Colorectal neoplasia in PSC–IBD patients: are times changing?


  • Harry Sokol,

    1. Service de Gastroenterologie, Hôpital Saint-Antoine, Assistance Publique – Hôpitaux de Paris (APHP), Paris, France
    2. ERL INSERM U 1057/UMR7203, Faculté de Médecine Saint-Antoine, Université Pierre et Marie Curie (UPMC), Paris 6, France
    Search for more papers by this author
  • Olivier Chazouilleres

    1. Service d'Hépatologie, Centre de référence des Maladies Inflammatoires des Voies Biliaires, Hôpital Saint-Antoine, Assistance Publique – Hôpitaux de Paris (APHP), Paris, France
    2. UMR_S938, Faculté de Médecine Saint-Antoine, Université Pierre et Marie Curie (UPMC), Paris 6, France
    Search for more papers by this author

Primary sclerosing cholangitis (PSC) is a chronic cholestatic liver disease of unknown aetiology characterized by inflammation and obliterative fibrosis of the biliary tree. Although the course may be variable, PSC is often progressive, leading to biliary cirrhosis and its complications. In addition, patients with PSC are exposed to an increased risk of cholangiocarcinoma and of other malignancies. Overall, PSC is a rather severe disease with a median transplant-free survival between 12 and 18 years. Currently, there is no established effective medical treatment and liver transplantation (LT) is the only life-extending therapeutic alternative for patients with end-stage PSC [1, 2].

Fifty to 80% of primary sclerosing cholangitis patients are also diagnosed with an inflammatory bowel disease (IBD), mainly ulcerative colitis. It has been reported in many cohort studies that PSC patients, and particularly those with associated IBD (PSC–IBD), have an increased risk of colorectal cancer (CRC).

Our knowledge on the mortality and morbidity rate associated with primary sclerosing cholangitis derives typically from single referral centre cohort studies that expose to a risk for selection bias. In contrast, population based studies on the natural history of PSC are sparse and, in most cases, rather ancient. As a consequence, the study from de Valle et al. based on a large and recent Swedish population cohort is of particular interest. In a previous high quality study, the authors investigated the prevalence and temporal trends in the incidence of PSC and identified 199 PSC patients in a well defined geographical region in southern Sweden between 1992 and 2005 [3]. Here, they bring further information by reporting on mortality and cancer risk in these patients which were followed until March, 31, 2008. The transplant free survival at 5 and 10 years was 81% and 70% respectively and the standardized mortality ratio (SMR) was four-fold increased in PSC patients compared with the general population (SMR 4.20; 95% CI 3.01–5.69). These findings are in keeping with our current knowledge of PSC and with previously published population based cohort studies [4]. Risk factors of liver related death or LT were age, female gender, jaundice, cholangitis and high bilirubin level. Except for female gender that has been previously reported as a pejorative prognosis factor in only one study [5], all these factors have already been pointed out in other studies. In this regard, it should be underscored that there is no consensus on the optimal prognostic model for PSC and that the use of prognostic models for predicting clinical outcomes in an individual patient is currently not recommended [6].

Hepatobiliary cancer, especially cholangiocarcinoma, is constantly reported as one of the major causes of mortality in PSC patients. This is also the case in the present study. Indeed, cholangiocarcinoma was responsible for 47.2% of all the observed deaths in PSC patients (17 of 36) and its cumulative 10 year incidence was 11%. In accordance with the largest previously published population based cohort study, the standardized incidence ratio (SIR) of hepatobiliary cancer was 177 (868 for cholangiocarcinoma!) and age was the only associated factor [7]. The unique good news in the hepatobiliarypancreatic area was that the previously suggested increase in risk of pancreatic cancer [7], was not confirmed.

More original and somewhat challenging are the findings on colorectal cancer. In many studies, the risk of CRC has been shown to be increased in PSC patients and notably in PSC–IBD. In these patients, CRC and dysplasia are most of the time located in the right colon [8-10] and occur at younger age than in IBD only patients. Although there was a trend for an increased risk of CRC in PSC patients in the study from de Valle et al., the difference with the risk of the general population was not statistically significant (SIR 2.87; 95% CI 0.33–10.4). Indeed, a diagnosis of CRC was made in only two patients and the cumulative incidence of CRC, although not indicated, was likely to be low. Even keeping in mind that dysplasia was not considered as part of the diagnosis of CRC, these results are in sharp contrast with numerous studies [7, 10-14]. In the experience of our tertiary centre, we showed that PSC–IBD patients had a higher risk of colorectal adenocarcinoma than IBD only patients and that PSC was an independent risk factor for the development of CRC (OR = 10.8; 95%CI, 3.7–31.3) [10]. In a recent long-term dutch study, it was found that PSC patients with concomitant IBD had a high risk of colorectal adenocarcinoma (14% and 31% at 10 and 20 years respectively) [12]. In addition, the proportion of colorectal cancer among other cancers was very low in the study by de Valle et al. (6.9%, 2 of 29 cancers) compared with the one observed in other studies (Claessen: 16 of 39, 41% [12]; Fevery: 10 of 40, 25% [15]).

How to explain such discrepancies and what are the clinical consequences? Several explanations can be proposed. Firstly, median follow-up (6.5 years) was shorter than in most other studies [12]. The risk of CRC persists after OLT (involving 25 patients of the study, 11.7%) and some studies suggest an even higher risk [14]. The fact that malignancies diagnosed more than 6 months after OLT were not included as incident cases might play a role as most of the published cases of CRC following OLT were diagnosed between 24 and 30 months after OLT [14]. Secondly, screening procedure in the studied population is not described and colon neoplasia might be underdiagnosed. Thirdly, although the actuarial rate was not available in the article, colectomy involved 24 patients (12.4%) of the cohort. A high rate of colectomy might be a partial explanation to a low risk of CRC. Finally, the recentness of the cohort (1992–2005) probably associated with an appropriate management of PSC patients (although no data regarding this point are provided) is one of the possible reasons. Indeed, most recent studies suggest that the risk of CRC in IBD patients is decreasing, probably because of better colonoscopic screening, wider use of chemoprevention and use of colectomy in UC therapeutic strategy [16].

Despite several limitations, the study by de Valle et al. brings an optimistic message suggesting that a high-quality management of PSC–IBD and more generally of IBD patients might lead to a significant decrease of the CRC risk. However, these findings are in conflict with a large number of other studies and should be confirmed before impacting on clinical practice. Thus, at present, recommendations regarding colorectal neoplasia screening in PSC–IBD patients (screening colonoscopy with biopsies at 1–2 years intervals from the diagnosis of PSC) should not be modified [6, 17].