Introduction Pediatric liver disease PLD is a serious condit
Pediatric liver disease (PLD) is a serious condition causing repeated hospitalizations and long-term morbidity and has a potentially fatal outcome without curative or supportive treatment. The true incidence and prevalence are unknown, but an estimated 15,000 children are hospitalized for PLD in the US annually [1,2]. Based on the clinical pictures, PLD is essentially comprised of two main groups, i.e. acute liver failure  and chronic liver disease, although etiologies can overlap. Pediatric chronic liver disease has a more insidious course of varying degrees of Z-VAD-FMK and changes in the extracellular matrix often leading to fibrosis/cirrhosis .
Prognostication based on disease severity or etiology is difficult in PLD. Seemingly comparable children can present markedly different outcomes. Although some prognostic scoring systems appear promising, addressing prognosis in PLD remains challenging [, , ]. Prognostic markers range from general ones such as age, sex, and clinical appearance to clinical scoring systems, histology, functional liver tests, and biomarker analysis. Prognostication is especially important for PLD patients with a potential need for liver transplantation (LT). Liver transplantation can be lifesaving for PLD patients and should be reserved for patients who have lost their capacity for liver regeneration; however, a shortage of donor organs still exists . Prognostic markers enabling children requiring rapid and extensive medical attention possibly involving LT to be differentiated from patients with a favorable prognosis continue to be sought by physicians.
The exponential development of precision medicine and accordingly related molecular methods has shifted the focus of PLD research onto molecular biomarkers [9,10]. Biomarkers evaluated in liver biopsy (LB) have demonstrated the strongest diagnostic and prognostic values among various PLD diagnoses . Hence, despite being an invasive procedure, LB remains the gold standard for PLD diagnostics. Of note, while prognostic molecular liver tissue markers are well described in malignant PLD [, , ], there is only limited knowledge about markers in non-malignant PLD.
Methods The databases MEDLINE via PubMed and Embase via OVID were searched in August 2017. A systematic strategy based on the PICO criteria  was used. This approach included MeSH terms and text words for the PubMed search and Subject Heading and text words for the Embase search. Briefly, three main aspects for the search were defined: i) patient and diagnosis; ii) molecular marker; and iii) the prognostic outcome. For each of these aspects, the queries were MeSH term/Subject Heading search combined with a text word-based search combined with “OR”. All three aspects of the search were combined by “AND” to ensure identification of articles fulfilling all aspects. Studies published between August 2007 and August 2017 were included in the review. Applied search filters were “human studies”, “child 0–18 years”, “English or Danish language”. COVIDENCE systematic review software (Veritas Health Innovation, Melbourne, Australia. Available at www.covidence.org) was used for the abstract and full text screening process. All articles from the initial search were independently screened for study title and abstract by 2 authors (JN, OØ). 10,996 articles were excluded after title and abstract screening. Full text of remaining 74 articles (Table 1_online) was studied for inclusion or exclusion by 4 authors (JN, VB, MSK, OØ), and subsequently 23 articles fulfilled the inclusion criteria for the final review. An overview of the search is found in Fig. 1 and in the supplementary material (Method 1_online and Method 2_online). Decision concerning in- or exclusion were blinded between authors. Disagreements were solved by joint discussion of the article. Papers on molecular markers in liver tissue addressing the prognostic course of pediatric liver disease were included. Case reports were excluded and neoplasia, in vitro studies, polymorphism, germline mutation/variant, non-child, non-human, non-liver tissue were all exclusion criteria. Only studies identifying a significant up- or downregulation of marker(s) were included. Assessment of prognosis across PLD diagnoses is not a well-defined entity. Hence, overall survival, native liver survival, liver transplant free survival and jaundice free survival were all used to evaluate prognosis of liver disease. Moreover, hepatic fibrosis was used as prognostic surrogate [11,, , ], due to the proven risk of progression to cirrhosis dependent on fibrosis stage. In non-alcoholic fatty liver disease (NAFLD) and non-alcoholic steatohepatitis (NASH) inflammation was included as a prognostic surrogate since NASH is the inflammatory progression of NAFLD. Other disease specific scoring systems were excluded when evaluating prognosis.