all at age 18 (kg), median (IQR)73 (65–80)70 (63–77)NS
      Current BMI all, median (IQR)26 (24–29)24 (22–27)<.001
      BMI all at 18 years, median (IQR)22 (20–24)21 (18–24).008
      Age at diagnosis of PSC (y), median (IQR)NA34 (26–45)
      Liver transplanted, n (%)NA94 (39)
      Self-reported IBD (all types), n (%)3 (1)171 (73)<.001
      Register verified IBD (all types), n (%)NA183 (77)
       UCNA150 (64)
       Crohn diseaseNA23 (10)
       Indeterminate colitisNA10 (4)
      Liver-related neoplasia, n (%)NA8 (4)
       Cholangiocarcinoma in situNA5 (2)
       CholangiocarcinomaNA3 (1)
       Hepatocellular carcinomaNA1 (0.4)
      IQR, interquartile range; NS, not statistically significant (P > .05); SD, standard deviation.

       Coffee Consumption

      Patients with PSC reported less coffee consumption than control subjects both currently (mean, 2.7 vs 4.0 cups; P < .001) and at the age of 18 (mean, 1.0 vs 1.6 cups; P = .001) (Table 2). The proportion of patients with PSC drinking coffee daily at all was also lower than control subjects currently (76% vs 86%; P = .006) and at the age of 18 (35% vs 49%; P = .004). When performing subgroup analyses according to gender, there were only trends toward lower coffee consumption in female patients with PSC than in control subjects when measured as current number of cups per day (Table 2). In contrast, the coffee consumption in male patients with PSC was significantly lower for all measures (Table 2). Twenty patients with PSC were diagnosed before the age of 18 (8%), with no significant difference in reported coffee consumption at age 18 from patients with PSC diagnosed after the age of 18. Regarding the subgroups of PSC defined by a history of IBD, there was a trend toward lower coffee consumption in patients with than without IBD for rate of current daily coffee drinkers (73% vs 85%; P = .077), whereas at the age of 18, the difference was not significant (34% in PSC with IBD vs 39% in PSC without IBD). In a subset of patients (n = 126) with clinical chemistry available from the time of diagnosis, any coffee consumption at the age of 18 was associated with lower levels of alanine aminotransferase (114 U/L in coffee drinkers vs 177 U/L in nondrinkers; P = .029), alkaline phosphatases (340 vs 627 U/L; P = .013), and γ-glutamyltranspeptidases (249 vs 439; P = .011).
      Table 2Coffee, Tea, Smoking, and Alcohol Consumption of Patients and Control Subjects
      Control subjects

      (n = 245)
      Patients

      (n = 240)
      Unadjusted

      P value
      Current daily coffee consumption at all, n (%)207 (86)177 (76).006
       Men151 (89)126 (75).001
       Women56 (78)51 (76)NS
      Daily coffee consumption at all at age 18, n (%)117 (49)82 (35).004
       Men90 (53)58 (35).001
       Women27 (38)24 (38)NS
      Daily consumption of coffee (cups), mean ± SD4.0 ± 2.92.7 ± 2.6<.001
       Men4.4 ± 2.92.9 ± 2.8<.001
       Women3.0 ± 2.72.2 ± 2.0NS
      Daily consumption of coffee (cups) at 18, mean ± SD1.6 ± 2.41.0 ± 1.9.001
       Men1.8 ± 2.51.0 ± 2.0<.001
       Women1.2 ± 1.90.9 ± 1.5NS
      Current daily smoking, n (%)37 (15)5 (2)<.001
      Former daily smoking, n (%)69 (28)43 (18).009
      Ever daily smoking, n (%)106 (43)48 (20)<.001
      Never daily smoking, n (%)139 (57)190 (80)<.001
      Exposition to passive smoking, n (%)173 (72)148 (63).040
      Ever use of snuff, n (%)51 (21)30 (13).014
      Current daily tea consumption at all, n (%)82 (42)102 (50)NS
      Daily tea consumption at all at age 18, n (%)54 (27)56 (28)NS
      Daily consumption of tea (cups), mean ± SD0.8 ± 1.40.9 ± 1.2NS
      Daily consumption of tea (cups) at 18, mean ± SD0.6 ± 1.20.4 ± 0.8NS
      No alcohol consumption, n (%)23 (9)91 (38)<.001
      0–1 units of alcohol daily, n (%)162 (67)124 (52).001
      More than 1 unit of alcohol daily, n (%)58 (24)24 (10)<.001
      No alcohol consumption at 18, n (%)54 (22)75 (31).022
      0–1 units of alcohol daily at 18, n (%)118 (49)103 (43)NS
      More than 1 unit of alcohol daily at 18, n (%)71 (29)59 (25)NS
      NS, not statistically significant (P > .05); SD, standard deviation.

