Clinical efficacy and safety of lactulose for minimal hepatic encephalopathy: a meta-analysis Ming Luo a, Lei Li a, Chen-Zheng Lu b and Wu-Kui Cao c

Transcription

1 1250 Original article Clinical efficacy and safety of lactulose for minimal hepatic encephalopathy: a meta-analysis Ming Luo a, Lei Li a, Chen-Zheng Lu b and Wu-Kui Cao c Objective To evaluate the clinical efficacy of lactulose in patients with minimal hepatic encephalopathy (MHE). Methods Randomized controlled trials (RCTs) comparing lactulose with placebo or with no intervention in the management of MHE that were conducted from January 1990 to July 11 were searched from MEDLINE, EMBASE, SCI, Cochrane Controlled Trials Register, and China Biological Medicine Database. Studies with a Jadad score higher than 3 were included in the meta-analysis and evaluated using RevMan5.0 software for relative risk (RR) or weighted mean difference (WMD) with 95% confidence intervals (95% CI). Sensitivity analysis was performed on the ethnical differences and quality of the trials. Publication bias was observed using an inverted funnel plot. Results Nine studies with 434 patients were included in the meta-analysis. Compared with placebo or no intervention, lactulose significantly reduced the risk of no improvement in neuropsychological tests (RR: 0.52, 95% CI: , P < ), the time required for the completion of the number connection test-a (WMD: 26.95, 95% CI: to 16.10, P < ), and the mean number of abnormal neuropsychological tests (WMD: 1.76, 95% CI: 1.96 to 1.56, P < ). Furthermore, the meta-analysis also showed that lactulose prevented the progression to overt hepatic encephalopathy (RR: 0.17, 95% CI: , P = 0.002), reduced blood ammonia levels (WMD: 9.89 lmol/l, 95% CI: to 8.77 lmol/l, P < ), and improve health-related quality of life (WMD: 6.05, 95% CI: 6.30 to 5., P < ). However, no significant difference was observed in the mortality of patients with MHE (RR: 0.75, 95% CI: , P = 0.66), and lactulose significantly increased the incidence of diarrhea (RR: 4.38, 95% CI: , P = 0.01). Conclusion Lactulose has significant beneficial effects for patients with MHE compared with placebo or no intervention. Eur J Gastroenterol Hepatol 23: c 11 Wolters Kluwer Health Lippincott Williams & Wilkins. European Journal of Gastroenterology & Hepatology 11, 23: Keywords: lactulose, meta-analysis, minimal hepatic encephalopathy, treatment a Graduate School, Tianjin Medical University, b Department of Hepatology and c Tianjin Liver Disease Institute, Tianjin Infection Disease Hospital, Tianjin, China Correspondence to Dr Wu-Kui Cao, MD, Tianjin Liver Disease Institute, Tianjin Infection Disease Hospital, No.75, Sudi Road, Nankai District, Tianjin , China Tel: ; fax: ; caowukui@126.com Received 26 August 11 Accepted 13 September 11 Introduction Minimal hepatic encephalopathy (MHE) is the mildest form of hepatic encephalopathy (HE). Patients with MHE have no recognizable mental and neurologic symptoms associated with HE on standard clinical examination, but have several quantifiable neuropsychological and neurophysiological defects [1]. The prevalence of MHE reportedly varies between 30 and 84% in patients with cirrhosis without overt hepatic encephalopathy (OHE) [2]. MHE impairs the cognitive function, affects health-related quality of life (HRQoL), and results in progression to OHE. To date, no ideal test or gold standard for the diagnosis of MHE is available. According to the recommendation of the Working Party at the 11th World Congress of Gastroenterology, psychometric, and neurophysiologic methods have been the most trusted and widely used tests to diagnose MHE [1]. Ammonia is universally acknowledged to play a key role in the pathogenesis of MHE, which is believed to be similar to that of OHE [3]. Thus, various kinds of therapeutic interventions have been aimed at reducing ammonia production and