Objective: The non-invasive diagnosis of liver fibrosis is based mainly on biochemical markers. The main aim was to validate whether the prothrombin index is an indirect marker of the severity of liver fibrosis.
Patients and methods: The predictive value of the prothrombin index for liver fibrosis was first assessed in 243 patients with chronic liver disease, then validated in 193 other patients with chronic liver disease. The reproducibility of measurement of the prothrombin index in different laboratories was evaluated in 82 other patients.
Results: In the first group, the prothrombin index was predicted accurately by serum hyaluronate (R(2)= 0.67 at the first step by multiple regression). The relationship between the prothrombin index and the area of fibrosis was not influenced significantly by non-fibrotic pathological lesions. The prothrombin index began to decrease when the Metavir fibrosis score was 2 versus 3 for albumin. In the second group, the prothrombin index and the histological fibrosis score were well correlated (r= -0.70, P< 10(-4)). Prothrombin index < or =80% or < or =70% diagnosed severe fibrosis or cirrhosis, respectively, and prothrombin index > or =105% or > or =100% excluded a diagnosis of severe fibrosis or cirrhosis, respectively, at the 95% probability level. The prothrombin indices measured in different laboratories were similar (78+/-18% v. 78+/-14%) and well correlated (r= 0.91, P< 10(-4)).
Conclusions: The prothrombin index was well correlated with pathological liver fibrosis score, had a high diagnostic accuracy for severe fibrosis or cirrhosis especially due to alcohol, and was not influenced by other pathological lesions. The prothrombin index was reproducible. Thus, the prothrombin index expressed as a percentage is an accurate, reproducible, inexpensive and easily available marker of severe liver fibrosis.