Abstract
Background
Ammonia is implicated in hepatic encephalopathy (HE) and prognostic in cirrhosis. Venous ammonia concentration, yielding similar correlation with HE grades as arterial, has become the preferred practise but comparative data are limited.
Aim
To quantify effect of sampling site on ammonia concentration in healthy persons and patients with cirrhosis.
Methods
Ammonia concentrations were measured by arterial and femoral venous blood sampling in ten healthy men and ten male patients with cirrhosis before and during hyperammonaemia induced by ammonia infusion. Cubital vein samples were included during the infusion.
Results
At baseline, arterial-venous concentration gaps were similar (p = 0.15) in healthy persons [14 (10–19) and 8 (4–12) µmol/L] and patients with cirrhosis [53 (32–74) and 40 (23–57) µmol/L]. Ammonia infusion increased arterial-venous concentration gaps in both groups [115 (97–133) and 61 (31–90) vs. 175 (123–227) and 134 (65–203) µmol/L]. Mean ammonia concentration difference between groups during hyperammonaemia was 72 (42–103) µmol/L (p < 0.001) and independent of sampling site. Cubital and femoral vein concentrations were comparable (p = 0.26). In cirrhosis, calculated upper limit normal values (ULN) were comparable for arterial and venous blood at baseline [2.0 (1.2–2.8) and 2.1 (1.2–3.0), p = 0.74] and during hyperammonaemia [6.7 (4.7–8.7) and 6.2 (4.4– 8.1), p = 0.44].
Conclusions
We found clinically meaningful intra-individual arterial-venous concentration gaps in both healthy persons and patients with cirrhosis at any ammonia concentration. Inter-group concentration differences after induced hyperammonaemia were relatively constant across sampling sites which supports clinical use of venous sampling. ULN-normalised ammonia concentrations were valid for both arterial and venous sampling.
Keywords: Ammonia, Sampling site, Arterial-venous concentration differences, Cirrhosis, Upper limit normal
Introduction
Ammonia, being neurotoxic, is implicated in the development of hepatic encephalopathy (HE). HE is a challenging diagnosis due to its nonspecific clinical features – and measurement of blood ammonia is especially relevant in ruling out HE owing to its high negative predictive value (European Association for the Study of the Liver 2022). Though elevated ammonia levels correlate with the severity of HE, the correlation is only moderate with substantial overlap across the various grades of HE (Ong et al. 2003; Nicolao et al. 2003). Moreover, ammonia measurements holds specific risk of errors related to blood collection on ice and requires rapid processing and analysis (Seligson and Hirahara 1957; Bajaj et al. 2020).
Therefore, the clinical utility of ammonia has been questioned (Bajaj et al. 2020; Haj and Rockey 2020; Tapper and Rahimi 2020). However, mounting evidence points to the prognostic potential of ammonia in liver disease with regard to risk of developing HE, decompensation and mortality in both in- and outpatients with liver cirrhosis (Vierling et al. 2016; Shalimar et al. 2019; Tranah et al. 2022; Balcar et al. 2023; Ballester et al. 2023). Thus, ammonia measurement is still relevant for both clinical and research purposes.
One aspect of the difficulties of ammonia measurements is the impact of blood sampling site. Inter-sample variation between arterial and venous sampling sites was studied by Stahl in the 1950s and 1960s (Stahl 1963). Based on these studies and from a theoretic point of view regarding arterial-delivered ammonia to the brain, arterial analysis was long considered gold standard (Stahl 1963). By arterial sampling, one also prevented falsely elevated ammonia levels related to the tourniquet of venous sampling. Since more recent studies examining correlations between ammonia levels and grades of HE showed similar results using arterial and venous sampling (Ong et al. 2003; Nicolao et al. 2003), eventually clinicians turned towards the easier obtainable venous sampling. Still, robust data regarding the effect of sampling site on ammonia concentration are scarce. Also, it remains unclear whether the possible effect of sampling site is comparable in healthy individuals and patients with cirrhosis and in conditions with low and high ammonia levels; not least given that cirrhosis associated sarcopenia could potentially decrease arterial-venous ammonia concentration gaps (Lattanzi et al. 2019).