       Smoking

      Five patients with PSC (2%) were current daily smokers compared with 37 (15%) of the control subjects (P < .001; Table 2). Ever (current or former) daily smoking was reported by 20% of the patients and 43% of the control subjects (odds ratio [OR], 0.33; 95% confidence interval [CI], 0.22–0.50; P < .001). There was no difference in the prevalence of ever daily smoking between patients with PSC with and without a diagnosis of IBD (19% vs 22%, respectively; P = .64). There was also no difference between different IBD subtypes; 22% of patients with PSC diagnosed with Crohn disease reported being ever daily smokers compared with 20% of patients with UC. The patients with PSC started smoking later than the control subjects (median age, 18 vs 17 years; P = .007). Most smokers, both patients with PSC and control subjects, started smoking early (<25 years of age), several years before median onset of PSC (Figure 1). Using logistic regression adjusting for current age, gender, and education level, both ever daily smoking (P < .001) and being a coffee drinker at the age of 18 (P = .048) were independently negatively associated with reduced risk of PSC. There were also trends toward less exposure to passive smoke (63% vs 72%; P = .040) and ever daily use of snuff (13% vs 21%; P = .014) in patients with PSC than control subjects, but these exposures were not independent from ever daily smoking. Ever daily smoking before the diagnosis of PSC was associated with older age at diagnosis compared with never daily smoking (median 42 vs 32 years; P < .001).
      Figure thumbnail gr1
      Figure 1Initiation of daily smoking in patients with PSC and control subjects. The proportions of control subjects (black line) and patients with PSC (dashed line) being ever daily smokers are plotted according to age, showing that initiation of smoking in general is an event early in life (<25 years). Individuals are censored (shown as vertical dots) at the age of the present study. For comparison, the grey line shows the proportion of the patients with PSC diagnosed with PSC at a given age, showing a median age at diagnosis of 34 years, making it unlikely that the no smoking status of patients with PSC is a consequence of acquiring PSC.

       Other Exposures

      Patients with PSC reported lower alcohol consumption both currently and at the age of 18 (Table 2). There were no differences between patients and control subjects in regard to reports of previous workplace exposure to oils, asphalt, anesthetic gas, or organic solvents, nor frequency of nail polish use estimated from use over the last year (data not shown). The frequency of domestic animals in childhood was similar in control subjects and patients (76%), although the patients with PSC reported a higher frequency of exposure to farm animals (horse, 13% vs 4%, P = .001; cows, 10% vs 5%, P = .052).

       Questionnaire-reported Comorbidities

      At least one immune-mediated disease (excluding IBD) was reported by 25% of patients and 11% of control subjects (P < .001). The most commonly reported immune-mediated comorbidities in PSC were psoriasis, celiac disease, hypothyroidism, and sarcoidosis (Table 3). Appendectomy and tonsillectomy rates were similar in patients and control subjects (Table 3). Regarding other diseases, 8% of the patients reported having been hospitalized for an abdominal infection before the age of 18 compared with 2% of the control subjects (Table 3), but after exclusion of patients with IBD the difference was not significant. Adherence to the children’s vaccine program was near complete (Table 3). The most notable other non–immune-mediated disease among patients was osteoporosis, which was reported at a high prevalence both in transplanted (25%) and nontransplanted (11%) patients with PSC, compared with 0.4% of the control subjects. Type 2 diabetes was more prevalent among transplanted patients with PSC than nontransplanted and control subjects (14% transplanted vs 4% nontransplanted [P = .005] and 2% among control subjects). Considering the validity of the patient-reported diseases, 171 (73%) patients reported having an IBD diagnosis. After verification using patient charts and the PSC registry, this increased to 183 (77%), corresponding to a kappa of agreement between the patient-reported and hospital chart diagnoses of 0.92, commonly interpreted as “almost perfect agreement.”
      • Landis J.R.
      • Koch G.G.
      The measurement of observer agreement for categorical data.
      For osteoporosis, there was a perfect overlap between patient reports and hospital charts.
      Table 3Current and/or Previous Comorbidities Reported by Patients and Control Subjects
      Control subjects (n = 245)Patients (n = 240)Unadjusted P value
      Type 1 diabetes, n (%)0 (0)6 (3).012
      Psoriasis, n (%)14 (6)14 (6)NS
      Rheumatoid arthritis, n (%)3 (1)5 (2)NS
      Ankylosing spondylitis, n (%)2 (1)5 (2)NS
      Sarcoidosis, n (%)0 (0)8 (3).004
      Myasthenia gravis, n (%)0 (0)0 (0)
      Multiple sclerosis, n (%)0 (0)1 (0.4)NS
      Celiac disease, n (%)0 (0)12 (5)<.001
      Systemic lupus erythematosus, n (%)0 (0)0 (0)
      Sjögren syndrome, n (%)0 (0)2 (1)NS
      Hypothyroidism, n (%)7 (3)10 (4)NS
      Hyperthyroidism, n (%)2 (1)6 (3)NS
      1 autoimmune except IBD, n (%)27 (11)59 (25)<.001
      1 autoimmune including self-reported IBD, n (%)30 (12)187 (78)<.001
      1 autoimmune including verified IBD, n (%)30 (12)197 (82)<.001
      Myocardial infarction, n (%)3 (1)2 (1)NS
      Angina pectoris, n (%)2 (1)4 (2)NS
      Heart failure, n (%)0 (0)0 (0)NS
      Other heart disease, n (%)4 (2)9 (4)NS
      Stroke, n (%)5 (2)2 (1)NS
      Type 2 diabetes, n (%)4 (2)19 (8).001
      Kidney disease, n (%)4 (2)18 (8).001
      Asthma, n (%)14 (6)21 (9)NS
      Chronic obstructive pulmonary disease, n (%)3 (1)2 (1)NS
      Osteoporosis, n (%)1 (0.4)38 (16)<.001
      Osteoarthritis, n (%)12 (5)9 (4)NS
      Migraine, n (%)39 (16)27 (12)NS
      Mononucleosis, n (%)33 (14)37 (17)NS
      Tonsillectomy, n (%)55 (23)55 (23)NS
      Appendectomy, n (%)31 (13)39 (17)NS
      Hospitalized for infection <18 years, n (%)27 (11)29 (13)NS
       Pneumonia8 (3)7 (3)NS
       Kidney/urinary tract4 (2)1 (0.4)NS
       Abdominal4 (2)18 (8).002
       Meningeal/brain1 (0.4)2 (1)NS
       Other15 (6)7 (3)NS
      Root infection, n (%)54 (22)45 (19)NS
      Gingival infection, n (%)13 (5)23 (10)NS (.074)
      None or incomplete vaccination, n (%)2 (1)1 (0.4)NS
      IQR, interquartile range; NS, not statistically significant (P > .05).

