Prevention of further hepatic decompensation has long been approached by ‘symptomatic’ measures focused on mitigating clinical consequences of advanced chronic liver disease (ACLD), acting downstream of a pathophysiological cascade [1]: Despite effective hypovolemia being a hallmark of ascites formation, diuretics and large-volume paracentesis are used to prevent or treat the recurrence of ascites, while varices are obliterated to treat the bleeding episode or prevent rebleeding.

At the same time, we have learned that there are two key disease-driving mechanisms underlying both first and further hepatic decompensation and subsequent mortality: Portal hypertension (PH) and systemic inflammation (SI, Fig. 1). While both coincide and cannot strictly be separated as there is a bidirectional relationship [1,2], their prognostic utility differs across disease stages [3]. While clinically significant PH (CSPH) is considered the predominant predisposing factor for first hepatic decompensation, SI aggravates at later stages (i.e., when CSPH is ubiquitous), in particular in those with ascites, driving the development of acute-on-chronic liver failure and death [4,5].

PH has long been broadly accepted as a therapeutic target to modify the course of disease [6]. More recently, the modulation of SI has emerged as a promising approach, which is based on evidence that amelioration of SI is linked with a better prognosis [7].

Significant PH improvement is commonly achieved by removal/suppression of the primary etiological factor [8-10]. In contrast, the evidence supporting an improvement in SI is comparatively limited (e.g., in abstinent vs. non-abstinent patients) [11], although an improvement of SI seems highly likely. Non-selective betablockers (NSBB) have been shown to impact not only PH, but also SI [12,13]. However, they are limited by a low potency and possible side effects in advanced disease (i.e., therapeutic window) [14]. Poorly/non-absorbable antibiotics (i.e., norfloxacin/rifaximin) have shown some signals towards an improvement of PH and SI [15]. While albumin does not affect PH and mainly improves circulatory dysfunction in patients with ascites, it has several non-oncotic/anti-inflammatory properties and has been associated with improved markers of SI [16,17].

Finally, TIPS is the most effective treatment to reduce PH, and has shown survival benefit in subgroups of patients with ascites and variceal bleeding [18-20] with numbers needed to treat of as low as 2.5-5 in well-selected patients [19,21]. While the convincing survival benefit in specific patient groups has primarily been attributed to the profound PH-lowering effect, signals of improved bacterial translocation [22] and markers of SI [23] have previously been observed and have recently been significantly expanded:

Tiede et al. [24] prospectively studied 43 soluble inflammatory markers (SIM) and sCD14/sCD163 as markers of bacterial translocation in 59 patients undergoing TIPS placement, 78% of whom had refractory ascites (RA) and 22% with a history of variceal bleeding as TIPS indication (i.e., excluding patients with preemptive or rescue TIPS). Several aspects appear to be particularly insightful regarding our understanding of SI and, provocatively speaking, TIPS as an ‘anti-inflammatory’ treatment strategy:

First, 30/43 SIM were upregulated in decompensated cirrhosis compared to healthy controls, highlighting the proinflammatory state. Interestingly, they did not differ based on portal pressure gradient (PPG) before TIPS, suggesting that while SI is associated with PH, it seems to be uncoupled in advanced disease stages, especially in RA. Interestingly, RA was characterized by significantly higher levels of IL-6 at comparable PPG, suggesting a more pronounced SI in these patients. Nevertheless, the power to detect differences in these 59 patients after correcting for multiple testing might be an issue and will stimulate further studies on this topic. At the same time, IL-6 levels, hepatocyte growth factor, and IL-12 were significantly higher in the portal vein as compared to hepatic vein, indicating a likely splanchnic origin of these cytokines.

Next, the authors studied changes of SI over 12 months after TIPS. While little change in SI was observed after the first month, the majority of SIM (including IL-6 and sCD14) subsequently decreased at 3, 6, and 12 months after TIPS, suggesting a sustained improvement in SI and bacterial translocation after an initial consolidation. Importantly, these changes at 6 months did not correlate with the reduction in PPG measured at the TIPS procedure, again suggesting that improvements in SI are independent of PH improvements. As a note of caution, this sample only included patients surviving and/or being followed for 12 months after TIPS, introducing survivorship bias. Linking baseline patterns or dynamics within the first months with subsequent outcomes will be interesting to understand the prognostic implications of these changes in SI. Nevertheless, IL-6 and CRP levels were associated with persistent severe ascites up to 2 years after TIPS in those followed-up, suggesting that failure to resolve SI might coincidence with a failure to control ascites by TIPS, independently of PH improvement. From another perspective, we would like to hypothesize that inconsistent findings regarding the impact of TIPS on survival, with favorable outcomes in patients with better liver function and less advanced disease [20], could be due to an ‘uncoupling’ of SI from PH in very advanced patients in whom TIPS placement reverses PH, but cannot reverse SI.

Even more recently, these data on SI after TIPS were complemented by an analysis of 109 prospectively studied patients from both Hannover and Vienna [25] which evaluated IL-6, CRP, and lipopolysaccharide-binding protein (LBP) before TIPS and at 3, 6, and 9 months. Again, SI was significantly decreased already after 3 months, remaining low thereafter. Reassuringly, relative changes in IL-6 did neither correlate with PH severity before TIPS nor relative change in PPG. Notably, PPG dynamics in both studies were measured directly at TIPS placement, which might not be fully reflective of the PPG at 1 or 3 months according to recent data [26]. Additional analyses on chronic hemodynamic changes will be required to further substantiate these findings.

In the latter study, changes in IL-6 correlated with changes in LBP, a well-established marker of bacterial translocation, which together with previously reported decrease in sCD14 indeed indicates an improved intestinal barrier function following reduction of PPG. Most importantly, decreases in IL-6 at 3 months were independently predictive of a decreased risk of subsequent ACLF and liver-related death while adjusting for age, serum albumin, and MELD. In light of the previously discussed data by Tiede et al. [24] on failure to control ascites, this could suggest that a prolonged proinflammatory state after TIPS might be an early prognostic marker of worse clinical outcomes and could identify patients who need particularly close monitoring. In this regard, identifying those specific patients at a high likelihood of persistent SI, or those in whom the disease-modifying properties and clinical benefit of TIPS are expected to be less pronounced already prior to TIPS placement, would be of high interest, given that other studies have highlighted a poor prognosis in patients with pronounced inflammation/immune activation at the time of TIPS placement [23,27-29]. Future studies will also have to investigate how the above-mentioned treatment strategies to modulate SI could be utilized to amend the clinical course of patients in whom SI does not improve after TIPS.

In summary, these emerging data on improvement of SI after TIPS as well as its prognostic significance are indeed stimulating and will clearly prompt further studies to explore the underlying mechanisms of (non-)improvement in SI and to exploit their prognostic potential.