Accordingly, the purpose of this study was to quantify the effect on ammonia levels of different sampling sites in healthy persons and patients with cirrhosis before and during induced hyperammonaemia. We, therefore, measured ammonia concentrations in arterial and femoral venous blood before and after infusion of ammonia as well as cubital veinous blood during the ammonia infusion.
Materials and methods
Study participants
Data obtained from a previous study using a constant ammonia infusion were available from ten healthy men and ten male patients with alcohol-related cirrhosis (Eriksen et al. 2023). The study included outpatients excluding patients with Child-Pugh score > 12, overt HE or more than one admission with HE within the past year. Informed consent was obtained from participants included in the study.
Experimental set-up, blood sampling and ammonia analysis
At resting and fasting conditions, controlled hyperammonaemia was induced by a constant intravenous infusion of 0.25 mmol/kg/h ammonia chloride. At baseline and at steady state hyperammonaemia, 80 min after infusion start, blood was drawn synchronously from the radial artery and the femoral vein. In addition, cubital vein blood was drawn during hyperammonaemia, at t = 80 min.
Radial arterial and femoral venous blood was obtained by installed catheters and cubital venous blood by venepuncture with special care given to obtain blood stasis free after release of the tourniquet. By use of gas-tight adapters, after waste blood was drawn, blood was sampled and cooled directly on ice in di-potassium EDTA tubes. Plasma was immediately recovered after cooled centrifuging and stored at -40 °C until batch analysis the following day by enzymatic photometry (Siemens Healthcare, Erlangen, Germany; analytical variability for ammonia determination of 10–16% (2CV, twice the coefficient of variation). As previous described ammonia concentrations were adjusted for possible ammonia concentration drift due to plasma deamination processes after batch freezing and thawing (Eriksen et al. 2023).
Statistics
Data were analysed using Stata version 14.2 (StataCorp, CollegeStation, TX) and presented as means with 95% CIs. Student’s paired and unpaired t test were used for comparisons within and between the groups, respectively. For skewed data, non-parametric Wilcoxon-Mann-Whitney test was used for comparisons between groups. The effects of sampling site and cirrhosis on ammonia concentration during induced hyperammonaemia were investigated using a multivariate linear mixed model. Equality of standard deviations and correlations in the model was considered as appropriate. Model validation was performed by comparing residuals and fitted values and by using QQ plots. The significance level was set at 0.05 in a two-tailed test. Normalised ammonia upper limit of normal (ULN) values for the respective sampling sites were calculated as the mean + 1.96 × SD of the measured value in healthy persons before ammonia infusion.
Results
Basic characteristics of the study participants are presented in Table 1. The patients were older, had lower muscle mass and biochemical blood tests reflective of their liver disease and portal hypertension.
Table 1.
Clinical and biochemical characteristics
| Healthy persons (n = 10) |
Patients with cirrhosis (n = 10) |
p | |
|---|---|---|---|
| Age (years) | 39 (22–66) | 58 (43–73) | < 0.001 |
| Body weight (kg) | 85 (79–91) | 77 (67–87) | 0.17 |
| Total muscle mass (kg) | 31 (29–34) | 26 (21–31) | 0.02 |
| Child-Pugh score# | - | 7 (6–9) | - |
| MELD-Na score | - | 12 (10–14) | - |
| Ascites (0/1/2/3) | - | 8/0/2/0 | - |
| Biochemistry | |||
| Alanine aminotransferase (U/L) | 20 (14–26) | 33 (23–43) | 0.02 |
| Bilirubin (µmol/L) | 12 (8–17) | 23 (15–31) | 0.02 |
| Albumin (g/L) | 40 (38–42) | 28 (23–34) | < 0.001 |
| Creatinine (µmol/L) | 75 (67–84) | 66 (55–77) | 0.16 |
| Haemoglobin (mmol/L) | 8.5 (8.3–8.7) | 7.1 (6.0–8.2) | 0.02 |
| Thrombocytes (x109/L) | 236 (195–277) | 87 (58–116) | < 0.001 |
| INR# | 1.1 (1.0–1.1) | 1.4 (1.2–2.2) | 0.003 |
MELD-Na, model for end-stage liver disease-sodium; INR, international normalised ratio. Values are given as means with their 95% CIs.#Given as median (range)
Baseline arterial ammonia was lower in healthy persons as compared with the patients with cirrhosis (14 [10–19] µmol/L vs. 53 [32–74] µmol/L, p < 0.001, Table 2). Femoral vein concentrations were significantly lower in both groups (p ≤ 0.03) [8 (4–12) and 40 (23–57) µmol/L, respectively, < 0.001). The arterial-venous ammonia concentration gap did not differ significantly between the groups (p = 0.15). During ammonia infusion at t = 80 min, ammonia concentrations in the radial artery and the cubital and femoral veins were 115 (97–133), 61 (31–90), 47 (32–59) µmol/L, respectively, in healthy persons and 175 (123–227), 134 (65–203), 120 (84–155) µmol/L, respectively, in the patients with cirrhosis (Table 2).