       Female Reproductive Health

      In women, both age at menarche and number of children were similar in patients with PSC and control subjects (Table 4). Fewer patients with PSC than control subjects reported ever use of hormonal contraception (51% vs 85%; P < .001) or nonhormonal contraception (27% vs 48%; P = .009), but there was no difference in duration of use for the women who had used contraception. In multiple logistic regressions in the females, adjusting for education level (P = .018) and ever daily smoking (P < .001), use of hormonal contraception was independently negatively associated with PSC (P = .001). The questionnaire did not allow for distinguishing effects of different types of hormonal contraception (eg, progesterone alone or combinations with estrogen). Female patients with PSC with IBD had a higher rate of use of hormonal contraception than females with PSC without IBD, but the difference was not significant (53% vs 40%; P = .32). There was no difference between patients and control subjects in rate reporting having ever been pregnant or ever having had an abortion. There were no differences regarding age at birth or duration of breast-feeding for any child. However, we observed a strong linear relationship between increasing number of children before the onset of PSC and increasing age at PSC diagnosis (r = 0.63; P < .001; Figure 2).
      Table 4Female Reproductive Health Factors in Patients and Control Subjects
      Control subjects(n = 73)Patients (n = 69)Unadjusted P value
      Age at menarche, median (IQR)13 (12–14)14 (12–14)NS
      Ever use of hormonal contraception, n (%)61 (85)35 (51)<.001
      Years with hormonal contraception, mean ± SD9 ± 68 ± 6NS
      Ever use of nonhormonal contraception, n (%)34 (48)18 (27).009
      Years with nonhormonal contraception, mean ± SD7 ± 55 ± 4NS
      Previous abortion, n (%)22 (31)23 (33)NS
      Number of abortions, mean ± SD2.0 ± 1.01.3 ± 0.6.006
      Hormonal treatment for infertility, n (%)8 (11)1 (2).020
      Ever pregnant, n (%)58 (81)54 (78)NS
      Number of children, mean ± SD1.5 ± 1.21.8 ± 1.3NS
      IQR, interquartile range; NS, not statistically significant (P > .05); SD, standard deviation.
      Figure thumbnail gr2
      Figure 2Relationship between number of children and age at PSC diagnosis. The figure shows the age at diagnosis of PSC (n = 69 female patients) according to the number of children born before the diagnosis, showing a strong linear relationship between these variables (r = 0.63; P < .001).

