Risk Factors for Primary Sclerosing Cholangitis

Kirsten Boonstra; Elisabeth M. G. de Vries; Nan van Geloven; Karel J. van Erpecum; Marcel Spanier; Alexander C. Poen; Carin M. van Nieuwkerk; Ben J. Witteman; Hans A. Tuynman; Anton H. Naber; Paul J. Kingma; Ulrich Beuers; Cyriel Y. Ponsioen


Liver International. 2016;36(1):84-91. 

In This Article


Study Inclusion

The case-finding exercise yielded 697 PSC patients, 254 HC's and 404 IBD controls. Four-hundred and fifty PSC patients were alive at time of inclusion (between 1 January 2008, and 31 December 2011) and were asked to give informed consent and to fill out a 10-item questionnaire. In total, 343 PSC patients completed the questionnaire resulting in a response rate of 76%. A total of 232 (92%) HC's and 370 (92%) IBD controls completed the questionnaire.

At time of diagnosis 590 of the 697 identified PSC patients lived in the geographically defined area, creating the population-based Epi PSC PBC cohort as described previously.[1] The subgroup of 343 PSC patients analysed in this study was not different from the entire population-based cohort, and is thereby representative of the PSC population in the Netherlands. (Table S1 http://onlinelibrary.wiley.com/doi/10.1111/liv.12894/suppinfo) The subgroup was only selected on the basis of being alive at the time of inclusion in the database and receipt of the questionnaire that had been sent to them.

Analysis of geographical distribution was performed in 218 patients that were diagnosed with PSC between 2000 and 2008 while living in the geographically defined area (Fig. S1 http://onlinelibrary.wiley.com/doi/10.1111/liv.12894/suppinfo).

Patient Characteristics

Patient characteristics are summarized in Table 1. Of the 343 PSC patients, 168 (49%) had concomitant UC, 50 (15%) CD and the remaining 125 (36%) patients did not have IBD. Sixty-five per cent of PSC patients were male and the median age at time of inclusion was 48 years (IQR 38–59 years). Healthy controls had a median age at inclusion of 55 years (IQR 42–65 years) and 46% were male. Of IBD controls 40% were male and the median age at inclusion was 43 years (IQR 31–55).

Risk Factors

Smoking. As shown in Table 1, there were less current smokers (8%) in the PSC group than in the HC group (19%). PSC patients were also less often former smokers (23%) compared to healthy controls (36%), with a statistically significant effect in multivariable analysis (OR 0.52 95% CI 0.35–0.75; P = 0.001) (Table 2). A similar trend was observed in the subgroup of PSC patients without IBD (n = 125) compared to healthy controls (OR 0.65 95% CI 0.40–1.06; P = 0.085) (Table 2).

When comparing IBD patients to PSC-IBD patients, UC patients were more often current smokers (19%), and former smokers (54%) compared to PSC-UC patients (6%, 22% respectively). Similar results were seen for CD patients compared to PSC-CD patients. When assessing the independent effect in multivariable analysis, former smoking was significantly associated with a lower risk for the development of PSC in both UC (OR 0.21; 95% CI 0.12–0.34; P < 0.001) and CD patients (OR 0.17; 95% CI 0.08–0.39; P < 0.001) (Table 2).

Appendectomy. An equal proportion of PSC patients and HC underwent appendectomy (13% in both groups). In one PSC patient the appendix was removed as part of an ileocecal resection. More PSC-UC patients (13%) than UC controls (6%) had undergone appendectomy (Table 1). In these subgroups all appendectomies were performed for clinical suspicion of acute appendicitis. When correcting for possible interacting variables using multivariable modelling, the positive association persisted in this subgroup (OR 2.51; 95%CI 1.04–6.07; P = 0.041) (Table 2). Appendectomy as risk factor for PSC was not assessed in the subgroups PSC-CD vs. CD, because appendectomy was the result of ileocecal resection in the majority of patients in these subgroups.

Family History. Seventeen (10%) PSC-UC patients and 28 (17%) UC controls had a first-degree family member with IBD. No differences in IBD prevalence were found between families of PSC-CD patients and CD controls (16% vs. 17% respectively) (Table 1). Multivariable analysis showed that family history of IBD was not an independent risk factor for PSC in the PSC-UC – UC nor PSC-CD – CD subgroups (OR 0.57; 95%CI 0.28–1.16; P = 0.12 and OR 0.94; 95% CI 0.37–2.38; P = 0.90 respectively) (Table 2). The percentages of first-degree relatives with autoimmune liver diseases were similar for PSC-UC patients (1%) and UC controls (1%). Autoimmune liver diseases were more prevalent in families of PSC-CD patients than in families of CD controls, however, numbers were very small (2% vs. 0%) (Table 1). Family history was not assessed in the subgroups PSC vs. HC, as the HC group was composed partly from healthy family members accompanying PSC patients to the outpatient clinic, which would create an unrepresentative high percentage of HC's with PSC family members.

Geographical Distribution. When combining municipalities according to degree of urbanization, the observed number of PSC patients was very similar to the expected number of PSC patients resulting in standardized incidence ratio's close to unity in each category (Table 3). There was no evidence for a trend in incidence over the degrees of urbanization (P = 0.97).