The CANONIC study has shown that among patients with acute decompensation (AD) of cirrhosis, those who had Acute-on-Chronic Liver Failure (ACLF, characterized by intense systemic inflammation, organ failures and poor outcome) were clearly distinct from those with traditional AD (who do not have organ failures and have good outcome).
The evidence-based definition of ACLF provided by this study introduced the notion of a common final pathway in the conceptual toolkit of hepatology, which expanded our understanding of the mechanisms of death in cirrhosis. However, besides the 'simple' dichotomy (ACLF vs. traditional AD), the CANONIC study also revealed the diversity existing among patients with AD, in particular with respect to the number of precipitating factors (which ranged from 0 to 2 or 3) and their nature (the commonest being acute bacterial infection, followed by excessive alcohol consumption and variceal hemorrhage). Moreover, a given precipitating factor -such as bacterial infection- has very different outcome (no sepsis, sepsis, severe sepsis, and septic shock) for reasons that currently remain elusive. Similarly, it is unclear why among patients with alcoholic hepatitis or variceal hemorrhage, some develop a very severe disease others don't. Patients' heterogeneity can also be related to differences in the treatment chosen among the available arsenal, including non-selective beta-blockers, curative antibiotics, prophylactic antibiotics, corticosteroids, albumin, vasopressors, renal-replacement therapy, MARS dialysis, mechanical ventilation. High-throughput '-omic' technologies permit the capture of human diversity with unprecedented precision. Leveraging these technologies in clinical decision making will help to bring about the long-heralded personalization of medicine (Nat Immunol. 2015;16:435-9). 'Omics' include genomics, transcriptomics, metabolomics, metagenomics, among others. The interest of using 'omics' in AD of cirrhosis is illustrated by the example of AD precipitated by bacterial infections. First some studies looking at the presence of pre-specified single-nucleotide polymorphisms (SNPs) in genes such as NOD2, TLR2, and FXR have shown that some SNPs were associated with an increased risk of developing spontaneous bacterial peritonitis (SBP) (reviewed in Semin Liver Dis 2016;36:133-40). These findings suggest that SNPs can contribute to the risk of infection in cirrhosis. However, studies published so far investigated a very limited number of SNPs and, obviously, this is one of their main weaknesses. The best way to address the diversity of infectious risk among patients with cirrhosis is to perform genome-wide association studies (GWASs), that take into account thousands of SNPs. Second, SNPs in the NOD2 genes were found to be associated not only with the risk of SBP but also with the risk of death (Hepatology 2010;51:1327-33), suggesting that these SNPs might contribute to severe inflammatory response and subsequent ACLF. Conversely, it has been shown that among patients with AD, some SNPs in the IL1 locus were associated with decreased systemic inflammation and low risk of ACLF (Hepatology 2017;65:202-16). These findings suggest that SNPs can contribute to the severity of infection by modulating, for example, the intensity of the systemic inflammatory response to pathogens. Here again, these studies had weaknesses because they investigated very few SNPs, highlighting the urgent need for GWASs assessing the risk of infection-induced developing excessive inflammation. Next, one should have in mind that SBP develops because bacteria translocate from the intestinal lumen to blood and then ascites. Changes in gut microbiome (i.e., metagenome) can be investigated by using metagenomics in patients with cirrhosis, in particular those at risk of developing SBP. Finally, infection is associated with increased levels of systemic and/or local metabolites that may reflect dramatic changes in gene induction (i.e., induction of metabolite-producing enzymes). Interestingly, some metabolites (endogenous or produced by microbes) can impact gene expression in immune cells and, therefore, contribute to the overall response to infection. These findings indicate that metabolomics would produce important information on metabolic changes associated with infection and, in this context, eventually differentiate patients with traditional AD from those with ACLF. Through different studies (CANONIC, PREDICT, ACLARA, ALADDIN), the Grifols Chair will use genomics, transcriptomics, metabolomics and metagenomics with the aim to address the diversity of patients with AD and, ultimately, develop personalized monitoring to uncover molecular networks that stratify patients with AD.