absorption. Lactulose can significantly lower ammonia levels and has served as the primary therapy for OHE. A recent consensus conference also promulgated lactulose as the first-line drug in the treatment of MHE, in concordance with previous study data and the American Association for the Study of Liver Diseases (AASLD) survey [4]. However, a Cochrane systematic review reported by Als-Nielsen et al. [5] demonstrated the lack of data supporting the efficacy of lactulose and concluded that insufficient evidence was available to recommend or refute the use of lactulose for HE. Furthermore, although lactulose is viewed as the standard therapy for MHE, its efficacy and safety has never been evaluated in a meta-analysis of randomized controlled trials (RCTs). Therefore, we conducted the meta-analysis to evaluate the efficacy and safety of lactulose in the management of MHE and to provide useful information for clinical practice. Materials and methods Search strategy Two independent reviewers (M.L. and L.L.) searched for data from MEDLINE, EMBASE, SCI, the China X c 11 Wolters Kluwer Health Lippincott Williams & Wilkins DOI: /MEG.0b013e32834d1938

2 Lactulose for minimal hepatic encephalopathy Luo et al Biological Medicine Database (CBM), and the Cochrane Controlled Trials Register. The search was limited to human studies on adult patients from January 1990 to July 11 without language restrictions. The following search terms were used: MHE, latent hepatic encephalopathy, subclinical hepatic encephalopathy, lactulose, placebo-controlled trials, and RCTs. The abstracts of all articles were evaluated, and most full-text papers were reviewed. When the clinical trials or the results were unclear, the authors and lead researchers were contacted for more details. Selection criteria All RCTs that compared lactulose with placebo or with no intervention were collected, including all abstracts and most full texts. The inclusion criteria were as follows: (a) patients with liver cirrhosis that have a definitive diagnosis of MHE according to neuropsychological tests; (b) the study should be an RCT comparing the therapeutic effects of lactulose with placebo or no intervention; (c) the study has to report the clinical outcomes of patients with MHE. The exclusion criteria were as follows: (a) studies involving patients with OHE; (b) uncontrolled or nonrandomized trials; (c) reviews, case reports, and preclinical studies. Data abstraction and quality assessment The review was conducted in accordance with the guidelines of the QUOROM statement [6]. Two independent reviewers (M.L. and L.L.) evaluated trial methodological quality and extracted the data. The last observed record was carried forward in instances where patients had missing outcome data. If reviewers had any disagreements, a third party (W.-K.C.) resolved the inconsistency by discussion. The following data were extracted and recorded: type of HE, number of patients, number of patients randomly assigned to each group, diagnostic tests used for diagnosing MHE, trial design and quality, duration of treatment, number of dropouts, and side-effect profile. Studies were graded for methodological quality using the Jadad five-point scoring system, which includes three itemized scoring systems: randomization (0 2 points), blinding (0 2 points), and dropouts and withdrawals (0 1 point) [7]. The quality scale ranges from 0 to 5. Studies with a total quality score of 3 or more were considered high-quality studies and were then included in the meta-analysis, and those with a score of less than 3 were considered low-quality and thereby excluded. Data analysis The meta-analysis was conducted in accordance with the Cochrane Reviewers Handbook recommended by the Cochrane Collaboration. All statistical analyses were performed using Review Manager Version 5.0 (provided by the Cochrane Collaboration, Oxford, UK). Outcomes are