Article Text

Download PDFPDF

Does the severity of primary sclerosing cholangitis influence the clinical course of associated ulcerative colitis?
  1. Laura Marelli1,
  2. Elias Xirouchakis1,
  3. Georgios Kalambokis1,
  4. Evangelos Cholongitas1,
  5. Mark I Hamilton2,
  6. Andrew K Burroughs1
  1. 1Liver Transplantation and Hepatobiliary Unit, Royal Free Hospital, London, UK
  2. 2Centre for Gastroenterology, Royal Free Hospital, London, UK
  1. Correspondence to Professor Andrew K Burroughs, Liver Transplantation and Hepatobiliary Medicine Unit, Royal Free Hospital, Pond Street, London NW3 2QG, UK; andrew.burroughs{at}nhs.net

Footnotes

  • See Commentary, p 1165

  • Linked article 240309.

  • LM—analysis and interpretation of data, study concept and design. EX—acquisition of data. GK—acquisition of data. EC—acquisition of data. MIH—critical revision of the manuscript for important intellectual content. AKB—study concept and design, study supervision.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

View Full Text

Statistics from Altmetric.com

Significance of this study

What is already known about this subject?

  • There is a strong association between primary sclerosing cholangitis (PSC) and ulcerative colitis (UC).

  • UC associated with PSC (PSC/UC) often runs a clinically quiescent course, but is characterised by pancolonic involvement and higher risk of colorectal cancer compared with UC alone.

  • PSC/UC may be a distinct phenotype of UC.

What are the new findings?

  • Clinically progressive PSC requiring liver transplantation is associated with a milder clinical course of UC (reduced disease activity, less use of steroids, azathioprine and surgery).

  • Patients with PSC/UC who have severe PSC have milder UC activity on histology and reduced incidence of dysplasia and colon carcinoma compared with patients with PSC/UC with mild liver involvement.

  • The severity of PSC may have a ‘protective’ effect on UC activity.

How might it impact on clinical practice in the foreseeable future?

  • Our results may help to clarify the pathogenesis and the relationship between PSC and UC, which are still largely unknown. This should allow further studies regarding lymphocytes trafficking between the liver and bowel, and vice versa.

Introduction

There is a strong association between primary sclerosing cholangitis (PSC) and inflammatory bowel disease (IBD), most commonly ulcerative colitis (UC). The prevalence of UC in patients with PSC varies widely from 23% in Japan up to 80% in Sweden.1 Patients with Crohn's disease affected by PSC are a smaller group ranging between 1.3% and 14%, and these are usually only patients with extensive colonic involvement.1

UC associated with PSC (PSC/UC) often runs a clinically quiescent course, but is endoscopically characterised by pancolonic involvement with more right-sided colitis,2 rectal sparing and backwash ileitis, and it also has a high risk of pouchitis postcolectomy.3 In addition, a meta-analysis has shown that PSC/UC has a higher risk of colorectal dysplasia and/or malignancy compared with UC alone.4 These clinical features seem to distinguish PSC/UC from UC alone and suggest that PSC/UC may be a distinct phenotype of UC.3

To date there is little information about a possible association between the activity of UC and severity of patients with PSC who have PSC/UC.

The aim of our study was to evaluate the clinical, endoscopic and histological features of PSC/UC in regards to treatment and outcome and to compare the course of PSC/UC in patients with stable liver disease, not needing liver transplantation (LT), and those with severe liver disease who required LT.

Patients and methods

We evaluated 109 consecutive patients with PSC and associated IBD seen at the Royal Free Hospital between 1990 and October 2009. They were initially referred by general practitioners (GPs) or local hospitals to either the hepatology clinic or to the gastroenterology clinic, depending on the initial symptoms or diagnosis of PSC and IBD, and then followed up in both clinics with 6 monthly clinical review. Among these 109 patients with PSC/IBD, 13 patients had Crohn's disease and they were excluded from this study. Thus, we enrolled 96 patients with PSC/UC in our analysis.

The diagnosis of PSC was established by characteristic biochemical findings and radiographic changes at cholangiogram (either percutaneous or retrograde) or, more recently, by magnetic resonance cholangiopancreatography. Conventional colonoscopic and histological criteria were used to diagnose IBD.