Table 2.
Ammonia concentrations before and after induced hyperammonaemia
| Healthy persons (n = 10) |
Patients with cirrhosis (n = 10) |
ULN values (for patients with cirrhosis) |
|
|---|---|---|---|
| Arterial, baseline | 14 (10–19) | 53 (32–74) | 2.0 (1.2–2.8) |
| Femoral venous, baseline | 8 (4–12) | 40 (23–57) | 2.1 (1.2–3.0) |
| Arterial, infusion | 115 (97–133) | 175 (123–227) | 6.7 (4.7–8.7) |
| Femoral venous, infusion | 47 (32–59) | 120 (84–155) | 6.2 (4.4– 8.1) |
| Cubital venous, infusion | 61 (31–90) | 134 (65–203) | - |
Mean ammonia concentrations for healthy persons and patients with cirrhosis (µmol/L) and calculated upper limit normal (ULN) values for patients with cirrhosis with their respective 95% CIs
Data collected after 80 min of ammonia infusion were analysed by multivariate model testing showing no significant interaction between the effects of sampling site and cirrhosis (p = 0.75). Thus, the effect of sampling site were similar for healthy persons and cirrhosis patients during induced hyperammonaemia. Therefore, an additive model was chosen accepting parallel slopes between the two groups (Fig. 1). The average ammonia concentration difference between healthy persons and patients with cirrhosis across all three sampling sites was 72 (42–103) µmol/L p < 0.001. Table 3 presents ammonia concentration gaps between the radial artery, the cubital vein and the femoral vein. During induced hyperammonaemia, ammonia concentrations were non-significantly higher in the cubital vein as compared with the femoral vein (10 [-8–29] µmol/L, p = 0.26). In both groups, arterial concentrations were ~ 50 µmol/L (p < 0.001) higher than venous concentrations (Table 3).
Fig. 1.
Interaction plots displaying the effect of sampling site (red: radial artery, blue: cubital vein, green: femoral vein) on ammonia concentration (μmol/L) during hyperammonaemia in healthy persons (white) and patients with cirrhosis (black). The additive multivariate model (larger circles) and the measured ammonia concentrations (smaller circles) are shown. Y-axis increments increases at y = 200. Mean ammonia concentration difference (μmol/L) between groups: 72 (42–103) (p < 0.001) and between sampling sites: a. radialis vs. v. cubitalis 50 (37–64) (p < 0.001), a. radialis vs. v. femoralis 61 (46–76) (p < 0.001), v. cubitalis vs. v. femoralis 10 (-8–29) (p = 0.26)
Table 3.
Ammonia concentration differences between radial artery, cubital vein and femoral vein during induced hyperammonaemia
| Sampling site comparison | Δ [Ammonia] | p |
|---|---|---|
| a. radialis vs. v. cubitalis | 50 (37–64) | < 0.001 |
| a. radialis vs. v. femoralis | 61 (46–76) | < 0.001 |
| v. cubitalis vs. v. femoralis | 10 (-8–29) | 0.26 |
Mean ammonia concentration differences with their 95% CIs (Δ, µmol/L)
In the patients with cirrhosis, ammonia concentrations normalised to ULN yielded comparable values for arterial and venous blood at baseline [2.0 (1.2–2.8) and 2.1 (1.2–3.0) times ULN, respectively, p = 0.74] and during induced hyperammonaemia [6.7 (4.7–8.7) and 6.2 (4.4–8.1) times ULN, respectively, p = 0.44] (Table 2).