expressed as relative risk (RR) for dichotomous data and as weighted mean difference (WMD) for continuous data, both reported with 95% confidence intervals (CI). Heterogeneity was evaluated using both w 2 and I 2 statistics, and P values less than 0.05, along with I 2 values greater than 50%, were considered to show significant heterogeneity. When the data were homogenous (P > 0.05, I 2 < 50%), the results were estimated using the Mantzel Haenzel (M H) method, and the fixed effects model was used. If the data had significant heterogeneity (P < 0.05, I 2 > 50%), the DerSimonian Laird (D L) method and the random effects model were used for outcomes. Subgroup analysis was performed to determine the reason for the significant heterogeneity. A sensitivity analysis was performed to evaluate whether ethnic difference and the quality of the studies influenced the efficacy of lactulose and to verify the results of meta-analysis with higher quality studies (scores greater than 4 in the Jadad quality scoring system). Publication bias was observed using an inverted funnel plot. Results Primary outcome Nine studies with a total of 434 patients met the specified criteria and were included in the meta-analysis (Fig. 1). All nine studies enrolled patients with cirrhosis with different etiologies, including viral, alcohol, and so on, reported the methods used for the diagnosis of MHE, and compared lactulose with placebo or with no intervention in the management of MHE. Six studies [8,9,11,12,15,16] used a parallel group design, and all the studies were described as randomized. However, only four studies [9,11,15,16] reported the methods used for randomization. Treatment allocation was concealed in four studies [8,10,12,15], and double blinding was used in two studies [8,15]. The characteristics and Jadad quality scores of studies included in the meta-analysis were presented in Table 1. All studies assessed oral lactulose. The daily mean doses of lactulose ranged from to 80 g, and the median dose was 60 g. In five studies [8,9,12,13,15], the dose was adjusted to obtain two to three semisoft stools per day. The duration of treatment ranged from 2 to 24 weeks, and the median duration was 8 weeks. None of the studies conducted patient follow-ups after the end of treatment. Risk of no improvement in neuropsychological tests All nine studies reported a risk of no improvement in neuropsychological tests after lactulose therapy, and the data were homogenous (w 2 = 7.12, P = 0.52, I 2 = 0%). Compared with placebo or no intervention, lactulose reduces the risk of no improvement in neuropsychological tests (RR: 0.52, 95% CI: , P < ) (Fig. 2).

3 1252 European Journal of Gastroenterology & Hepatology 11, Vol 23 No 12 Changes in mean number of abnormal neuropsychological tests Four studies [8,11,15,16] reported changes in the mean number of abnormal neuropsychological tests. All four studies used two or more neuropsychological tests to evaluate the therapeutic effect of lactulose. Changes in the mean number of abnormal neuropsychological tests were defined as Dabnormal neuropsychological tests, which served as a measure of improvement in cognitive function. The result of meta-analysis shows that lactulose significantly lowered the mean number of abnormal neuropsychological tests in patients with MHE (WMD: Fig , 95% CI: 1.96 to 1.56, P < ) with no significant heterogeneity (w 2 = 4.56, P = 0.21, I 2 = 34%) (Fig. 4). Reduction in number connection test-a time A number connection test (NCT) is a neuropsychological test used to assess cognitive function in patients with MHE. Two studies [13,14] reported the time required for the completion of the NCT-A. No heterogeneity was observed among the studies (w 2 = 0.15, P = 0.70, I 2 = 0%). The meta-analysis shows that lactulose significantly reduced the time required for the completion of NCT-A in patients with MHE (WMD: 