We constructed a database using medical records, including GP records where necessary, to collect the following data: demographic information (patients sex and age at last follow-up or LT); UC duration from diagnosis, colonic extent of UC; UC activity in the last 5 years of the follow-up period; type of UC treatment used during the whole follow-up period (steroid, azathioprine and surgery); severity of disease at last colonoscopy (before last follow-up or before LT), assessed both macroscopically and histologically; prevalence of colorectal dysplasia and/or carcinoma; PSC duration from diagnosis; severity of PSC at last follow-up or before LT (serum albumin, bilirubin values and presence of ascites); prevalence of cholangiocarcinoma during the whole follow-up period; outcome (patient alive at last follow-up, dead or had LT). After 1998, surveillance colonoscopy was performed every year in each patient with PSC/UC.

Activity of UC (in the last 5 years before last follow-up or LT) was defined as ‘quiescent’ in the absence of UC-related symptoms (eg, diarrhoea, rectal bleeding or abdominal pain), ‘requiring drugs’ when drugs, other than maintenance mesalazine, were required to control disease activity (steroid or immunosuppressant) and ‘surgery’ when colectomy was required. UC relapse was defined by the presence of clinical symptoms, with or without endoscopic confirmation, requiring a change in medical treatment.

The follow-up period started from the date of diagnosis and was censored at the time of the last clinical review, date of LT or death.

Our 96 patients with PSC/UC were divided into two groups:

  • No-LT group: 46 patients with PSC/UC who did not require LT

  • LT group: 50 patients with PSC/UC who required LT.

The course of UC was compared between these two groups using the Pearson χ2 test to compare categorical variables and the Mann–Whitney test to compare continuous variables.

Results

The median follow-up time (from date of original diagnosis of PSC and/or UC) was 132 months for the LT group (n=50) and 150 months for the no-OLT group (n=46). The patients' characteristics in the two groups are shown in table 1.

Table 1

Patient data

They were similar in terms of sex, follow-up time, IBD duration from diagnosis, PSC duration from diagnosis and colonoscopic extent of UC. The LT group was significantly younger than the no-LT group: mean age was 39±14 years compared with 51±16 years, respectively (p<0.0001). At diagnosis or at time of initial referral to our hospital, the severity of PSC was similar in the two groups, in terms of no significant difference in the median level of serum albumin (39 g/l in the LT group vs 42 g/l in the no-LT group) and bilirubin (33 μmol/l in the LT group vs 15 μmol/l in the no-LT group). While, as expected, at the end of the follow-up period or just before LT, the LT group had significantly lower serum albumin (median 33 g/l vs 41 g/l; p=0.003), significantly higher serum bilirubin (median 159 μmol/l vs 21 μmol/l; p=0.001) and more frequent ascites (54% vs 9%; p<0.0001) compared with the no-LT group.

The clinical features of UC in the two groups are shown in figures 1 and 2. UC activity (in the last 5 years before LT or last follow-up) was significantly more quiescent in the LT group (54% vs 20%) and more active (requiring medications) in the no-LT group (60% vs 36%) (p=0.002).

Figure 1

Ulcerative colitis activity (last 5 years follow-up).

Figure 2

Number of ulcerative colitis relapses throughout follow-up. LT, liver transplantation.

During the whole follow-up period, UC relapses requiring treatment with steroids were significantly more frequent in the no-LT group compared with the LT group; the number of UC exacerbations was >3 in 26% versus 4%, respectively (p=0.04). However, the number of patients requiring one or more hospital admission due to UC flare-ups (during the whole follow-up) was not significantly different between the two groups (24% in the no-LT group vs 29% in the LT group).

The number of steroid courses required in the two groups since diagnosis is shown in figure 3. The LT group had required significantly fewer courses of steroids and was given steroids for significantly shorter periods of time compared with the no-LT group. Steroids were never used in 44% (LT group) versus 17% (no-LT group). More than three courses were given in 12% (LT group) versus 31% (no-LT group) of patients (p=0.025). The duration of steroid courses was a median of 1 month (LT group) compared with 8 months (non-LT group) (p=0.022). There was no difference in terms of percentage of patients with steroid dependence (17% in the LT group vs 21% in the no-LT group). The use of azathioprine was significanlty more frequent in the no-LT group compared with the LT group (42% vs 12%, p=0.003), suggesting that relapses of UC were more severe and/or more prolonged and/or more difficult to control in the no-LT group.

