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.