      Discussion

      In this single-center study of nongenetic risk factors in PSC, a negative association with coffee consumption was observed, in particular in male patients, suggesting that coffee may protect against PSC development. The study also provides strong support for a protective effect of smoking against PSC development, independent from IBD and IBD subtype, and also leading to delayed onset of disease. Less contraceptives use and a strong correlation between number of children and increasing age at diagnosis suggest that hormonal factors may influence risk or onset of PSC.
      The patients with PSC reported lower rate of coffee consumption both currently and at the age of 18, measured as the proportion being coffee drinkers and mean number of cups daily. The negative association with coffee consumption was mainly attributable to the difference between male patients with PSC and control subjects, whereas the difference was not as evident in women. There are no previous data regarding the role of coffee in PSC. However, a protective effect against gallstone disease has been observed,
      • Leitzmann M.F.
      • Willett W.C.
      • Rimm E.B.
      • et al.
      A prospective study of coffee consumption and the risk of symptomatic gallstone disease in men.
      and increased bile flow and reduced gallbladder absorption are among the suggested mechanisms.
      • Lillemoe K.D.
      • Magnuson T.H.
      • High R.C.
      • et al.
      Caffeine prevents cholesterol gallstone formation.
      Coffee drinking has been associated with liver biochemistry levels in population studies,
      • Nilssen O.
      • Forde O.H.
      • Brenn T.
      The Tromso Study. Distribution and population determinants of gamma-glutamyltransferase.
      • Honjo S.
      • Kono S.
      • Coleman M.P.
      • et al.
      Coffee consumption and serum aminotransferases in middle-aged Japanese men.
      • Torres D.M.
      • Harrison S.A.
      Is it time to write a prescription for coffee? Coffee and liver disease.
      and in the present study lower levels of alanine aminotransferase, alkaline phosphatases, and γ-glutamyltranspeptidase significantly associate with coffee consumption. The literature also suggests a protective effect against hepatocellular carcinoma,
      • Bravi F.
      • Bosetti C.
      • Tavani A.
      • et al.
      Coffee drinking and hepatocellular carcinoma: an update.
      and reduced hepatic fibrosis and slower disease progression in patients with hepatitis C
      • Freedman N.D.
      • Everhart J.E.
      • Lindsay K.L.
      • et al.
      Coffee intake is associated with lower rates of liver disease progression in chronic hepatitis C.
      and nonalcoholic steatohepatitis.
      • Molloy J.W.
      • Calcagno C.J.
      • Williams C.D.
      • et al.
      Association of coffee and caffeine consumption with fatty liver disease, nonalcoholic steatohepatitis, and degree of hepatic fibrosis.
      This points to liver-specific health benefits of coffee drinking irrespective of disease etiology.
      • Masterton G.S.
      • Hayes P.C.
      Coffee and the liver: a potential treatment for liver disease?.
      In murine models, coffee has been shown to down-regulate both the profibrogenic cytokine transforming growth factor-β
      • Moreno M.G.
      • Chavez E.
      • Aldaba–Muruato L.R.
      • et al.
      Coffee prevents CCl(4)-induced liver cirrhosis in the rat.
      and cytokine gene expression,
      • Fukushima Y.
      • Kasuga M.
      • Nakao K.
      • et al.
      Effects of coffee on inflammatory cytokine gene expression in mice fed high-fat diets.
      implying that antifibrotic or anti-inflammatory mechanisms could be involved. The effects of coffee could also be more general as illustrated by the association between coffee consumption and reduction of all-cause mortality.
      • Freedman N.D.
      • Park Y.
      • Abnet C.C.
      • et al.
      Association of coffee drinking with total and cause-specific mortality.
      In the latter study there was an inverse association between coffee and most major causes of death in both genders, except for cancer deaths among women.
      • Freedman N.D.
      • Park Y.
      • Abnet C.C.
      • et al.
      Association of coffee drinking with total and cause-specific mortality.
      The mechanisms causing gender differences are not known, but gender-specific effects of coffee have previously been observed in Parkinson disease.
      • Ascherio A.
      • Weisskopf M.G.
      • O'Reilly E.J.
      • et al.
      Coffee consumption, gender, and Parkinson's disease mortality in the cancer prevention study II cohort: the modifying effects of estrogen.
      Participants in our study reported coffee consumption daily both currently and at the age of 18, the latter to quantify the consumption at a time point, which for most cases was before the diagnosis of PSC onset to reduce the problem of coffee intake being influenced by disease (ie, “reverse” causality). This time point was selected to minimize risk of recall bias, because reaching the legal age of 18, the last year of school, and acquiring a driver’s license are events people may be likely to remember. Importantly, the proportion of patients not at all drinking coffee, which could represent a more robust parameter than the number of cups, was also lower than in control subjects. Coffee intake between the age of 18 years and now was not investigated in our study. This may be of significance, because quantification of coffee consumption before the onset of disease would give more information than consumption at a single time point. Our questionnaire was also not designed to separate the effects of caffeinated versus decaffeinated coffee, or filtered versus unfiltered coffee, which is important to address in later studies because it may help decipher the effect of caffeine and other substances in coffee. There are very limited data in the literature suggesting that coffee consumption per se influences IBD risk.
      • Ng S.C.
      • Bernstein C.N.
      • Vatn M.H.
      • et al.
      Geographical variability and environmental risk factors in inflammatory bowel disease.
      • Persson P.G.
      • Ahlbom A.
      • Hellers G.
      Diet and inflammatory bowel disease: a case-control study.
      In the present study, patients with and without a history of IBD reported similar coffee consumption at the age of 18, whereas current coffee consumption was less prevalent in patients with IBD. The latter observation could be speculated to be caused by coffee, which has been reported to worsen IBD symptoms.
      • Cohen A.B.
      • Lee D.
      • Long M.D.
      • et al.
      Dietary patterns and self-reported associations of diet with symptoms of inflammatory bowel disease.
      Unfortunately, it was not possible to assess the relationship between coffee exposure and the onset or activity of IBD, and to definitely conclude that the effects of coffee on PSC are independent from IBD requires additional studies. Smoking and educational level are other possible confounders of coffee consumption
      • Berger J.
      • Wynder E.L.
      The correlation of epidemiological variables.
      but the observed effects were robust to adjustment for these factors. Taken together, the results are highly suggestive of a protective effect of coffee against PSC in men, but no conclusion can be drawn in women. Other studies are required to confirm these observations, preferably using other instruments to quantify coffee consumption.
      The previously observed protective effects of cigarette smoking are strongly reinforced in the present study. Of interest is that key substances in coffee and cigarette smoke (caffeine and nicotine, respectively) have similar effects; both are sympathomimetic and may increase the levels of intracellular cyclic adenosine monophosphate.
      • Smits P.
      • Temme L.
      • Thien T.
      The cardiovascular interaction between caffeine and nicotine in humans.
      The point estimate of the OR of PSC for ever daily smoking (0.35) is within the range observed in previous studies from the United States, the Netherlands, and the United Kingdom (0.33–0.54).