Discussion
In this study we examined the impact of sampling site on ammonia concentrations before and after hyperammonaemia was induced by ammonia infusion. In both groups arterial ammonia was higher than femoral venous ammonia and this difference increased 3–4 fold during induced hyperammonaemia. Ammonia concentration differences between healthy individuals and patients with cirrhosis were comparable across the different sampling sites.
Higher arterial than venous ammonia concentrations are in accordance with a priori expectations as the resting skeletal muscle mass is a net ammonia-consuming organ (Huizenga et al. 1996; Lockwood 2004; Dam et al. 2011). Our observation on arteriovenous concentration gaps across the leg at basal conditions in healthy individuals align with reports from previous studies ranging from 5 to 14 µmol/L (Warter et al. 1983; Eriksson et al. 1985; Dam et al. 2011) as compared to 6 µmol/L in our study. As regards patients with cirrhosis, comparison between studies, not least between a controlled experimental set-up like ours and spontaneous ammonia concentration measurements from the clinic including patients with hepatic encephalopathy, is problematic. Nevertheless, similar relative arteriovenous concentration gaps for cirrhosis patients have been reported (Tyor et al. 1960; Stahl 1963; Ong et al. 2003; Nicolao et al. 2003; Dam et al. 2011; Manjunath et al. 2014); ranging from 10 to 35% as compared to 23% at baseline and 30% during hyperammonaemia in our study.
During induced hyperammonaemia we found clinically relevant differences between arterial and venous sampling sites of ~ 50 µmol/L for both healthy persons and cirrhosis patients, highlighting the sampling site for ammonia as being important when assessing hyperammonaemia. The small difference between ammonia concentration in cubital and femoral blood most likely reflect the fact that blood flow is proportional to muscle mass which thereby gives rise to similar fractional ammonia removal across the respective muscle group (arm vs. leg). Accordingly, depending on the clinical setting, different venous sampling sites may be used interchangeably for ammonia concentration measurement.
Of importance, during induced hyperammonaemia the ammonia concentration differences between healthy persons and patients with cirrhosis were tolerably constant across the three sampling sites in our analysis. Accordingly, discrimination between healthy persons and patients with cirrhosis was valid and quantitatively unaffected by venous sampling. The observation that arteriovenous differences during ammonia infusion were similar in healthy persons and patients with cirrhosis although the ammonia levels were higher in cirrhosis is interesting and needs further exploration that will also include measurement of muscle blood flow. We speculate that patients with cirrhosis have lower muscle mass and lower resting muscle blood flow resulting in reduced muscle ammonia removal but further explorations are needed.
Being a key pathological feature of liver disease, the primary driver of hepatic encephalopathy and prognostic in cirrhosis, we believe that ammonia measurement is still valuable even though the diagnostic specificity in hepatic encephalopathy is limited. Optimal and standardised sampling conditions and sampling site specific normal values is key. Our results indicates that the use of ULN values, as have been proposed (Tranah et al. 2022; Balcar et al. 2023; Ballester et al. 2023), is valid for both arterial and venous sampling sites, given that arterial- and venous- specific ULN values are used.
A major strength of our study is that we measured arterial and venous ammonia concentration gaps by synchronous sampling in each participant during standardised metabolic conditions during a controlled experimentally-induced hyperammonaemic steady state. Even though such hyperammonaemia is unphysiological, its impact on ammonia concentration in arterial versus venous sampling sites will likely mimic the clinical situation where systemic ammonia is increased in the patient with liver disease. Also, the careful sampling procedure limited preanalytical and analytical factors so our conclusions are not confounded this way. One limitation of our study is the small sample size that introduces the risk of type II error, e.g. regarding the importance of the numerically small effect of cubital vs. femoral sampling. Moreover, the study was not powered to explore the mechanistic explanation for the ammonia concentration differences observed between groups and sampling sites. Accordingly, a proper exploration of e.g., a possible correlation to liver dysfunction or muscle mass within the group of cirrhosis patients was not possible. We emphasise the importance of our healthy persons and controls being in the resting state; during physical exercise, ammonia production in the working muscle increases the ammonia concentration in venous blood draining these muscles and reduce arteriovenous concentration gaps (Eriksson et al. 1985). Thus, our results cannot be generalised to studies during exercise. Also, prolonged hyperammonaemia as observed in patients with hepatic encephalopathy due to e.g. a gastrointestinal bleed might exhaust muscle ammonia buffer capacity by depletion of carbon skeletons as glutamine is formed (from alpha-ketoglutarate and glutamate) which could potentially lower the arteriovenous concentration gaps.
