Introduction: In children presenting with jaundice, presence of underlying chronic liver disease (CLD) has prognostic implications. Hence at the time of presentation, identifying those with underlying CLD is important. We studied children presenting with acute on chronic liver disease (ACLD) and compared their clinical and biochemical profile with those of acute viral hepatitis (AVH).

Method: Children presenting with jaundice of <6 mo duration were studied. ACLD was diagnosed by positive test for viral markers (IgM HAV or HEV or HBc) in patients with established CLD. AVH was diagnosed on the basis of positive test for viral markers (IgM HAV or HEV or HBc). We compared clinical and biochemical features of ACLD with those of AVH using Fisher's exact test, Mann Whitney U test and did logistic regression analysis to identify variables predicting ACLD or AVH. Children with fulminant hepatic failure were excluded.

Results: Between June 1998 and June 2005, of the 151 children, 21 (14%) and 130 (86%) were diagnosed to have ACLD and AVH respectively. We found statistically significant differences between the two groups with respect to duration of symptoms (7.8 Vs 2.5 weeks), frequency of prodrome (57% Vs 81%) and frequency of ascites (57% Vs 16%) and, biochemical values viz. S.bilirubin (13.8 Vs 8.1 mg/dL), PT (27.7 Vs 15 Sec), AST (486 Vs 1145 IU/L), ALT (548 Vs 1395 IU/L), AST/ALT ratio (1.5 Vs 0.9) and S.albumin, in ACLD and AVH respectively. In terms of frequency, more number of children in ACLD had PT prolongation> 6 sec above control (76.2% Vs 11.5% in AVH) and serum albumin values <3 g/dL (66.7% Vs 28.7%).

We did logistic regression analysis including clinical and biochemical variables. Only duration of illness and prothrombin time were found to be significant on univariate analysis and subsequent multivariate analysis showed these to be independent predictors. Furthermore, increment of 10 sec in PT, increases likelihood of ACLD by 2.2 times (95% CI 0.277-0.879) and increment of every week in duration of symptom increases the likelihood of ACLD by 1.2 times (95% CI 0.68-0.89). Prediction model incorporating duration of illness and PT is correct 88% of times. At a cut off> 6 sec above control for PT, prediction of likelihood of ACLD was 76.2% times correct and for AVH, 88.5 times correct (overall correct prediction for the model using both duration and PT 88.8%).

Conclusion: In children with jaundice of <6 mo duration, ACLD can be correctly predicted 88% of times with increased duration of disease and prolonged PT.

Pedgastro 2005 - Conference Abstracts.Pediatric Oncall [serial online] 2006 [cited 15 March 2006(Supplement 3)];3. Available from:
http://www.pediatriconcall.com/fordoctor/Conference_abstracts/
jaundice.asp