      • Mitchell S.A.
      • Thyssen M.
      • Orchard T.R.
      • et al.
      Cigarette smoking, appendectomy, and tonsillectomy as risk factors for the development of primary sclerosing cholangitis: a case control study.
      • Loftus Jr., E.V.
      • Sandborn W.J.
      • Tremaine W.J.
      • et al.
      Primary sclerosing cholangitis is associated with nonsmoking: a case-control study.
      • van Erpecum K.J.
      • Smits S.J.
      • van de Meeberg P.C.
      • et al.
      Risk of primary sclerosing cholangitis is associated with nonsmoking behavior.
      The negative association between smoking and PSC was observed in patients both with and without a history of IBD, in line with the findings in the United Kingdom,
      • Mitchell S.A.
      • Thyssen M.
      • Orchard T.R.
      • et al.
      Cigarette smoking, appendectomy, and tonsillectomy as risk factors for the development of primary sclerosing cholangitis: a case control study.
      pointing to specific protective effect of cigarette smoking against PSC. There was also no difference between IBD subtypes in PSC, as would have been expected in IBD without PSC,
      • Lakatos P.L.
      • Szamosi T.
      • Lakatos L.
      Smoking in inflammatory bowel diseases: good, bad or ugly?.
      perhaps supporting that PSC-IBD is a separate IBD subtype.
      • Loftus Jr., E.V.
      • Harewood G.C.
      • Loftus C.G.
      • et al.
      PSC-IBD: a unique form of inflammatory bowel disease associated with primary sclerosing cholangitis.
      Interestingly, ever daily smoking was associated with higher age at diagnosis of PSC, suggesting it may protect against PSC also by delaying the disease process. This is in line with a later age of onset of UC in smokers observed in some studies
      • Lakatos P.L.
      • Szamosi T.
      • Lakatos L.
      Smoking in inflammatory bowel diseases: good, bad or ugly?.
      and daily smoking thus represents the first exogenous factor shown to affect the age of PSC onset. Sociodemographic factors as measured by educational level are known to be associated with smoking behavior and represent possible confounders. However, given the higher educational level among the control subjects this actually suggests an even stronger protective effect of smoking.
      This study provides new information regarding female reproductive health and PSC. Pregnancy is known to affect the course of several immune-mediated diseases including autoimmune hepatitis
      • Heneghan M.A.
      • Norris S.M.
      • O'Grady J.G.
      • et al.
      Management and outcome of pregnancy in autoimmune hepatitis.
      and PBC.
      • Poupon R.
      • Chretien Y.
      • Chazouilleres O.
      • et al.
      Pregnancy in women with ursodeoxycholic acid-treated primary biliary cirrhosis.
      In the present study, there was a strong linear relationship between parity and age at PSC diagnosis, suggesting that pregnancies may delay the disease process. One theoretical explanation could be that the number of children was limited by acquired disease. However, the median age at diagnosis in the groups with 2, 3, and 4 children was after typical childbearing age (40, 43, and 54, respectively), suggesting that the correlation between number of children and age at diagnosis is not influenced by the diagnosis of PSC per se. Parity did not seem to affect the risk of developing PSC; the number of children in patients with PSC and the control subjects were similar, in line with a previous study in PSC, where normal fertility and no serious maternal or fetal complications were reported.
      • Wellge B.E.
      • Sterneck M.
      • Teufel A.
      • et al.
      Pregnancy in primary sclerosing cholangitis.
      The use of contraception was less prevalent among female patients with PSC than control subjects, and could be speculated to be a protective factor. However, it was not possible to assess in this study whether contraception use was lower because of the presence of disease. Overall, the observations regarding pregnancies and hormonal contraceptives indicate that hormonal factors may be protective in the pathogenesis of PSC and could be speculated to contribute to the gender differences in PSC.
      Several limitations need to be taken into account in the interpretation of the current study. Because selection of patients is done from a tertiary referral center, the patient cohort includes a high rate of previously transplanted patients and patients with severe phenotype. The response rate was 73% and it is not possible to rule out a selection bias, because the nonresponders were younger than the responders. Still, the overall characteristics of the included patients with PSC regarding gender, age, and IBD were as expected. Although patients with small-duct affection only or with established IgG4-related disease were excluded, routine IgG4 investigations have only been initiated recently in our institution. Our patient population represents the typical PSC phenotype described in the early 1980s,
      • Chapman R.W.
      • Arborgh B.A.
      • Rhodes J.M.
      • et al.
      Primary sclerosing cholangitis: a review of its clinical features, cholangiography, and hepatic histology.
      and it is therefore likely that a proportion has elevated serum-IgG4
      • Mendes F.D.
      • Jorgensen R.
      • Keach J.
      • et al.
      Elevated serum IgG4 concentration in patients with primary sclerosing cholangitis.
      but the significance of this is uncertain.
      Selection of control subjects for this patient population is particularly challenging. The control group was selected from previous blood donors volunteering for inclusion for bone marrow donor registry, and could be speculated to be healthier than the general population. However, the immune-related comorbidities seem comparable with the general population; the 12% prevalence among control subjects was driven by cases of psoriasis and hypothyroidism, 2 diseases that do not exclude from inclusion as a blood donor, and for which the prevalences were in line with findings from previous Norwegian questionnaire-based investigations.
      • Bo K.
      • Thoresen M.
      • Dalgard F.
      Smokers report more psoriasis, but not atopic dermatitis or hand eczema: results from a Norwegian population survey among adults.
      This was the case, despite the higher likelihood of acquiring diagnosis for patients already in the health care system, a phenomenon likely to inflate the differences regarding osteoporosis, which is often screened for in cholestatic diseases and after liver transplantation. For case and control matching, gender was prioritized, whereas the age matching resulted in a borderline significant 2-year age difference, and a wider age distribution in the patients than the control subjects (interquartile range, 20 vs 13 years, respectively). However, both smoking and coffee consumption were independently negatively associated with PSC in an age-adjusted model. Educational level was the main measure of sociodemographic factors in this questionnaire; however, future studies should collect even more sociodemographic data to improve case-control matching and enable more complex modeling of effects. The items included in our questionnaires were selected from several previously published questionnaires. Some of these items were modified, in particular to assess exposure both now and at the age of 18. These modifications have not been validated and because these relate to exposures many years previously, the data must be interpreted with caution. Still, the most important observations in this study are based on information on questions that intuitively are rather robust, such as coffee consumption or not, smoking, and number of children.
      In conclusion, our observations suggest that coffee consumption and hormonal factors influence the risk of development or onset of PSC, warranting further studies for confirmation of the results. In addition, this study strongly supports that smoking protects against PSC.

