Implementing Ammonia Measurement in Clinical Practice: Time to Forge Ahead

Even with the recent advancements for the treatment of hepatic encephalopathy (HE) over the last decade, HE still remains one of the most frequent and debilitating complications of cirrhosis. For several decades, it has been clearly established that the dysfunction of the cirrhotic liver to metabolise and detoxify ammonia leads to a substantial risk for hyperammonemia. Consequently, an influx of blood borne ammonia across the blood–brain barrier becomes neurotoxic, leading to deleterious effects on the brain [2]. For this, hyperammonemia has, for close to a century, been declared as the primary factor in the pathophysiology of HE [3, 4]. Since then, ammonia has remained a cornerstone of mechanistic understanding, even though its role in diagnosis, severity and prognosis has frequently been debated and the usefulness of measuring plasma ammonia in clinical practice is often questioned. In patients hospitalised with clinical symptoms of encephalopathy, it is recommended that an ammonia measurement confirm the diagnosis of HE since a normal ammonia value would suggest alternative pathogenic factors (associated with liver disease) or differential diagnoses (unrelated to liver disease) [5]. Recently, Deplats and colleagues elegantly demonstrated in close to 33% of hospitalised patients with cirrhosis that overt encephalopathy was associated with normal levels of ammonia (defined as non-HE) [6]. It was determined that this cohort of patients was exposed to significantly more psychotropic medications and had a higher occurrence of infections compared to patients with overt HE. This significant number of patients with non-HE clearly identifies the heterogeneity of pathogenic factors that cause encephalopathy in patients with liver disease and justifies measuring ammonia in patients with encephalopathy. Additionally, this important finding demonstrates that encephalopathy in patients with cirrhosis should not be a proxy for treatment with lactulose. Therefore, measuring ammonia might prevent lactulose overuse avoiding severe side effects, such as dehydration, electrolyte imbalance or perianal skin irritation due to diarrhoea. The relationship between ammonia levels and HE has forever been disputed. The study conducted by Ong and colleagues initiated this debate when the authors demonstrated a significant correlation between blood levels of ammonia and severity of HE (grades based on West-Haven criteria) [7]. However, the presented scatter plot graph clearly demonstrated patients with grades ranging from 1 to 4 had normal levels of blood ammonia. It is important to note that not all precipitating events of encephalopathy trigger an increase in blood ammonia. Interestingly, in the same study, infection represented 21% of the defined precipitating factors. It has been shown that inflammation amplifies the deleterious effects of hyperammonemia [8] however whether infection provokes hyperammonemia remains undefined. In addition, as mentioned above, medications (psychoactive drugs) can lead to encephalopathy Therefore, many factors can induce encephalopathy (liver-disease related and unrelated) and therefore including all patients with encephalopathy will provide an inexact conclusion on whether blood ammonia levels impact HE severity. The relationship between the severity of HE and blood ammonia levels is further weakened since outpatients with no clinical symptoms of HE present with above the upper limit of normal (ULN) levels of blood ammonia [9]. The fact that the presence of hyperammonaemia does not cause overt HE suggests inter-individual sensitivity to ammonia which can be a result of age, the presence of inflammation/oxidative stress and comorbidities [10]. As an example, this can be frequently seen in patients with elective TIPS insertion: although ammonia frequently exceeds the ULN not all patients develop post-TIPS HE [11]. However, intriguingly, increased blood ammonia (independent of the severity of HE) has emerged as a useful biomarker for prognostication. Within the last decade, 4 studies have been published demonstrating the important risk of death with high levels of hyperammonemia. Vierling and colleagues (178 patients) found that an elevated level of ammonia (> 1.5 × ULN) led to a significantly higher risk of death over the following 125 days [12]. Patwardhan et al. (494 patients) found that a cut-off of 60 μM of ammonia led to a higher risk of death over 90 days [13]. In a study with 498 patients, Shalimar et al., obtained results revealing an ammonia cut-off of 79.5 μM led to a significantly higher threat of death over 90 days [14]. Lastly, a study from the United Kingdom, found that a cut-off