IBD patients with PSC are more likely to have total colonic involvement than those without PSC

Reference: O'Toole, A., et al., Primary sclerosing cholangitis and disease distribution in inflammatory bowel disease. Clin Gastroenterol Hepatol, 2012. 10(4): p. 439-41.

Brief Summary: Primary sclerosing chogangitis (PSC) is a liver disease cuased by a scarring and inflamation of the the liver, and many of these patients also have IBD.  The exact relationship between intestinal site inflammation and and primary sclerosing cholangitis is not entirely clear. The main objective of this study was to determine the disease extent relationship for IBD patients with and without PSC.  The study was performed in two parts. The first part incluced 2754 IBD patients attenting a University Hospital between 1983 and 2011, and the second part included 82 PSC patients from the Irish National Liver Transplant Unity.  For the first part, IBD and PSC were diagnosed using standard protocals. 

Results: PSC was significantly associated with disease extent in both Ulcerative Colitis (UC) and Chron’s Disease (CD) patients.  Particuarly, two-thirds of UC patients with PSC had total colonic involvement wheras only 43 percent of UC patients without PSC had total colonic involvement.  Furthermore, males were more likely to have PSC than females. There was not found to be any signficant association with age.  

Implications for Practice: If a patient has PSC and IBD, there is a high liklihood of total colonic involvement vs. just left or rectal involvement. 

Discussion:  Good study here and I like the statistical design which I will get to in a bit.  It will be interesting to see follow-up studies on the the causal relationship between IBD and PSC, but little is known right now. As the authors noted, there is likely to be more than one causal pathway to PSC, since PSC can develop in the absense of IBD. 

It was interesting that males have a greater probability of contracting PSC than females, and I wonder why.  This would be another good follow up study.  This paper leaves more questions than answers, which is always a good thing!  

Commentary on Statistics and Study Design: The statistical and study design layout of the paper was good.   I like how the author looked at the relationship between several extraneous clinical and demographic variables in their association with PSC or no-PSC.  This is always an important thing to do, because it often leaves open new questions to look at, and helps fulfill the overall knowledge. 

I am a little confused, because the paper says the authors used a binary logistic regression model to generate the p-values in Table 3 but then in the table it-self, it says that various statistical measures were used such as the Mann-Whitney t-test.  Either way, the p-values would be roughly equivlant regardless of the method used, so it’s good either way.

A beer in the pub for our guys in Ireland on this.