      Acknowledgments

      The authors thank the patients and control subjects participating in the study. Mona Bjørnstad, Kristian Holm, and Hege Dahlen Sollid at the Norwegian PSC research center, and Marte Wendel Gustavsen, Marthe Mæhlen, Marte K. Viken, and the members of the multiple sclerosis research group at Oslo University Hospital Ullevål and the immunogenetics group at University of Oslo are acknowledged for contributions to questionnaire design, distribution, collection, and/or scanning. The authors thank Jennifer Follestad for English translation of the questionnaire.

      Supplementary Material

      References

        • Karlsen T.H.
        • Schrumpf E.
        • Boberg K.M.
        Update on primary sclerosing cholangitis.
        Dig Liver Dis. 2010; 42: 390-400
        • Saarinen S.
        • Olerup O.
        • Broome U.
        Increased frequency of autoimmune diseases in patients with primary sclerosing cholangitis.
        Am J Gastroenterol. 2000; 95: 3195-3199
        • Lamberts L.E.
        • Janse M.
        • Haagsma E.B.
        • et al.
        Immune-mediated diseases in primary sclerosing cholangitis.
        Dig Liver Dis. 2011; 43: 802-806
        • Saich R.
        • Chapman R.
        Primary sclerosing cholangitis, autoimmune hepatitis and overlap syndromes in inflammatory bowel disease.
        World J Gastroenterol. 2008; 14: 331-337
        • Okolicsanyi L.
        • Fabris L.
        • Viaggi S.
        • et al.
        Primary sclerosing cholangitis: clinical presentation, natural history and prognostic variables: an Italian multicentre study. The Italian PSC Study Group.
        Eur J Gastroenterol Hepatol. 1996; 8: 685-691
        • Takikawa H.
        • Takamori Y.
        • Tanaka A.
        • et al.
        Analysis of 388 cases of primary sclerosing cholangitis in Japan. Presence of a subgroup without pancreatic involvement in older patients.
        Hepatol Res. 2004; 29: 153-159
        • Boberg K.M.
        • Aadland E.
        • Jahnsen J.
        • et al.
        Incidence and prevalence of primary biliary cirrhosis, primary sclerosing cholangitis, and autoimmune hepatitis in a Norwegian population.
        Scand J Gastroenterol. 1998; 33: 99-103
        • Kingham J.G.
        • Kochar N.
        • Gravenor M.B.
        Incidence, clinical patterns, and outcomes of primary sclerosing cholangitis in South Wales, United Kingdom.
        Gastroenterology. 2004; 126: 1929-1930
        • Bambha K.
        • Kim W.R.
        • Talwalkar J.
        • et al.
        Incidence, clinical spectrum, and outcomes of primary sclerosing cholangitis in a United States community.
        Gastroenterology. 2003; 125: 1364-1369
        • Kaplan G.G.
        • Laupland K.B.
        • Butzner D.
        • et al.
        The burden of large and small duct primary sclerosing cholangitis in adults and children: a population-based analysis.
        Am J Gastroenterol. 2007; 102: 1042-1049
        • Takikawa H.
        • Manabe T.
        Primary sclerosing cholangitis in Japan: analysis of 192 cases.
        J Gastroenterol. 1997; 32: 134-137
        • Escorsell A.
        • Pares A.
        • Rodes J.
        • et al.
        Epidemiology of primary sclerosing cholangitis in Spain. Spanish Association for the Study of the Liver.
        J Hepatol. 1994; 21: 787-791
        • Hurlburt K.J.
        • McMahon B.J.
        • Deubner H.
        • et al.
        Prevalence of autoimmune liver disease in Alaska Natives.
        Am J Gastroenterol. 2002; 97: 2402-2407
        • Bergquist A.
        • Montgomery S.M.
        • Bahmanyar S.
        • et al.
        Increased risk of primary sclerosing cholangitis and ulcerative colitis in first-degree relatives of patients with primary sclerosing cholangitis.
        Clin Gastroenterol Hepatol. 2008; 6: 939-943
        • Folseraas T.
        • Melum E.
        • Rausch P.
        • et al.
        Extended analysis of a genome-wide association study in primary sclerosing cholangitis detects multiple novel risk loci.
        J Hepatol. 2012; 57: 366-375
        • Liu J.Z.
        • Hov J.R.
        • Folseraas T.
        • et al.
        Dense genotyping of immune-related disease regions identifies nine new risk loci for primary sclerosing cholangitis.
        Nat Genet. 2013; 45: 670-675
        • Melum E.
        • Franke A.
        • Schramm C.
        • et al.
