Increased Risk of Primary Sclerosing Cholangitis and Ulcerative Colitis in First-Degree Relatives of Patients With Primary Sclerosing Cholangitis

Transcription

1 CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2008;6: Increased Risk of Primary Sclerosing Cholangitis and Ulcerative Colitis in First-Degree Relatives of Patients With Primary Sclerosing Cholangitis ANNIKA BERGQUIST,* SCOTT M. MONTGOMERY,, SHAHRAM BAHMANYAR, ROLF OLSSON, ÅKE DANIELSSON, STEFAN LINDGREN,** HANNE PRYTZ, ROLF HULTCRANTZ,* LARS LÖÖF, HANNA SANDBERG GERTZÉN, SVEN ALMER, JOHAN ASKLING, ANNA EHLIN, and ANDERS EKBOM *Department of Gastroenterology and Hepatology, Karolinska University Hospital, Karolinska Institute, Huddinge and Solna, Stockholm; Clinical Epidemiology Unit, Department of Medicine, Karolinska Hospital, Karolinska Institute, Stockholm; Clinical Research Centre, Örebro University Hospital, Örebro; Department of Medicine, Sahlgrenska University Hospital, Gothenburg; Department of Medicine, Section for Gastroenterology and Hepatology, University Hospital, Umeå; **Department of Gastroenterology and Hepatology, University Hospital, Malmö; Division of Gastroenterology and Hepatology, University Hospital, Lund; Centre for Clinical Research, Central Hospital, Västerås; Division of Gastroenterology and Hepatology, Department of Medicine, Medical Center Hospital, Örebro; and Division of Gastroenterology and Hepatology/IKE, University Hospital, Linköping, Sweden Background & Aims: The importance of genetic factors for the development of primary sclerosing cholangitis (PSC) is incompletely understood. This study assessed the risk of PSC and inflammatory bowel disease (IBD) among first-degree relatives of patients with PSC, compared with the first-degree relatives of a cohort without PSC. Methods: Subjects from the national Swedish cohort of PSC patients (n 678) were matched for date of birth, sex, and region to up to 10 subjects without a diagnosis of PSC (n 6347). Linkage through general population registers identified first-degree relatives of subjects in both the PSC and comparison cohorts (n 34,092). Diagnoses among first-degree relatives were identified by using the Inpatient Register. Results: The risk of cholangitis was statistically significantly increased in offspring, siblings, and parents of the PSC patient cohort, compared with relatives of the comparison cohort, with the hazard ratios and 95% confidence intervals, 11.5 ( ), 11.1 ( ), and 2.3 ( ), respectively. The hazard ratios for ulcerative colitis (UC) among first-degree relatives of all PSC patients was 3.3 ( ) and for Crohn s disease 1.4 ( ). The risk of UC for relatives of PSC patients without IBD was also increased, 7.4 ( ). Conclusions: First-degree relatives of patients with PSC run an increased risk of PSC, indicating the importance of genetic factors in the etiology of PSC. First-degree relatives of PSC patients without IBD are also at an increased risk of UC, which might indicate shared genetic susceptibility factors for PSC and UC. Primary sclerosing cholangitis (PSC) is an idiopathic chronic cholestatic liver disease of presumed autoimmune etiology, characterized by diffuse fibrosing inflammation of the intrahepatic and/or extrahepatic bile ducts, resulting in bile duct obliteration, biliary cirrhosis, and eventually hepatic failure. 1 Approximately 80% of all PSC patients have concomitant ulcerative colitis (UC). The association between inflammatory bowel disease (IBD) and PSC, the fact that patients with PSC are at increased risk of other autoimmune disorders, 2 and the association with the HLA B8-DR3 haplotype 3 suggest that the immune system plays a critical role in the etiology of PSC. The genetic basis of human autoimmune diseases is receiving increased attention. It is well-established that genetic factors play important roles in the susceptibility to autoimmune diseases, and familial occurrence of PSC has previously been reported. 4 In recent years the knowledge about the importance of genetic polymorphisms or variants in PSC has increased. With respect to a potential genetic link between PSC and IBD, a number of IBD susceptibility genes were recently evaluated in 365 Scandinavian PSC patients and in healthy controls, without any evidence of association between any of the investigated genetic IBD risk variants and overall PSC risk. 5 Studies of polymorphisms of the adhesion molecules mucosal addressin cellular adhesion molecule-1 and intercellular adhesion molecule-1, which have been implicated in the pathogenesis of PSC, 6 have not provided convincing evidence that these polymorphisms contribute to PSC susceptibility. 