26.95, 95% CI: to 16.10, P < ) compared with placebo or no intervention (Fig. 3). MEDLINE, EMBASE, CBM, SCI Cochrane Controlled Trials Register (January 1990 to July 11) Potentially relevant articles indentified (56) Potentially appropriate trials (21) Review articles (19) Case reports (7) Animal experiments (5) Repeat publications (4) Reduction in blood ammonia levels Five studies [9,12 14,16] reported that lactulose appeared to reduce blood ammonia levels, but Nie et al. [12] and Horsmans et al. [9] did not report blood ammonia levels after lactulose therapy. Three studies [13 15] reported a reduction in blood ammonia levels, and the data were homogenous (w 2 =2.85, P =0.24, I 2 = 30%). The metaanalysis indicates that lactulose significantly lowered blood ammonia levels (WMD: 9.89 mmol/l, 95% CI: to 8.77mmol/l, P < ) (Fig. 3). RCT (lactulose versus placebo or no intervention) (9) Involve OHE (7) Not controlled (2) Other interventions (3) QUOROM flow diagram of included and excluded studies on the use of lactulose in the management of minimal hepatic encephalopathy. Improvement in health-related quality of life The impact of MHE on health-related quality of life in patients with liver cirrhosis has been paid increasing attention. used the Quality of Life Scale recommended by WHO to assess the HRQoL of patients with MHE treated by lactulose. Prasad et al. [15] and used the Sickness Impact Profile (SIP) questionnaire (Medical Outcome Trust, Boston, Table 1 Characteristics of included randomized controlled trials for meta-analysis Study Location Sample size Number of patients without improvement/total Experiment/control intervention Lactulose Control Diagnostic tests Treatment duration (weeks) Number of dropouts/total Lactulose Control Jadad score Watanabe Japan 36 Lactulose/no treatment 12/22 11/14 NCT, SDT, BDT, 8 2/22 1/14 5 et al. [8] WAIS-R Horsmans et al. [9] Belgium 14 Lactulose/lactose 2/7 6/7 NCT, RCT 2 0/7 0/7 4 Li et al. [10] China 86 Lactulose/supportive treatment 22/48 27/38 NCT, DST, NH 3 4 Unknown 3 Dhiman India 26 Lactulose/no treatment 6/14 12/12 NCT, FCT, PCT, BDT 12 4/14 4/12 4 et al. [11] Nie et al. [12] China 43 Lactulose/vitamin B 6/21 15/21 NCT, DST, SEP, NH /21 2/22 3 Xing and China 48 Lactulose/vitamin B 11/24 22/24 NCT, DST, EEG, 4 1/24 2/24 3 Liu [13] NH 3 Zeng and China Lactulose/vitamin B 7/ 12/ NCT, DST, SEP, NH 3 8 1/ 2/ 3 Li [14] Prasad India 61 Lactulose/no treatment 6/31 /30 NCT, FCT, PCT, BDT 12 6/31 10/30 5 et al. [15] India 80 Lactulose/no treatment 21/ 36/ NCT, FCT, BDT, PCT 12 5/ 9/ 5 BDT, block design test; DST, digit symbol test; EEG, electroencephalogram; FCT, figure connection test; NCT, number connection test; NH3, blood ammonia levels; PCT, the picture completion test; SDT, the symbol digit test; SEP, brain stem evoked potential; WAIS-R, Wechsler Adult Intelligence Scale.

4 Lactulose for minimal hepatic encephalopathy Luo et al Fig. 2 Lactulose Control Risk ratio Study or subgroup Events Total Events Total Weight M H, Fixed, 95% CI Risk of no improvement in neurophychological tests Prasad et al. [15] % 0.29 [0.14, 0.62] Horsmans et al. [9] % 0.33 [0.10, 1.12] Nie et al. [12] % 0.38 [0.18, 0.80] Dhiman et al. [11] % 0.45 [0.25, 0.81] Xing and Liu [13] % 0.50 [0.32, 0.79] % 0.58 [0.29, 1.17] % 0.58 [0.43, 0.80] Li et al. [10] % 0.65 [0.45, 0.93] Watanabe et al. [8] % 0.69 [0.43, 1.11] [0.44, 0.62] Total events Heterogeneity: χ 2 = 7.12, d.f. = 8 (P = 0.52); l 2 = 0% Test for overall effect: Z = 7.49 (P< ) Risk ratio M H, Fixed, 95% CI Risk of progression to OHE 1 8 Dhiman et al. [11] Prasad et al. [15] Xing and Liu [13] Total events 2 18 Heterogeneity: χ 2 = 0.23, d.f. = 4 (P = 0.99); l 2 = 0% Test for overall effect: Z = 3.12 (P = 0.002).3% 13.5% 13.5% 12.6%.1% 0.13 [0.02, 0.91] 0.17 [0.01, 3.29] 0.17 [0.01, 3.29] 0. [0.01, 3.96] 0.25 [0.03, 2.14] 0.17 [0.06, 0.52] Incidence of diarrhea Horsmans