Assessment of liver cirrhosis severity in 1015 patients of the Euricterus database with Campbell-Child, Pugh-Child and with ascites and ascites-nutritional state (ANS) related classifications. Euricterus Project Management Group.

Background/aims: The assessment of disease stage in cirrhosis is important for the individual patient (prognosis, timing and risk for requiring surgical intervention) and also for population comparisons and trials. There are several established methods, and we have aimed at comparison of the methods within a large cirrhosis population.

Methodology: In the European Union Euricterus database, there are 1015 patients with a "certain" diagnosis of cirrhosis, each of whom in one session had a protocol work-up of history, physical examination and all laboratory investigations needed for this study. The Child-Turcotte (CT), Campbell-Child (C) and Pugh-Child (P) classifications, as well as ascites/no ascites, ascites 1, 2, 3 (no, therapy responsive, nonresponsive) and ascites/nutritional state (ANS, 1-9) scores were used. CT and C have the same 5 variables, P has prothrombin time instead of nutritional state. CT, C and P variables score 1-3 each. C and P furthermore have variable range scores of 5-15. CT, C and P have classes A-C. The variables used were ascites, nutritional state, encephalopathy, bilirubin, albumin and prothrombin time.

Results: Only 53 patients (5%) fit within the CT criteria. C and P variable range scores (5-15) correlated strongly (r = 0.84). Cross-over calculation showed slightly different results in the P and C choice of variables, while the variable ranges (1-3) did not matter. Different selection of score ranges for the A-C classes in C and P resulted in 69% class C in P (35% in C) and 3% A in P (19% in C). The patients with ascites (70%) had worse bilirubin, albumin, nutritional states and C and P 5-15 scores (p < 0.0001). Patients with ascites 3 had all variables and also C, P 5-15 scores worse than those with ascites 2 (p < 0.02). ANS scoring showed wasting in 33% of the patients without ascites (ANS 3), 50% of the patients with ascites 2 (ANS 6) and 60% with ascites 3 (ANS 9) (p < 0.0003), and C and P scores were higher in the 3 ANS scores with wasting.

Conclusions: Campbell and Pugh 5-15 scores correlated closely and can be used interachangeably. As C does not contain the more elaborate prothrombin time determination, it probably can be used anywhere in the world. Ascites (degree) and Ascites/Nutritional State (ANS) scoring only use history and physical examination and are, or remain, although less refined, clinically relevant.