7 Polymorphisms outside the HLA complex have been evaluated including the multidrug resistance gene 3 8 and the steroid and xenobiotic receptor, 9 but there is no convincing evidence that they are associated with PSC susceptibility. Although convincing evidence is lacking to support the role of a genetic component for PSC susceptibility, this needs to be explored further. The present study was designed to assess the risk of PSC, UC, and Crohn s disease among first-degree relatives of patients with PSC, compared with the first-degree relatives of a cohort without these diseases. The study made use of our national cohort of patients with a diagnosis of PSC, as well as Swedish general population register data with linkage between generations. Subjects and Methods Index Cohort of Patients With Primary Sclerosing Cholangitis The study was performed in collaboration with the Swedish Internal Medicine Liver Club, representing hepatologists from all Swedish University hospitals. The cohort of patients with PSC comprised all registrations of the disease between 1970 and 2003 in all 10 University hospitals including all 3 transplant centers. Inclusion criteria were a diagnosis of PSC Abbreviations used in this paper: CI, confidence interval; IBD, inflammatory bowel disease; ICD, International Classification of Diseases; PSC, primary sclerosing cholangitis; UC, ulcerative colitis by the AGA Institute /08/$34.00 doi: /j.cgh

2 940 BERGQUIST ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 6, No. 8 Table 1. Clinical Characteristics of 678 Patients With PSC No. of patients (%) (n 678) Male sex 465 (69) Mean age at onset of PSC ( SD) 36 ( 12) y IBD 545 (80) UC 463 (68) Crohn s disease 59 (9) Indeterminate colitis 23 (3) Colectomy 145 (21) Liver transplanted 142 (21) Dead at end of the study 166 (25) on the basis of biochemical, clinical, and cholangiographic features. 10 A total of 700 patients with this diagnosis were identified, but 22 were excluded because they had emigrated and could not be matched with the comparison cohort. Thus, 678 patients with PSC were available for analysis. Table 1 shows the clinical characteristics of the PSC patient cohort. All of these patients records were scrutinized to confirm the diagnosis and onset of PSC and to record information on IBD and colectomy. Onset of PSC was defined as the time of the first cholangiographic investigation consistent with a diagnosis of PSC. The diagnosis of IBD was based on typical history and characteristic endoscopic and histologic findings. 11 Comparison Cohort A comparison cohort of subjects without a diagnosis of PSC was identified, with up to 10 individuals matched with each patient with PSC. Matching was for date of birth, sex, and region (health care district). Members of the comparison cohort were selected at random in the general population register by Statistics Sweden, the government organization responsible for national statistics. A total of 6347 individuals were identified. First-Degree Relatives of Patients With Primary Sclerosing Cholangitis and of the Comparison Subjects To identify first-degree relatives of subjects in the PSC cohort and in the comparison cohort, we linked these individuals (using the unique identification number issued to all Swedish Table 2. Characteristics of First-Degree Relatives Relatives of PSC cohort Relatives of comparison cohort Total ,953 Offspring Male (%) 496 (53.3) 5035 (51.5) Siblings ,529 Male (%) 540 (52.0) 5352 (50.8) Parents ,651 Male (%) 576 (49.2) 5211 (48.9) Mean age at first record of PSC, y Offspring Siblings Parents Mean follow-up time, y residents) to the Multi-generation Register. In this register, parents, siblings, and offspring of individuals born in 1932 or later and alive in 1961 or later can be identified. Linkage with the Swedish Emigration and Cause of Death Registers allowed assessment of vital status throughout our study period, which ended in Follow-up among the relatives was from 1965 or, if born after this date, from their date of birth. Relatives who died or emigrated before follow-up were excluded (n 204), and an additional 602 (9%) relatives were excluded as a result of data inconsistencies precluding unambiguous follow-up. Figure 1 describes how the cohorts were created. Table 2 shows the characteristics of the relatives. The number of relatives identified per patient/control, the sex- and age-distribution of these relatives, and their mean time under follow-up were similar for patients and controls. The sex ratios for each type of relative were not significantly different, indicating that the results were not confounded by sex. There was no significant difference in average follow-up time for relatives of the 3 cohorts. Occurrence of Cholangitis, Ulcerative Colitis, and Crohn=s Disease All first-degree relatives discharged from a hospital with a diagnosis of cholangitis, UC, and/or Crohn=s disease between Figure 1. The cohort of first-degree relatives of patients with PSC and of controls.