Figure 3

Number of steroid courses throughout follow-up. LT, liver transplantation.

Patients who required surgical treatment for UC are shown in table 2. The no-LT group required colectomy significantly more frequently than the LT group (33% vs 10%; p=0.006). There was no difference between the two groups regarding the indication for surgery (control of acute flare-ups non-responsive to medical treatment, colorectal dysplasia or cancer) and type of surgery performed (proctocolectomy with ileostomy or colectomy with ilealanal pouch formation).

Table 2

Surgical treatment of ulcerative colitis

The results of the last colonoscopy performed before LT in the LT group, or before last clinical follow-up (or death) for the no-LT group are summarised in figure 4. There were no significant differences between the two groups in terms of macroscopic disease activity and the extent of disease seen at colonoscopy. However, the LT group had significantly milder disease histologically, compared with the no-LT group (84% vs 57% of patients had quiescent or mild colitis, p=0.011).

Figure 4

Endoscopic and histological features of ulcerative colitis at last colonoscopy (before liver transplantation (LT) or last follow-up).

The prevalence of dysplasia and cancer in the two groups is shown in table 3. The no-LT group had significantly higher prevalence of dysplasia (both low and high grade) compared with the LT group (low grade dysplasia in 13% vs 4%, and high grade dysplasia in 9% vs 0%, p=0.023). In the no-LT group four patients (9%) had developed colorectal cancer, while in the LT-group no patient developed colorectal cancer; two patients (4%) in the no-LT group had cholangiocarcinoma versus five patients (10%) in the LT group (three patients were diagnosed at explant and two patients with possible cholangiocarcinoma were transplanted in 1993 and 1997); however, these differences were not statistically significant. Conversely, the prevalence of colorectal cancer combined with high grade dysplasia was significantly higher in the no-LT group compared with the LT group (15% vs 0%, respectively; p=0.004).

Table 3

Prevalence of dysplasia and cancer throughout follow-up

Discussion

The association between PSC and IBD, mainly UC, is well recognised. In a study from the Mayo Clinic,3 PSC/UC was described as a unique form of IBD with distinguishing features such as pancolonic involvement, more right-sided colitis, rectal sparing and backwash ileitis. Several studies have shown that PSC/UC has a milder clinical course compared with UC alone2 3 5–9; however, a meta-analysis showed that PSC/UC carries an increased risk of developing colorectal dysplasia and/or cancer compared with UC alone, with an OR of 4.79.4 In addition, patients with PSC/UC have an increased risk of postcolectomy complications such as pouchitis10 11 and peristomal varices.12

However, associations between severity of liver disease and UC activity have not been reported. In this study, we analysed the features of colonic disease in a consecutive series of 96 patients with PSC/UC, seen at our centre, comparing a group that underwent LT due to severe liver disease and a group that was not transplanted who had similar length of follow-up and intervals from diagnosis of both PSC and UC. Selection bias due to referral practices is unlikely to have occurred as median values of bilirubin and albumin at diagnosis or at time of initial referral to our hospital were similar in the two groups. However, at the last follow-up or at the time of LT, median bilirubin was significantly lower and median albumin significantly higher in the non-LT group, who also had significantly less ascites. This confirms that the two groups had distinctly different patterns of progression of PSC.

Our results showed that the patients with PSC/UC with severe liver involvement requiring LT compared with no-LT had a significantly milder clinical course of UC (in the last 5 years of follow-up) and required fewer drugs (other than mesalazine), and less surgery to control UC activity (p=0.002) throughout follow-up. The LT group also had significantly fewer UC relapses (p=0.04), required fewer courses of steroids (p=0.025) for a shorter period of time (p=0.022), and fewer patients required azathioprine (p=0.025), compared with the no-LT group.

This milder UC in patients with PSC/UC with severe liver disease undergoing LT was confirmed on histology: biopsy samples taken at the last colonoscopy before LT showed significantly milder UC activity compared with the last colonoscopy in patients with PSC/UC who were not transplanted (p=0.011).