        Genome-wide association analysis in primary sclerosing cholangitis identifies two non-HLA susceptibility loci.
        Nat Genet. 2011; 43: 17-19
        • Hov J.R.
        • Lleo A.
        • Selmi C.
        • et al.
        Genetic associations in Italian primary sclerosing cholangitis: heterogeneity across Europe defines a critical role for HLA-C.
        J Hepatol. 2010; 52: 712-717
        • Mitchell S.A.
        • Thyssen M.
        • Orchard T.R.
        • et al.
        Cigarette smoking, appendectomy, and tonsillectomy as risk factors for the development of primary sclerosing cholangitis: a case control study.
        Gut. 2002; 51: 567-573
        • Loftus Jr., E.V.
        • Sandborn W.J.
        • Tremaine W.J.
        • et al.
        Primary sclerosing cholangitis is associated with nonsmoking: a case-control study.
        Gastroenterology. 1996; 110: 1496-1502
        • van Erpecum K.J.
        • Smits S.J.
        • van de Meeberg P.C.
        • et al.
        Risk of primary sclerosing cholangitis is associated with nonsmoking behavior.
        Gastroenterology. 1996; 110: 1503-1506
        • Rutgeerts P.
        • D'Haens G.
        • Hiele M.
        • et al.
        Appendectomy protects against ulcerative colitis.
        Gastroenterology. 1994; 106: 1251-1253
        • Russel M.G.
        • Dorant E.
        • Brummer R.J.
        • et al.
        Appendectomy and the risk of developing ulcerative colitis or Crohn's disease: results of a large case-control study. South Limburg Inflammatory Bowel Disease Study Group.
        Gastroenterology. 1997; 113: 377-382
        • Munger K.L.
        • Levin L.I.
        • Hollis B.W.
        • et al.
        Serum 25-hydroxyvitamin D levels and risk of multiple sclerosis.
        JAMA. 2006; 296: 2832-2838
        • Ng S.C.
        • Bernstein C.N.
        • Vatn M.H.
        • et al.
        Geographical variability and environmental risk factors in inflammatory bowel disease.
        Gut. 2013; 62: 630-649
        • Sugiyama D.
        • Nishimura K.
        • Tamaki K.
        • et al.
        Impact of smoking as a risk factor for developing rheumatoid arthritis: a meta-analysis of observational studies.
        Ann Rheum Dis. 2010; 69: 70-81
        • Gershwin M.E.
        • Selmi C.
        • Worman H.J.
        • et al.
        Risk factors and comorbidities in primary biliary cirrhosis: a controlled interview-based study of 1032 patients.
        Hepatology. 2005; 42: 1194-1202
        • Chapman R.W.
        • Arborgh B.A.
        • Rhodes J.M.
        • et al.
        Primary sclerosing cholangitis: a review of its clinical features, cholangiography, and hepatic histology.
        Gut. 1980; 21: 870-877
        • Hedstrom A.K.
        • Baarnhielm M.
        • Olsson T.
        • et al.
        Tobacco smoking, but not Swedish snuff use, increases the risk of multiple sclerosis.
        Neurology. 2009; 73: 696-701
        • Lund E.
        • Dumeaux V.
        • Braaten T.
        • et al.
        Cohort profile: The Norwegian Women and Cancer Study–NOWAC–Kvinner og kreft.
        Int J Epidemiol. 2008; 37: 36-41
        • Maniaol A.H.
        • Brunborg C.
        • Tallaksen C.M.
        Development and validation of a self-administered questionnaire for myasthenia gravis patients.
        Neuroepidemiology. 2010; 34: 253-261
        • Krokstad S.
        • Langhammer A.
        • Hveem K.
        • et al.
        Cohort profile: The HUNT Study, Norway.
        Int J Epidemiol. 2013; 42: 968-977
        • Pugliatti M.
        • Casetta I.
        • Drulovic J.
        • et al.
        A questionnaire for multinational case-control studies of environmental risk factors in multiple sclerosis (EnvIMS-Q).
        Acta Neurol Scand Suppl. 2012; : 43-50
        • Landis J.R.
        • Koch G.G.
        The measurement of observer agreement for categorical data.
        Biometrics. 1977; 33: 159-174
        • Leitzmann M.F.
        • Willett W.C.
        • Rimm E.B.
        • et al.
        A prospective study of coffee consumption and the risk of symptomatic gallstone disease in men.
        JAMA. 1999; 281: 2106-2112
        • Lillemoe K.D.
        • Magnuson T.H.
        • High R.C.
        • et al.
        Caffeine prevents cholesterol gallstone formation.
        Surgery. 1989; 106 (discussion 406–407): 400-406
        • Nilssen O.
        • Forde O.H.
        • Brenn T.