3 August 2008 HEREDITY IN PSC 941 Table 3. PSC Among First-Degree Relatives of Patients With PSC: Events and Risks Compared With First-Degree Relatives of a Comparison Cohort First-degree relatives of patients with PSC Cholangitis a Cholangitis excluding gallbladder disease Cholangitis and UC Events Risk (95% CI) Events Risk (95% CI) Events Risk (95% CI) Offspring b ( ) ( ) e ( ) Siblings ( ) ( ) ( ) c Parents ( ) ( ) ( ) All first-degree relatives d ( ) ( ) ( ) a All patients with cholangitis, including gallbladder disease. b Cox proportional hazard modeling was used, with adjustment for attained age and stratified by risk set, producing hazard ratios (with 95% CIs). c Estimated without stratification as a result of sparse data. d Conditional logistic regression was used, stratified on risk set and producing odds ratios (with 95% CIs). e Because of the structure of the data, this estimate did not change after the exclusions and December 31, 2004 were identified from the Inpatient Register by using the relevant codes according to International Classification of Diseases (ICD). The population-based Swedish Inpatient Register contains information on all inpatient care since 1964 (nationwide since 1987). About 99% of all hospitalizations are included. 12 This includes all cases in whom an endoscopic retrograde cholangiography was conducted but does not include PSC patients diagnosed as outpatients by using only magnetic resonance cholangiography unless they were later identified after an inpatient care episode. Cholangitis was identified as ICD-7: , ICD-8: , ICD-9: 576B, ICD-10: K830; UC as ICD-7: 572.2, ICD-8: 563.1, ICD-9: 556, ICD-10: K51; and Crohn s disease as ICD-7: 572.0, ICD-8: 565.0, ICD-9: 555, ICD-10: K50. Statistical Analysis We used Cox proportional hazards models, with attained age as the underlying time scale to estimate the risk of PSC among first-degree relatives of the PSC cohort compared with relatives of the comparison cohorts. The offspring, siblings, and parents were the units of analysis, and these groups were investigated separately. The analysis was stratified by risk set defined by index members of the original cohorts. One estimate could not be generated with internal stratification as a result of sparse data. Therefore, this was calculated with a nonstratified model, and the resulting estimate is identified in Table 3. Final follow-up was defined as first diagnosis of outcome disease, liver transplant, death, or emigration; whichever came first and before December 31, The outcome diseases were PSC, UC, and Crohn s disease. The codes for cholangitis are not specific to PSC and include other entities such as secondary sclerosing cholangitis. Therefore, we used 3 separate diagnostic models. The first model used diagnosis of cholangitis as the dependent variable. For more specific definitions of PSC, we identified cholangitis without concomitant gallbladder disease, and we also defined PSC as cholangitis with a diagnosis of UC. To assess the risk of the outcome diseases among all firstdegree relatives (parents, offspring, and siblings), conditional logistic regression was used where the members of the original cohorts were the unit of analysis. Thus, the relative risk of having any first-degree relative with a diagnosis of the outcome disease was examined. These analyses were conditioned on risk set. The pedigrees for all families with a positive event were examined to establish whether multiple affected members of the same families could have influenced the results (non-independence). To tackle potential concerns about surveillance or information bias, the analysis of PSC risk was repeated, excluding relatives who appeared in the original cohort of patients with PSC. To assess whether associations with IBD rather than PSC account for the results, the analyses were repeated among relatives of the subset of the PSC cohort (and their matched comparison subjects) who did not have a diagnosis of Crohn s disease or UC. All statistical analyses were performed with SAS, release 9.1 (SAS Institute, Cary, NC); PROC PHREG was used for the regression models. All statistical tests were 2-sided, and statistical significance was defined as 95% confidence intervals (CIs) that did not include The Ethics Committee at Karolinska University hospital, Huddinge approved this study. Results Risk of Primary Sclerosing Cholangitis