of 1.4 times the ULN of plasma ammonia caused significant death within a year [9]. Furthermore, in outpatients, blood ammonia levels affect the development of a first episode of overt HE. This testifies to the value of measuring ammonia as a prognostic marker and for purposes of differential diagnosis in patients, in both the outpatient setting and during hospitalisations. In the current issue of Liver International, Erminelli and colleagues present a well-conducted retrospective study addressing this important question [1]. The analysis included 104 patients with overt HE treated at the University of Padua, Italy. Ammonia levels were assessed at the discretion of the treating physician before, during, and after hospitalisation. The cohort was well balanced with respect to disease severity, comprising patients with OHE grade II (30%), grade III (66%), and grade IV (7%). Ammonia levels did increase with higher OHE grades; however, substantial overlap between groups was observed. Notably, OHE grade IV was associated with a longer length of hospital stay compared with lower grades. During hospitalisation, ammonia levels were measured at least once in the majority of patients. Consistent with clinical expectations, levels decreased under treatment compared to those obtained at admission. Interestingly, patients who did not fully recover from their OHE episode, as well as those receiving chronic psychoactive medication (e.g., benzodiazepines, antidepressants, or opioids), exhibited significantly lower admission ammonia levels. Consequently, the authors raise the hypothesis that in patients under treatment with the aforementioned medication(s), lower ammonia levels are needed to proxy the overt HE phenotype. The study represents a rigorous contribution to the field; nevertheless, its retrospective design limits causal inference. Prospective, multicentre validation in a controlled setting with measurements following a predefined protocol would therefore be highly desirable. The data reported by Erminelli et al. underscores that the pathophysiology of overt HE is not uniform across patients. In individuals receiving chronic psychoactive medication, comparatively lower ammonia concentrations appear sufficient to precipitate an overt HE phenotype, highlighting the increased vulnerability of this subgroup. These observations are consistent with the recent study by Deplats et al., in which 33% of the relevant subset of hospitalised patients with cirrhosis had non-hyperammonemia encephalopathy, characterised by low ammonia levels [6]. Identifying patients with non-hyperammonemia encephalopathy will improve management and response to treatment. In the phase 2b randomised controlled trial evaluating the efficacy and safety of the ammonia-lowering strategy ornithine phenylacetate for the treatment of overt HE, negative results were yielded due to the inadvertent inclusion of patients with encephalopathy unrelated to hyperammonemia [15]. From this perspective, it is not surprising that an ammonia-lowering strategy failed to demonstrate efficacy in patients whose encephalopathy was not driven by hyperammonemia. Following secondary analysis, it was demonstrated that a reduction in blood ammonia led to a faster improvement in HE severity. Taken together, these findings—along with the study by Erminelli et al. published in this issue—support the clinical relevance of ammonia measurement in hospitalised patients with cirrhosis to better define the diagnosis of HE with a need for ammonia lowering therapy. The data lay the foundation for planning a prospective multicentre trial to finally answer the question of whether measuring ammonia is necessary to guide treatment decisions. In addition to further understand the correlation between present levels of blood ammonia and severity of HE, studies are also required to investigate the relationship between improvements in HE severity and blood ammonia levels. The temporal resolution between normalising ammonia levels and improvement in HE might not be linear; therefore, there might be a lag (consequence and resolution) on the changes in ammonia and impact on the brain. C.F.R.: Consulting fees: Genfit. Research support: Genfit, Satellite Bio. C.L.: Lecture and consultant fees: Merz Therapeutics, Norgine, Alfasigma, Intercept, Gilead Sciences, Abbvie, Ipsen, Falk Foundation e.V., CSL Behring, Boehringer Ingelheim. Research grants: Merz Therapeutics, Norgine, Schwiete Foundation. This article is linked to Erminelli et al. paper. To view this article, visit https://doi.org/10.1111/liv.70365. Data sharing not applicable to this article as no datasets were generated or analysed during the current study.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.