Predictors of nonresponse to lactulose for minimal hepatic encephalopathy in patients with cirrhosis

Abstract

Background/Aims: Minimal hepatic encephalopathy (MHE) impairs health-related quality of life and predicts overt hepatic encephalopathy (HE) in cirrhotic patients. Lactulose is effective in the treatment of MHE. However, not all patients respond to lactulose. We evaluated predictors of nonresponse to lactulose. Patients and methods: Consecutive 110 cirrhotic patients without HE were evaluated for MHE by psychometry, P300 auditory event-related potential (P300ERP), venous ammonia and critical flicker frequency (CFF). MHE was diagnosed by abnormal psychometry and P300ERP (>2 SD). MHE patients were treated with lactulose for 1 month. Response was defined by normalization of the abnormal test parameters (both psychometric tests and P300ERP). Results: Sixty patients (54.5%) were diagnosed as having MHE: 17/39 (44%) in Child's A, 21/42 (50%) Child's B and 22/29 (76%) in Child's C. There was a significant difference between Child's C's vs Child's A's and B's (P<0.05). Abnormal psychometric tests and abnormal P300ERP were seen in 74 (67%) and 74 (67%) patients respectively. Of 60 patients with MHE, after treatment, psychometry remained abnormal in 22 (36.6%) and P300ERP in 21 (35%) patients. CFF was<38 Hz in 34 (57%) and 11 (18%) patients, respectively, before and after treatment in MHE patients. There was a significant difference between the baseline serum sodium level (134.7±2.6 vs 131.1±2.2 mmol/L, P=0.001) and the venous ammonia level (76.6±20.7 vs 113.4±22.8 μmol/L, P=0.001) between responders vs nonresponders. Receiver operating characteristic analysis was performed to identify the cutoff for venous ammonia [cutoff 93.5 mmol/L, area under the curve (AUC) 0.892 (0.814-0.970)] and for the serum sodium level [cutoff 132.5 mmol/L, AUC 0.874 (0.779-0.998)]. Taking a cutoff of 93.5 mmol/L for ammonia patient had a sensitivity of 88.5% and a specificity of 79.4%, respectively, and a cutoff of 132.5 mmol/L for serum sodium patient had a sensitivity of 76.5% and a specificity of 88.5% for nonresponse to lactulose. On univariate analysis and multivariate analysis, serum sodium and venous ammonia were the only two parameters associated with nonresponse to lactulose. Conclusion: The prevalence of MHE was 55% and MHE improved in 57% patients with lactulose. Baseline low serum sodium and high venous ammonia were highly predictive of nonresponse to lactulose therapy.