in First-Degree Relatives A diagnosis of cholangitis was found in 0.5% (16/3130) among first-degree relatives of patients with PSC and in 0.1% (41/30,953) among first-degree relatives of the comparison cohort (Figure 1). The risk of cholangitis was statistically significantly increased in offspring, siblings, and parents of the PSC patient cohort, compared with relatives of the comparison cohort (Table 3). To improve the specificity of the PSC diagnosis among first-degree relatives, we also performed analyses for cholangitis without gallbladder disease and cholangitis with UC. The risk of cholangitis without gallbladder disease was statistically significantly increased for offspring and siblings of the PSC cohort, whereas there was a nonstatistically significant increase in risk among parents. The risk of cholangitis with UC was statistically significantly increased among the offspring of PSC patients, but the data were rather sparse for estimation among siblings and parents. When all relatives were combined, there was a statistically significant increase in risk for all definitions of PSC, particularly where it was defined as cholangitis with UC. There were no families with more than 2 members with PSC, so non-independence caused by several patients in the same family was not a concern. When members of the

4 942 BERGQUIST ET AL CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 6, No. 8 Table 4. IBD Among First-Degree Relatives of Patients With PSC: Events and Risks Compared With First- Degree Relatives of a Comparison Cohort First-degree relatives of patients with PSC UC Crohn s disease Events Risk (95% CI) Events Risk (95% CI) Offspring a ( ) ( ) Siblings a ( ) ( ) Parents a ( ) ( ) All first-degree relatives b ( ) ( ) a Cox proportional hazard modeling was used, with adjustment for attained age and stratified by risk set, producing hazard ratios (with 95% CIs). b Conditional logistic regression was used, stratified on risk set and producing odds ratios (with 95% CIs). patient cohort were excluded from relatives at risk of PSC (among all first-degree relatives) to deal with concerns about surveillance or information bias, the odds ratios for PSC remained statistically significant, 2.5 (95% CI, ) for cholangitis and 2.2 (95% CI, ) for cholangitis without gallbladder disease. This exclusion is likely to underestimate the true association. The data in this restricted group were too sparse to produce an estimate for cholangitis with UC. Risk of Inflammatory Bowel Disease in First- Degree Relatives The occurrence of UC among first-degree relatives of patients with PSC was 0.9% (29/3139) and 0.3% (98/30953) in the relatives of the comparison cohort. There was a statistically significant increased risk of UC among offspring, siblings, and parents of PSC patient cohort compared with relatives of the comparison cohort (Table 4). There was a statistically significant risk of Crohn s disease among offspring but not other first-degree relatives. A total of 3 families had multiple members with UC (none had more than one with Crohn s disease). In none of the 3 families with several members with UC was there more than one with the diagnosis among the same type of first-degree relatives (offspring, siblings, or parents). Because the association was statistically significant for each of these groups separately, this minimizes the concern that the results are driven by multiple occurrence of the disease in the same family. Disease Risk in Relatives of Patients With Primary Sclerosing Cholangitis but Not Inflammatory Bowel Disease Among relatives of the 135 PSC cohort members without IBD and the relatives of the matched comparison cohort, 9 had PSC, 18 had UC, and 13 had a diagnosis of Crohn s disease. Sparse data only permitted an estimate of PSC, UC, and Crohn s disease risk among parents and all first-degree relatives combined. Statistically significant increased risks were observed for PSC, UC, and Crohn s disease among relatives of the PSC cohort without IBD (Table 5). Discussion This study found a nearly 4-fold increased risk of PSC in first-degree relatives of patients with PSC. The familial occurrence of PSC is well-established, and we have previously reported 5 index cases with PSC who had a first-degree relative with PSC among 145 PSC patients. 4 The previous study indicated a 100-fold increased risk of PSC among firstdegree relatives compared with the PSC prevalence reported in the general Scandinavian population (8/100,000). 4,13 However, this substantial increase in risk is possibly an overestimate and might at least partly be explained by underestimation of PSC prevalence in the general population. Many PSC patients are asymptomatic, have fluctuations in liver function tests down to normal levels, 14 and a cholangiogram is mandatory for the diagnosis. 10 All of these factors lead to underestimation of the true prevalence of PSC. It is also possible that the previous study might have been influenced by surveillance bias, leading to a higher detection rate of PSC in close relatives of patients with PSC. These individuals might have been more likely to have undergone liver function tests and, if these tests were abnormal, to have magnetic resonance cholangiography performed. The study reported here is not similarly affected by the potential sources of bias that might have affected previous results. The case-control setting and the use of the Inpatient Register counteract these biases, so the estimates of association are more reliable. The use of the Inpatients Register for identification of cholangitis might indeed underestimate the occurrence of PSC both among PSC relatives and in the comparison group because some of the patients never have been hospitalized. This does not, however, influence our results differentially, because the detection in PSC patients and controls was the same. In addition, when family members of the index cohort were excluded from the cohort of relatives at risk of PSC, the odds Table 5. PSC and IBD Among First-Degree Relatives of Patients With PSC Without IBD: Events and Risks Compared With First-Degree Relatives of a Comparison Cohort First-degree relatives of patients with PSC Cholangitis UC Crohn s disease Events Risk (95% CI) Events Risk (95% CI) Events Risk (95% CI) Parents a ( ) ( ) 0 No estimate b All first-degree relatives c ( ) ( ) ( ) a Cox proportional hazard modeling was used, with adjustment for attained age and stratified by risk set, producing hazard ratios (with 95% CIs). b Not estimated because of empty cells. c Conditional logistic regression was used, stratified on risk set and producing odds ratios (with 95% CIs).

5 August 2008 HEREDITY IN PSC 943 ratios for PSC remained statistically significant. This shows that surveillence bias is not a big concern. For monozygotic disorders the relative risk for the disease in siblings compared with the general population ranges from several hundred to thousands, whereas in complex traits it is usually below 100. Our findings are consistent with the definition of PSC as a polygenic disease. The lower risk of PSC among parents compared with the risk in siblings and offspring is most likely due to poorer rates of PSC diagnosis among the older population. PSC was not commonly diagnosed until the late 1970s and 1980s when endoscopic retrograde cholangiography became more commonly available. Today with increased knowledge of the disease and use of the noninvasive magnetic resonance cholangiography, PSC is diagnosed more frequently. Although there is no conclusive evidence of an increasing incidence of PSC, 15 we cannot exclude the possibility that PSC is less common among earlier birth cohorts. A disadvantage of the present study is the use of the ICD diagnostic codes that do not differentiate between PSC and bile duct disorders of other origins. To deal with this, we used 3 separate definitions of PSC. The least specific definition identified cholangitis, but we also defined PSC as cholangitis with UC, because approximately 80% of Swedish PSC patients have concomitant IBD. 14 Although the reduced numbers by using the definition of cholangitis with UC made estimation for some strata impossible, the consistency of results indicate that our estimates are reliable. It should be noted that the potential concern about diagnostic specificity of PSC applies only to identification of the disease among first-degree relatives. The diagnosis is highly reliable among the index cohort of PSC patients because these subjects were identified by gastroenterologists, and the diagnosis was confirmed through scrutiny of medical records. This study revealed more than a 3-fold increased risk of UC among first-degree relatives of patients with PSC. There was also some evidence of an increased risk