Serum ammonia does not guide management and is overutilised in patients with cirrhosis in hospital settings.

Background: Hepatic encephalopathy is a confusional state associated with cirrhosis. Serum ammonia levels are neither sensitive nor specific for the diagnosis.

Aims: We audited the ordering location and hospital unit whilst assessing the impact on management at a major Australian tertiary centre.

Methods: We conducted a single-centre retrospective chart review of the ordering of serum ammonia levels between 1 March 2019 and 29 February 2020 at The Royal Melbourne Hospital, a tertiary-referral centre in Melbourne, Victoria. Demographic, medication and pathology results, including serum ammonia measurements, were collected. The main outcomes assessed were ordering location, sensitivity, specificity and impact on management.

Results: A total of 1007 serum ammonia tests were ordered in 425 patients. Nearly all ammonia ordering was by non-gastroenterologists, 24.2% by the intensive care unit, 23.1% by general medicine and 19.5% by the emergency department (ED). Only 21.6% of patients had a history of cirrhosis, with hepatic encephalopathy diagnosed in 13.6%. On subgroup analysis, 217 ammonia tests were performed in 92 patients with cirrhosis. Cirrhotic patients were older (64 vs 59 years, P = 0.012) and had higher median ammonia levels (64.46 vs 59 μmol/L, P < 0.001) compared with non-cirrhotic patients. In cirrhotic patients, the sensitivity and specificity for serum ammonia and diagnosis of hepatic encephalopathy were 75% and 52.3% respectively.

Conclusion: We affirm the poor utility of serum ammonia levels for guiding management of hepatic encephalopathy within the Australian context. ED and general medical units account for the majority of test ordering within the hospital. Understanding where ordering occurs provides a target for targeted education.