Introduction Staging of liver fibrosis by liver biopsy informs prognosis and aids treatment decisions but is associated with morbidity. A recent publication validated a new non-invasive fibrosis index, P2/MS in Korean patients with chronic viral hepatitis.1 This model is novel in relying upon haematological parameters rather than the levels of serum proteins to predict the risk of cirrhosis.

We sought to validate the predictive value of P2/MS in identifying liver cirrhosis and fibrosis in a cohort of patients with chronic viral hepatitis in a UK hospital outpatient clinic.

Methods We used the full blood count of subjects with compensated chronic viral liver disease undergoing liver biopsy prospectively recruited into a study of serum markers of liver fibrosis to calculate P2/MS. The ability of P2/MS to predict fibrosis and cirrhosis was compared to other validated models using the χ² test for equality of area under the ROC curve.

Results Results were available for 181 patients, of whom 145 had HCV and 36 HBV. The median age was 46 years, 62% were male and 52% Caucasian. Significant fibrosis (Ishak stage >2) was present in 81 (45%), of whom 35 had cirrhosis (19%). P2/MS was superior to AST/ALT ratio at predicting fibrosis and cirrhosis. It was inferior to hyaluronic acid in the diagnosis of cirrhosis and to APRI in detecting bridging fibrosis (Abstract 045). P2/MS diagnosed cirrhosis with a sensitivity of 57% (P2/MS<45) and excluded bridging fibrosis with a specificity of 33% (P2/MS>115). When the individual components of the P2/MS score were considered in isolation, liver fibrosis was not associated with monocyte or neutrophil fraction. The predictive ability of the P2/MS was due to the reduction in platelets in advanced liver disease.

Conclusion P2/MS has a modest ability to detect bridging fibrosis and cirrhosis in a cohort of patients with chronic viral liver disease in a UK clinic. We have not validated the association between altered leucocyte populations and liver fibrosis. This may be secondary to ethnic differences, or to the lower incidence of cirrhosis in our population relative to the index study. The utility of P2/MS in different populations should be investigated in further studies.