Objectives: The aim of this study was to determine whether the number connection test (NCT) times of a group of cirrhotic patients without clinically overt hepatic encephalopathy and a group of healthy patients without liver disease who were undergoing endoscopy were prolonged after sedation with short acting i.v. benzodiazepines.
Methods: All patients were administered the NCT in a standard fashion for 30 min before sedation for an upper GI endoscopy and then 2 h after sedation postprocedure. Two NCTs were carried out before and 2 h after sedation, and the mean of the tests pre- and postsedation calculated. Based on the upper limit of the 95% CI of the presedation NCT of patients without liver disease as the cut-off level for hepatic encephalopathy, the proportion of cirrhotic patients with subclinical encephalopathy before and after sedation were also determined.
Results: A total of 61 consecutive cirrhotic patients who underwent therapeutic upper GI endoscopy completed the study. The mean presedation NCT time was 43.5 s (95% CI = 39.0-48.1 s) and the mean postsedation NCT time 60.0 s (95% CI = 50.7-69.3 s). The difference between the mean pre- and postsedation NCT times was 16.4 s (95% CI = 9.8-23.1 s; p < 0.001). A total of 38 consecutive patients without clinical or biochemical evidence of liver disease who presented for upper GI endoscopy completed the NCT as described for the group of cirrhotic patients. The mean (+/- SD) baseline NCT time was 34.7+/-7.9 s (95% CI = 32.1-37.2 s), whereas the mean postsedation NCT time was 33.7+/-8.5 s (95% CI = 30.9-36.5 s). This difference was not statistically significant (p = 0.177). Using the upper limit of the 95% CI of the mean (37.4 s) of the presedation time in the patients without liver disease as the cut-off between normal and encephalopathy, the number of cirrhotic patients with abnormal presedation NCT times was 33 patients (54.1%), and this number rose to 46 patients (75.4%) after sedation with midazolam. This increase in proportion of cirrhotic patients with prolonged NCT time was statistically significant (p < 0.001).