of Crohn s disease among first-degree relatives, but this was less consistent. These findings indicate that PSC and IBD might share some genetic risk factors. Because the first-degree relatives of those with PSC but without IBD in this study were also at an increased risk of IBD, with risk estimates very close to those relatives with both PSC and IBD, this provides further evidence of a shared disposition with both PSC and IBD. Whether this is caused by genetic, environmental, or other factors remains elusive. However, the study by Karlsen et al 5 showing that none of the IBD candidate genes investigated were linked with PSC susceptibility does not provide evidence of genetic influences. In addition, we cannot ensure the absence of IBD in patients registered without an IBD because no information on whether colonoscopy was performed to rule out IBD is provided. UC in PSC patients has some characteristic features such as mild clinical course, rectal sparing, extensive colitis and backwash ileitis, and increased risk of colorectal malignancy, 16,17 which might indicate that UC with PSC represents a distinct phenotype. In conclusion, PSC is a complex trait, such that several genetic factors coupled with environmental factors are required for disease development. Individuals with a first-degree relative with PSC are at increased risk of PSC, indicating the importance of genetic factors in the etiology of PSC. Because the first-degree relatives of patients with PSC without IBD are at an increased risk of UC, there is further evidence that PSC and UC might share genetic susceptibility factors. References 1. Talwalkar J, Lindor K. Natural history and prognostic models in primary sclerosing cholangitis. Best Pract Res Clin Gastroenterol 2001;15: Saarinen S, Olerup O, Broomé U. Increased frequency of autoimmune diseases in patients with primary sclerosing cholangitis. Am J Gastroenterol 2000;95: Donaldson P, Norris S. Immunogenetics in PSC. Best Pract Res Clin Gastroenterol 2001;15: Bergquist A, Lindberg G, Saarinen S, et al. Increased prevalence for primary sclerosing cholangitis among first degree relatives. J Hepatol 2005;42: Karlsen T, Hampe J, Wiencke K, et al. Genetic polymorphisms associated with inflammatory bowel disease do not confer risk for primary sclerosing cholangitis. Am J Gastroenterol 2007;102: Eksteen B, Grant A, Miles A, et al. Hepatic endothelial CCL25 mediates the recruitment of CCR9 gut-homing lymphocytes to the liver in primary sclerosing cholangitis. J Exp Med 2004;200: Bowlus C, Karlsen T, Broomé U, et al. Analysis of MAdCAM-1 and ICAM-1 polymorphisms in 365 Scandinavian patients with primary sclerosing cholangitis. J Hepatol 2006;45: Pauli-Magnus C, Kerb R, Fattinger K, et al. BSEP and MDR3 haplotype structure in healthy Caucasians, primary biliary cirrhosis and primary sclerosing cholangitis. Hepatology 2004;39: Karlsen T, Lie B, Froslie K, et al. Polymorphisms in the steroid and xenobiotic receptor gene influence survival in primary sclerosing cholangitis. Gastroenterology 2006;131: Wiesner R, La Russo N. Clinicopathologic features of the syndrome of primary sclerosing cholangitis. Gastroenterology 1980; 74: Evans J, Acheson E. An epidemiological study of ulcerative colitis and regional enteritis in the Oxford area. Gut 1965;6: Centre for Epidemiology National Board of Health and Welfare. The Swedish Inpatient Register Stockholm: Boberg K, 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: Broomé U, Olsson R, Lööf L, et al. Natural history and prognostic factors in 305 Swedish patients with primary sclerosing cholangitis. Gut 1996;38: Bambha K KW, Talwalkar J, Torgerson H, et al. Incidence, clinical spectrum, and outcomes of primary sclerosing cholangitis in a United States community. Gastroenterology 2003;125: Loftus EJ, Harewood G, Loftus C, et al. PSC-IBD: a unique form of inflammatory bowel disease associated with primary sclerosing cholangitis. Gut 2005;54: Soetikno R, Lin O, Heidenreich P, et al. Increased risk of colorectal neoplasia in patients with primary sclerosing cholangitis and ulcerative colitis: a meta analysis. Gastrointest Endosc 2002;56: Address requests for reprints to: Annika Bergquist, MD, PhD, K63, Department of Gastroenterology and Hepatology, Karolinska University Hospital, S Stockholm, Sweden. annika.bergquist@ ki.se; fax: This study received grant support from Bengt Ihre Foundation.