Interestingly, the patients with PSC/UC who eventually needed LT had a significantly reduced prevalence of colorectal dysplasia and/or cancer throughout follow-up before LT, compared with those with PSC/UC with mild PSC not undergoing LT, despite a similar length of follow-up and interval from diagnosis of UC and PSC.

The pathogenesis of both PSC and UC is largely unknown and less is known about relationships between the two. The only data regarding the potential reciprocal influence between the two diseases come from studies involving patients with PSC/UC in whom the impact of either colectomy or LT was assessed to evaluate what happened when one diseased organ is removed.

Thus the effect of colectomy on PSC in patients with PSC/UC was addressed by Cangemini et al: proctocolectomy appeared to have no effect on liver biochemistry, histology or patient survival.13 Similar results were reported by Martin et al14 who found no difference in patient survival between a group of patients with PSC undergoing colectomy, compared with a group of unoperated patients with PSC.

Conversely, a retrospective study by Vera et al15 and a follow-up16 showed that PSC recurrence post-LT was decreased in patients with PSC/UC who had had colectomy before or during LT. Similar results were confirmed by our group17: patients with PSC with UC and without colectomy before LT compared with no colectomy had a higher risk of recurrent PSC post-LT. The presence of an inflamed colon may be important in the pathogenesis and/or progression of PSC. It has been hypothesised that the immune response responsible for the development of PSC may be initiated by the ingress of bacteria or other toxic metabolites through the diseased bowel wall.18 Eksteen et al19 showed that there is an overlapping expression of endothelial adhesion molecules and tissue-specific chemokines between the liver and the gut. They suggested that effector/memory lymphocytes are able to migrate to both sites, as part of an enterohepatic T cell recirculation. They proposed that effector T cells generated in the organised lymphoid tissue of the gut during active IBD persist as long-lived memory cells that may be recruited to the liver to trigger hepatic inflammation under specific circumstances. Alabraba et al16 used this model to explain why, in their update of the Birmingham series15 of patients with PSC/UC undergoing LT, an intact colon remained a significant risk factor for PSC recurrence.

Thus, the presence of an inflamed colon seems to be important for the development of PSC, but UC activity seems to be less important when PSC is already established.

However, the presence of PSC appears to have a modulating effect on UC, making it more quiescent compared with UC without PSC,2 3 5–9 although PSC/UC carries a higher risk of colorectal dysplasia and/or cancer.4

Our study has shown that in patients with PSC/UC, more severe PSC is associated with a milder course of UC, and vice versa. It is notable that several studies18 20–26 have shown that UC can become more aggressive after LT (ie, PSC is removed), despite the use of more immunosuppressive agents after LT, such as azathioprine and ciclosporin, which are themselves used to treat more active UC.

It is difficult to explain this ‘protective’ effect of PSC on UC activity. A hypothesis could be that severe liver disease, causing depression of T cell function, may prevent an immunological response against the colon; while after LT, a new balance of the immune system could favour an immune-mediated attack towards the colonic mucosa, enhancing UC activity. In addition, if PSC is determined by the recruitment of gut-homing lymphocytes into the liver, then PSC and UC share the same pool of lymphocytes able to cause organ damage. When there is severe liver disease these lymphocytes might all be sequestered in the liver, and the colon would be spared. When the affected liver is removed, these lymphocytes would be free to migrate to the colon and create an immune attack there.

In conclusion, our study shows that patients with PSC/UC with severe liver disease requiring LT have a milder course of UC, both clinically and histologically, and also have a reduced risk of colorectal dysplasia and/or cancer, compared with patients with PSC/UC with mild liver involvement. This relationship could result from possible links between the pathogenesis of liver and colonic disease, which, at present, is mainly unexplained and represents a challenge to be addressed in future studies.

References

View Abstract

Footnotes

  • See Commentary, p 1165

  • Linked article 240309.

  • LM—analysis and interpretation of data, study concept and design. EX—acquisition of data. GK—acquisition of data. EC—acquisition of data. MIH—critical revision of the manuscript for important intellectual content. AKB—study concept and design, study supervision.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles