Between October 2008 and December 2009, 360 patients with cirrhosis recovering from an episode of HE were screened; 235 patients (65.2%) who had recovered from HE and met the inclusion criteria were included in the study. The etiologies of cirrhosis were alcohol (n=94), chronic hepatitis B (n=49), chronic hepatitis C (n=36), primary biliary cirrhosis (n=2), autoimmune hepatitis (n=3), and cryptogenic cirrhosis (n=51). A total of 125 patients (34.8%) were excluded from the study because of a history of recent alcohol intake (n=42), renal impairment (n=38), hepatocellular carcinoma (n=16), recent use of drugs affecting psychomotor performance (n=1), or a severe medical problem (n=20) or because they were not willing to participate in routine follow-up (n=8). Patients with chronic hepatitis C were decompensated in the form of ascites and were not treated with interferon-α during the follow-up period. The clinical and demographic characteristics of the patients in the three groups were comparable (Table 1).
Recurrence of HE
Of 235 patients, 38 (16.2%) were lost to follow-up, with a median follow-up of 4 months (range 1–7 months), 12 in Gp-L, 13 in Gp-P, and 13 in Gp-N. The compliance with follow-up in our study was 83.8%; 68 patients in Gp-L, 64 patients in Gp-P, and 65 patients in Gp-N were followed up for 12 months.
During the study period, 77 (39.1%) of 197 patients developed an episode of HE (Figure 2). In all, 18 (26.5%) of 68 in Gp-L (grade 1, n=2; grade 2, n=9; grade 3, n=6; and grade 4, n=1), 22 (34.4%) of 64 in Gp-P (grade 1, n=2, grade 2, n=11; grade 3, n=6; and grade 4, n=3;), and 37 (56.9%) of 68 in Gp-N (grade 1, n=5, grade 2, n=21; grade 3, n=8; and grade 4, n=3;) developed HE (P=0.001). Lactulose therapy, as compared with no therapy, was significantly more effective in secondary prophylaxis (26.2% vs. 56.9%, P=0.001). Similarly, probiotics therapy, as compared with no therapy, was associated with a lower incidence in recurrence of HE (34.4% vs. 56.9%, P=0.02), but no significant difference was found between lactulose and probiotics therapy (26.2% vs. 34.4%, P=0.349). There was a significant difference in recurrence of HE on intention-to-treat analysis considering patients lost to follow-up who had developed HE (30 [37.5%] of 80 in Gp-L, 35 [45.4%] of 77 patients in Gp-P, and 50 [64.1%] of 78 patients in Gp-N; P=0.003). The etiologies of precipitating factors are shown in Table 2, and there was no significant difference among the causes of HE.
Probability of developing hepatic encephalopathy (HE) in patients receiving lactulose or probiotics treatment or no therapy.
There was no significant difference in patients who were on β-blockers and developed HE among the three groups (Gp-L, n=5; Gp-P, n=7; and Gp-N, n=9; P=0.631). Similarly, no difference was seen in large spontaneous shunts among the three patients who developed HE (Gp-L, n=4; Gp-P, n=4; and Gp-N, n=6; P=0.43).
Hospitalizations Other Than for HE
Hospitalizations other than for HE occurred in 19 (27.9%) of 68 patients in Gp-L (SBP, n=7; pneumonia, n=3; upper gastrointestinal bleed, n=3; urinary tract infection, n=1; and hepatorenal syndrome, n=5), 21 (32.8%) of 64 patients in Gp-P (SBP, n=8; pneumonia, n=2; upper gastrointestinal bleed, n=3; urinary tract infection, n=3; and hepatorenal syndrome n=5), and 28 (43.0%) of 65 patients in Gp-N (SBP, n=13; upper gastrointestinal bleed, n=4; urinary tract infection n=2; pneumonia, n=3; and hepatorenal syndrome, n=6) (P=0.134). A total of 40 patients died during follow-up (Gp-L:Gp-P:Gp-N=13:11:16; P=0.56). All patients who died were decompensated at baseline.
Results of Psychometric Tests and CFF
All patients could perform the BDT and DST, 187 patients could perform the NCT, and the remaining 48 performed FCT because of illiteracy. There was a high prevalence of abnormal psychometry test results (NCT-A, 71.5%; NCT-B, 69.2%; DST, 76.9%; and BDT, 85.2%), and FCT-A and B were abnormal in 35 of 48 patients (72.9%; Table 3). There was a high prevalence of two or more abnormal psychometry test in these groups (n=167, i.e., 71.0%). Patients who developed HE had significantly higher NCT (71.2±15.5 vs. 53.3±17.8 s, P=0.001), FCT (83.8±18.7 vs. 69.6±20.5 s, P=0.001), DST (128.6±21.4 vs. 107.1±19.1 s, P=0.001), and BDT raw scores (27.0±3.6 vs. 29.0±3.8,P=0.001) compared with patients who did not develop HE. Arterial ammonia level was not significantly higher in patients with two or more abnormal psychometric tests compared with patients with fewer than two abnormal psychometric tests (91.8±17.9 vs. 86.9±22.4 μmol/l, P=0.142).
All patients could perform CFF. CFF was <38 Hz in 118 patients (50.2%; Table 3). In all, 62 patients had psychometric impairment without CFF impairment, 12 had CFF impairment without psychometric impairment, and 106 had both CFF and psychometric impairments. Patients diagnosed with two or more abnormal psychometric tests had a CFF that was significantly lower than those with fewer than two abnormal tests (37.7±1.6 vs. 39.5±2.0 Hz, P=0.001). CFF, when used alone for the diagnosis of patients with two or more abnormal psychometry tests, had a sensitivity and specificity of 63.2% and 83.6%, respectively. There was a significant difference in baseline CFF in patients who developed HE vs. those who did not (37.7±1.6 vs. 38.5±1.96 Hz, P=0.008).
Arterial Ammonia Levels
Arterial ammonia levels were analyzed in all study patients at baseline and at the 3-month follow-up. Mean levels for arterial ammonia were 90.3±19.4 μmol/l. Analysis of the change in mean arterial ammonia levels in the study groups at baseline and follow-up of 3 months of treatment showed a significant decrease in ammonia levels in Gp-L (P=0.03) and Gp-P (P=0.04) but not in Gp-N (P=0.597; Table 4).
All patients were able to tolerate and remained compliant with lactulose therapy in Gp-L. Of 68 patients, 18 (26.4%) had diarrhea, 11 (16.2%) had abdominal bloating, and 12 (17.6%) had distaste to lactulose. In these patients, the dose was reduced but not stopped. In Gp-N, constipation was reported in 14 (21.5%) patients and was managed by dietary modifications. In the Gp-P group, 9 (14%) patients complained of abdominal distension and 14 (21.8%) of constipation managed with dietary advice and on-and -off use of proton pump inhibitors. None of the patients in the Gp-P group developed increased frequency of stools, fever, or rash related to probiotics.
Correlation of HE With Baseline Parameters
CFF, baseline arterial ammonia, two or more abnormal psychometric tests, and Child-Turcotte-Pugh scoring (CTP) score were significantly associated with the development of HE upon univariate analysis ( Table 5 ). On multivariate analysis, recurrence of overt HE was significantly associated with the presence of two or more abnormal psychometric test results and baseline arterial ammonia level ( Table 6 ).