Introduction Hepatocellular carcinoma (HCC) is a rare complication in primary biliary cirrhosis (PBC). Current clinical guidance from the British Society of Gastroenterology advises that only male patients with histologically advanced PBC are at sufficient risk of HCC to justify enrolment in surveillance programs. In this study, we review our experience of HCC in PBC and investigate the utility of serum (s-) bilirubin as a selection criterion in determining enrolment in an HCC surveillance program.
Method Patients with HCC and PBC managed in our centre were identified from the Hepatobiliary Cancer, Liver Transplant and Liver Histopathology Databases. Demographic, clinical, biochemical and outcome data were obtained by retrospective review of the patient's case notes.
Results 32 cases of HCC in PBC patients from 1999 to 2010 were identified, including 1 with concomitant alcohol-related liver disease and one patient with an autoimmune hepatitis–PBC overlap syndrome. No patient had documented hepatitis b surface or hepatitis c antibody positivity. 81% of patients were female. 92% of patients had received ursodeoxycholic acid (UDCA). Liver biopsy result was available in 15 patients, with 7 having stage 3–4 PBC. Median age at diagnosis of HCC was 66 years (range: 42–86) and median time from diagnosis of PBC to development of HCC was 13.5 years (4.3–16). Identification of HCC was made during surveillance in 16 and assessment for liver transplantation in four patients, following hepatic decompensation in four, and incidentally in five, including in the liver explant of three patients. Median α-fetoprotein at diagnosis was 28 kIU/l (IQR: 5–204), and was elevated (>20 kIU/l) in 15 patients (50%). At diagnosis of HCC, median Mayo Risk Score, MELD, MELD-Na, UKELD and Child–Pugh scores were 6.87 (6.08–8.09), 10 (8–14), 14 (10–19), 50 (48–55) and 7, respectively, with no difference between survivors and non-survivors (p=NS) or males and females (p=NS). Median s-bilirubin was 27 μmol/l (14–54), but in 13 (41%) patients the s-bilirubin was normal. 89% (25/28) patients had evidence of portal hypertension defined by the presence of varices, splenomegaly, or ascites; including nine patients with normal s-bilirubin. There was no correlation between the presence of jaundice and outcomes or the presence of portal hypertension (p=NS). However, those patients diagnosed with HCC during surveillance were less likely to be jaundiced (31%), when compared with those with symptomatic presentations (75%).
Conclusion In our patient cohort, s-bilirubin is not an appropriate indicator of HCC risk in PBC, as most PBC patients were not jaundiced when diagnosed with HCC by surveillance. Therefore, a normal s-bilirubin should not be used to exclude patients from HCC surveillance. Further we propose that surveillance should not be limited to male patients. We hypothesise that UDCA is modifying s-bilirubin levels in PBC without altering portal hypertension or HCC risk.