In conclusion, in the basal state and during hyperammonaemia clinically important arterial-venous ammonia concentration gaps exist in both healthy persons and patients with cirrhosis. However, the concentration differences between healthy persons and patients with cirrhosis seems to be relatively constant across sampling sites, in support of the use of venous sampling for ammonia measurement in the clinical setting as long as it is clear what sampling site was used and interpretation takes sampling site into account. Also, ammonia concentrations normalised to calculated ULN values were valid for both arterial and venous sampling sites.
Acknowledgements
The authors take this opportunity to express their gratitude to Ms. Mette Mejlby Hansen from the Laboratory at Department of Hepatology and Gastroenterology, Aarhus University Hospital, Denmark, for skilful technical and practical assistance.
Author contributions
Lars Djernes, Hendrik Vilstrup, Peter Ott and Peter Lykke Eriksen all contributed to the study conception and design. Experiments, data collection and analysis were performed by Lars Djernes and Peter Lykke Eriksen. The first draft of the manuscript was prepared by Lars Djernes and Peter Lykke Eriksen. All authors contributed to the editing and revision of the manuscript and approved the final version.
Funding
This work was supported by the Aase and Ejnar Danielsen’s Foundation (PLE, grant no. 20-10-0266), Inger Goldmann’s Fund (PLE, grant no. 252023-001), Torben and Alice Frimodt’s Foundation (PLE) and the Health Research Foundation of Central Denmark Region (PLE, grant no. A3306).
Open access funding provided by Aarhus University Hospital
Data availability
No datasets were generated or analysed during the current study.
Declarations
Competing interests
The authors declare no competing interests.
Ethics approval
The study conformed to the ethical guidelines of the Declaration of Helsinki and Danish and European Union legislation on data protection. The protocol was approved by the Central Denmark Region Committees on Health Research Ethics (110-72-165-19).
Consent to participate
Written informed was obtained from all individual participants included in the study.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- Bajaj JS, Bloom PP, Chung RT et al (2020) Variability and lability of Ammonia levels in healthy volunteers and patients with cirrhosis: implications for Trial Design and Clinical Practice. Am J Gastroenterol 115:783–785. 10.14309/ajg.0000000000000384 [DOI] [PMC free article] [PubMed]
- Balcar L, Krawanja J, Scheiner B et al (2023) Impact of ammonia levels on outcome in clinically stable outpatients with advanced chronic liver disease. JHEP Rep 5:100682. 10.1016/j.jhepr.2023.100682 [DOI] [PMC free article] [PubMed]
- Ballester MP, Tranah TH, Balcar L et al (2023) Development and validation of the AMMON-OHE model to predict risk of overt hepatic encephalopathy occurrence in outpatients with cirrhosis. J Hepatol 79:967–976. 10.1016/j.jhep.2023.05.022 [DOI] [PubMed]
- Dam G, Keiding S, Munk OL et al (2011) Branched-chain amino acids increase arterial blood ammonia in spite of enhanced intrinsic muscle ammonia metabolism in patients with cirrhosis and healthy subjects. Am J Physiol Gastrointest Liver Physiol 301:G269–277. 10.1152/ajpgi.00062.2011 [DOI] [PubMed]
- Eriksen PL, Djernes L, Vilstrup H, Ott P (2023) Clearance and production of ammonia quantified in humans by constant ammonia infusion - the effects of cirrhosis and ammonia-targeting treatments. J Hepatol 79:340–348. 10.1016/j.jhep.2023.03.042 [DOI] [PubMed]