        The Tromso Study. Distribution and population determinants of gamma-glutamyltransferase.
        Am J Epidemiol. 1990; 132: 318-326
        • Honjo S.
        • Kono S.
        • Coleman M.P.
        • et al.
        Coffee consumption and serum aminotransferases in middle-aged Japanese men.
        J Clin Epidemiol. 2001; 54: 823-829
        • Torres D.M.
        • Harrison S.A.
        Is it time to write a prescription for coffee? Coffee and liver disease.
        Gastroenterology. 2013; 144: 670-672
        • Bravi F.
        • Bosetti C.
        • Tavani A.
        • et al.
        Coffee drinking and hepatocellular carcinoma: an update.
        Hepatology. 2009; 50: 1317-1318
        • Freedman N.D.
        • Everhart J.E.
        • Lindsay K.L.
        • et al.
        Coffee intake is associated with lower rates of liver disease progression in chronic hepatitis C.
        Hepatology. 2009; 50: 1360-1369
        • Molloy J.W.
        • Calcagno C.J.
        • Williams C.D.
        • et al.
        Association of coffee and caffeine consumption with fatty liver disease, nonalcoholic steatohepatitis, and degree of hepatic fibrosis.
        Hepatology. 2012; 55: 429-436
        • Masterton G.S.
        • Hayes P.C.
        Coffee and the liver: a potential treatment for liver disease?.
        Eur J Gastroenterol Hepatol. 2010; 22: 1277-1283
        • Moreno M.G.
        • Chavez E.
        • Aldaba–Muruato L.R.
        • et al.
        Coffee prevents CCl(4)-induced liver cirrhosis in the rat.
        Hepatol Int. 2011; 5: 857-863
        • Fukushima Y.
        • Kasuga M.
        • Nakao K.
        • et al.
        Effects of coffee on inflammatory cytokine gene expression in mice fed high-fat diets.
        J Agric Food Chem. 2009; 57: 11100-11105
        • Freedman N.D.
        • Park Y.
        • Abnet C.C.
        • et al.
        Association of coffee drinking with total and cause-specific mortality.
        N Engl J Med. 2012; 366: 1891-1904
        • Ascherio A.
        • Weisskopf M.G.
        • O'Reilly E.J.
        • et al.
        Coffee consumption, gender, and Parkinson's disease mortality in the cancer prevention study II cohort: the modifying effects of estrogen.
        Am J Epidemiol. 2004; 160: 977-984
        • Persson P.G.
        • Ahlbom A.
        • Hellers G.
        Diet and inflammatory bowel disease: a case-control study.
        Epidemiology. 1992; 3: 47-52
        • Cohen A.B.
        • Lee D.
        • Long M.D.
        • et al.
        Dietary patterns and self-reported associations of diet with symptoms of inflammatory bowel disease.
        Dig Dis Sci. 2013; 58: 1322-1328
        • Berger J.
        • Wynder E.L.
        The correlation of epidemiological variables.
        J Clin Epidemiol. 1994; 47: 941-952
        • Smits P.
        • Temme L.
        • Thien T.
        The cardiovascular interaction between caffeine and nicotine in humans.
        Clin Pharmacol Ther. 1993; 54: 194-204
        • Lakatos P.L.
        • Szamosi T.
        • Lakatos L.
        Smoking in inflammatory bowel diseases: good, bad or ugly?.
        World J Gastroenterol. 2007; 13: 6134-6139
        • Loftus Jr., E.V.
        • Harewood G.C.
        • Loftus C.G.
        • et al.
        PSC-IBD: a unique form of inflammatory bowel disease associated with primary sclerosing cholangitis.
        Gut. 2005; 54: 91-96
        • Heneghan M.A.
        • Norris S.M.
        • O'Grady J.G.
        • et al.
        Management and outcome of pregnancy in autoimmune hepatitis.
        Gut. 2001; 48: 97-102
        • Poupon R.
        • Chretien Y.
        • Chazouilleres O.
        • et al.
        Pregnancy in women with ursodeoxycholic acid-treated primary biliary cirrhosis.
        J Hepatol. 2005; 42: 418-419
        • Wellge B.E.
        • Sterneck M.
        • Teufel A.
        • et al.
        Pregnancy in primary sclerosing cholangitis.
        Gut. 2011; 60: 1117-1121
        • Mendes F.D.
        • Jorgensen R.
        • Keach J.
        • et al.
        Elevated serum IgG4 concentration in patients with primary sclerosing cholangitis.
        Am J Gastroenterol. 2006; 101: 2070-2075
        • Bo K.
        • Thoresen M.
        • Dalgard F.
        Smokers report more psoriasis, but not atopic dermatitis or hand eczema: results from a Norwegian population survey among adults.
        Dermatology. 2008; 216: 40-45