L E Schmidt1, K Dalhoff. 1. Department of Hepatology A, Rigshospitalet, Copenhagen, Denmark. lars.schmidt@dadlnet.dk
Abstract
BACKGROUND: Hyperamylasaemia and even acute pancreatitis have been reported in patients with paracetamol poisoning. AIMS: To describe the incidence, clinical characteristics, and prognostic implications of hyperamylasaemia in paracetamol poisoning. PATIENTS: Six hundred and two patients transferred to a specialized unit with severe paracetamol poisoning and 212 unselected patients admitted from the local region. METHODS: Retrospective study based on hospital charts. The optimum threshold of serum amylase to discriminate non-survivors was identified. RESULTS: An elevated serum amylase (>100 U/L) occurred in 28 of the unselected patients (13%), in 218 of the transferred patients (36%), and in 118 of 148 patients (80%) with fulminant hepatic failure. Only 33 cases of paracetamol-associated acute pancreatitis were diagnosed. A threshold serum amylase of 150 U/L to discriminate non-survivors had sensitivity 76%, specificity 85%, positive predictive value 33%, and negative predictive value 97%. In a logistic regression analysis, a serum amylase > 150 U/L was associated with an excess mortality (odds ratio 5.0, 2.6-9.7). CONCLUSIONS: Hyperamylasaemia is frequent in patients with paracetamol poisoning, whereas clinical acute pancreatitis occurs rarely. The incidence of hyperamylasaemia increases with the degree of hepatic dysfunction. A serum amylase exceeding 1.5 times the upper normal limit indicates a poor prognosis.
BACKGROUND: Hyperamylasaemia and even acute pancreatitis have been reported in patients with paracetamolpoisoning. AIMS: To describe the incidence, clinical characteristics, and prognostic implications of hyperamylasaemia in paracetamolpoisoning. PATIENTS: Six hundred and two patients transferred to a specialized unit with severe paracetamolpoisoning and 212 unselected patients admitted from the local region. METHODS: Retrospective study based on hospital charts. The optimum threshold of serum amylase to discriminate non-survivors was identified. RESULTS: An elevated serum amylase (>100 U/L) occurred in 28 of the unselected patients (13%), in 218 of the transferred patients (36%), and in 118 of 148 patients (80%) with fulminant hepatic failure. Only 33 cases of paracetamol-associated acute pancreatitis were diagnosed. A threshold serum amylase of 150 U/L to discriminate non-survivors had sensitivity 76%, specificity 85%, positive predictive value 33%, and negative predictive value 97%. In a logistic regression analysis, a serum amylase > 150 U/L was associated with an excess mortality (odds ratio 5.0, 2.6-9.7). CONCLUSIONS: Hyperamylasaemia is frequent in patients with paracetamolpoisoning, whereas clinical acute pancreatitis occurs rarely. The incidence of hyperamylasaemia increases with the degree of hepatic dysfunction. A serum amylase exceeding 1.5 times the upper normal limit indicates a poor prognosis.