- Eriksson LS, Broberg S, Björkman O, Wahren J (1985) Ammonia metabolism during exercise in man. Clin Physiol 5:325–336. 10.1111/j.1475-097x.1985.tb00753.x [DOI] [PubMed]
- European Association for the Study of the Liver(2022) EASL Clinical Practice guidelines on the management of hepatic encephalopathy. J Hepatol 77:807–824. 10.1016/j.jhep.2022.06.001 [DOI] [PubMed]
- Haj M, Rockey DC (2020) Ammonia levels do not Guide Clinical Management of patients with hepatic Encephalopathy caused by cirrhosis. Am J Gastroenterol 115:723–728. 10.14309/ajg.0000000000000343 [DOI] [PubMed]
- Huizenga JR, Gips CH, Tangerman A (1996) The contribution of various organs to Ammonia formation: a review of factors determining the arterial Ammonia concentration. Ann Clin Biochem 33:23–30. 10.1177/000456329603300103 [DOI] [PubMed]
- Lattanzi B, D’Ambrosio D, Merli M (2019) Hepatic encephalopathy and Sarcopenia: two faces of the same metabolic alteration. J Clin Exp Hepatol 9:125–130. 10.1016/j.jceh.2018.04.007 [DOI] [PMC free article] [PubMed]
- Lockwood AH (2004) Blood ammonia levels and hepatic encephalopathy. Metab Brain Dis 19:345–349. 10.1023/b:mebr.0000043980.74574.eb [DOI] [PubMed]
- Manjunath R, NH N, Bhardwaj V (2014) Clinical co relation between arterial versus venous ammonia levels in hepatic encephlopathy in cirrhosis of liver. J Evol Med Dent Sci 3:5322
- Nicolao F, Efrati C, Masini A et al (2003) Role of determination of partial pressure of ammonia in cirrhotic patients with and without hepatic encephalopathy. J Hepatol 38:441–446. 10.1016/S0168-8278(02)00436-1 [DOI] [PubMed]
- Ong JP, Aggarwal A, Krieger D et al (2003) Correlation between ammonia levels and the severity of hepatic encephalopathy. Am J Med 114:188–193. 10.1016/s0002-9343(02)01477-8 [DOI] [PubMed]
- Seligson D, Hirahara K (1957) The measurement of ammonia in whole blood, erythrocytes, and plasma. J Lab Clin Med 49:962–974 [PubMed]
- Shalimar null, Sheikh MF, Mookerjee RP et al (2019) Prognostic role of Ammonia in patients with cirrhosis. Hepatology 70:982–994. 10.1002/hep.30534 [DOI] [PubMed]
- Stahl J (1963) Studies of the blood ammonia in liver disease: its diagnostic, prognostic, and therapeutic significance. Ann Intern Med 58:1. 10.7326/0003-4819-58-1-1 [DOI] [PubMed]
- Tapper EB, Rahimi RS (2020) Low-value levels: Ammonia Testing does not improve the outcomes of overt hepatic encephalopathy. Am J Gastroenterol 115:685–686. 10.14309/ajg.0000000000000454 [DOI] [PubMed]
- Tranah TH, Ballester M-P, Carbonell-Asins JA et al (2022) Plasma ammonia levels predict hospitalisation with liver-related complications and mortality in clinically stable outpatients with cirrhosis. J Hepatol 77:1554–1563. 10.1016/j.jhep.2022.07.014 [DOI] [PubMed]
- Tyor MP, Owen EE, Berry JN, Flanagan JF (1960) The relative role of extremity, liver, and kidney as ammonia receivers and donors in patients with liver disease. Gastroenterology 39:420–424 [PubMed] [Google Scholar]
- Vierling JM, Mokhtarani M, Brown RS et al (2016) Fasting blood Ammonia predicts risk and frequency of hepatic encephalopathy episodes in patients with cirrhosis.Clin Gastroenterol Hepatol 14:903–906e1. 10.1016/j.cgh.2015.11.018 [DOI] [PubMed]
- Warter JM, Brandt C, Marescaux C et al (1983) The renal origin of sodium valproate-induced hyperammonemia in fasting humans. Neurology 33:1136–1140. 10.1212/wnl.33.9.1136 [DOI] [PubMed]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
No datasets were generated or analysed during the current study.