- 1.Potential role of sarcopenia: We did not quantify severity of sarcopenia but we did perform a single-centre sub-analysis of clinically validated metrics of sarcopenia and malnutrition that included estimated dry BMI, handgrip strength, median arm circumference and triceps skinfold thickness in addition to the global nutrition score and saw no association between sarcopenia and either AMM-ULN levels or the development of liver-related complications.
- 2.Degree of spontaneous portosystemic shunts (SPSS): This was not measured in this study. Most patients included in our study (94%) had indirect evidence of clinically significant portal hypertension as measured by either a history of oesophago-gastric varices, use of non-selective beta blockers or splenomegaly. We would agree that SPSS are pathophysiologically relevant and that the mechanism of occurrence of hepatic encephalopathy (HE) in this setting is driven by hyperammonaemia. Thus, AMM-ULN measurement here may act as a biochemical surrogate of the severity of portal hypertension and portosystemic shunting. This association between ammonia levels and the degree of shunting has been explored in previous modelling studies of ammonia metabolism.
- 3.The presence or absence of previous episodes and role of psychometric tests: Although previous decompensation was a risk factor for the prediction of liver-related complications in a univariable analysis, and it was included in the best model in multivariable competing risk analyses, it did not reach statistical significance when incorporating other variables such as AMM-ULN. The role of psychometric tests in the occurrence of HE and mortality is an important question and is the subject of an ongoing study of the AMMON Consortium.
- 4.Reporting crude ammonia values: Both absolute values of ammonia and AMM-ULN values were reported in our paper. AMM-ULN performed better than crude ammonia levels in predicting both complications and mortality. Therefore, we propose that using AMM-ULN may harmonise the different absolute ammonia levels being reported in the literature.
- 5.Predictive ability of AMM-ULN compared to traditional severity scores: Despite no significant differences in mortality, AMM-ULN demonstrated improved predictive performance for the development of liver-related complications when compared against the MELD score both at 6 months and 1 year and showed a tendency to be higher than the Child-Pugh score. These data are described in the paper.
Conflicts of interest
- Multimedia component 1
- Plasma ammonia levels predict hospitalisation with liver-related complications and mortality in clinically stable outpatients with cirrhosis.J Hepatol. 2022; (online ahead of print)
- Ammonia and prognosis of cirrhosis: a new perspective for identifying high risk patients.J Hepatol. 2022; (online ahead of print)
- Arterial ammonia levels in cirrhosis are determined by systemic and hepatic hemodynamics, and by organ function: a quantitative modelling study.Liver Int. 2014 Jul; 34: e45-e55
Publication stageIn Press Journal Pre-Proof
Author names in bold designate shared co-first authorship