Introduction Recent follow-up cohort studies have identified biochemical parameters in UDCA-treated patients which identify “responders” with survival identical to normal control populations, in contrast to “non-responders” who have survival indistinguishable from that predicted by the Mayo Risk Score for untreated patients. The most widely accepted criteria for UDCA “response”, proposed by Corpechot, consist of alkaline phosphatase (ALP) <3× uln, alanine transaminase (ALT) <2× uln and bilirubin <1× uln. However, Bilirubin and ALP levels are independent predictors of outcome in PBC irrespective of UDCA treatment. ALT, in addition, is a biomarker for overlap syndrome, a prognostic feature itself.

Aim UDCA response criteria identify a group of patients with an inherently low risk rather than UDCA response per se.

Method In a comprehensive cohort of patients, geographically defined in 1999 followed up for 10 years, a group of patients (n=94) not treated with UDCA were identified. This cohort, matched prospectively to individual community case controls, represents an opportunity to study the natural history of non-UDCA treated PBC.

Results In the whole cohort (including UDCA treated patients), survival was significantly better in the UDCA-responding patients (defined using the Corpechot criteria) than in the non-UDCA-treated patients (p<0.05 log-rank test) although as in other studies survival was not the same as in age and sex match controls (p<0.05). Of the 94 non-UDCA receiving patients, 80 and 14 had, after 1 year of follow-up, biological features which, if they had received UDCA, would have been compatible with UDCA-response and UDCA-non response, respectively. Un-transplanted survival was significantly better in the non-UDCA patients meeting response criteria (58/80 (72%) at 10 years follow-up) than in non-UDCA treated patients meeting non-response criteria (5/14 (36%), p=0.01; χ² 9.4, p<0.01 log-rank test). The magnitude of this effect was similar to that associated with UDCA response in other series. The absence of UDCA treatment precludes, of course, this being a phenomenon of actual UDCA response.

Conclusion UDCA response criteria identify a group of patients who, at the time of criterion measurement, have a good prognosis. It does not appear to matter whether attainment of these parameters is a natural feature of untreated disease or a consequence of treatment. These criteria therefore have an important role in identifying high and low risk patients, and a sub-group of PBC patients with poor prognosis for whom additional therapies should be sought. They do not, however, provide specific information about the actions of UDCA.