et al. [9] 3 Nie et al. [12] 2 2 Watanabe et al. [8] 8 Total events Mortality of patients with MHE Prasad et al. [15] 0 31 Dhiman et al. [11] Total events Heterogeneity: χ 2 = 0.17, d.f. = 3 (P = 0.98); l 2 = 0% Test for overall effect: Z = 2.46 (P = 0.01) Heterogeneity: χ 2 = 2.53, d.f. = 2 (P = 0.28); l 2 = 21% Test for overall effect: Z = 0.44 (P = 0.66) % 15.8% 15.5% 37.8% 69.3% 21.0% 9.7% 3.00 [0., 22.30] 4.78 [0.24, 94.12] 5.00 [0.26, 98.00] 5.09 [0.71, 36.43] 4.38 [1.35, 14.25] 0.14 [0.01, 2.57] 1.71 [0.18, 16.65] 3.00 [0.13, 71.51] 0.75 [0.21, 2.72] Favours lactulose Favours control Forest plot illustrating the comparison of the clinical efficacy and safety of lactulose with placebo or no intervention. Massachusetts, USA) to evaluate the effect of MHE on HRQoL and the therapeutic effect of lactulose. Changes in the total SIP score (DSIP) presented an improvement or deterioration in overall HRQoL. No heterogeneity was observed between the studies (w 2 = 0.03, P = 0.86, I 2 = 0%). The pooled results show that lactulose

5 1254 European Journal of Gastroenterology & Hepatology 11, Vol 23 No 12 Fig. 3 Study or subgroup Lactulose Control Mean SD Total Mean SD Total Weight Mean difference IV, Fixed, 95% CI year Mean difference IV, Fixed, 95% CI Reduction in blood ammonia levels Xing and Liu [13] % 2.2% 13.9% [ 11.32, 8.88] [ 21.54, 6.46] 7.95 [ 10.96, 4.94] 9.89 [ 11.01, 8.77] Heterogeneity: χ 2 = 2.85, d.f. = 2 (P = 0.70); l 2 = 3% Test for overall effect: Z = (P < ) Reduction in NCT-A time 23 Xing and Liu [13] % 47.6% [ 43.99, 14.01] [.44, 8.96] [ 37.81, 16.10] Heterogeneity: χ 2 = 0.15, d.f. = 1 (P = 0.70); l 2 = 0% Test for overall effect: Z = 4.87 (P < ) Favours lactulose Favours control Forest plot illustrating the comparison of the clinical efficacy of lactulose with placebo or no intervention on blood ammonia levels and number connection test-a time. Fig. 4 Study or subgroup Lactulose Control Mean SD Total Mean SD Total Weight Changes in mean number of abnormal neuropsychological tests Watanabe et al. [8] % Dhiman et al. [11] Prasad et al. [15] 107 Heterogeneity: χ 2 = 4.56, d.f. = 3 (P = 0.21); l 2 = 34% Test for overall effect: Z = (P < ) % 56.5% 25.9% Mean difference IV, Fixed, 95% CI year 1.30 [ 1.85, 0.75] 1.90 [ 2.88, 0.92] 1.92 [ 2.19, 1.65] 1.62 [ 2.02, 1.22] 1.76 [ 1.96, 1.56] Mean difference IV, Fixed, 95% CI Improvement in HRQoL Prasad et al. [15] % 31.9% 6.10 [ 7.13, 5.07] 5.93 [ 7.43, 4.43] 6.05 [ 6.90, 5.] Heterogeneity: χ 2 = 0.03, d.f. = 1 (P = 0.86); l 2 = 0% Test for overall effect: Z = (P < ) Favours lactulose Favours control Forest plot illustrating the comparison of the clinical efficacy of lactulose with placebo or no intervention on mean number of abnormal neuropsychological tests and health-related quality of life. significantly lowered SIP scores and improved HRQoL (WMD: 6.05, 95% CI: 6.30 to 5., P < ) compared with placebo or no intervention (Fig. 4). The improvement in HRQoL was significant in terms of emotional behavior, ambulation, mobility, and sleep in both studies [15,16]. Risk of progression to overt hepatic encephalopathy MHE has been confirmed as an independent risk factor for the subsequent development of OHE [17,18]. Five studies [11,13 16] provided data on the number of patients with MHE that progressed to OHE at the end of treatment. No significant heterogeneity was found among the five studies (w 2 = 0.23, P = 0.99, I 2 = 0%). On the basis of the fixed effects model, the use of lactulose significantly reduces the risk of progression to OHE compared with placebo or no intervention (RR: 0.17, 95% CI: , P = 0.002) (Fig. 2). Incidence of diarrhea None of the patients in all studies reported adverse events related to lactulose treatment, and most of the side-effects presented in the gastrointestinal tract (diarrhea, abdominal pain, flatulence, or nausea). Patients in five studies [8,9,12,13,15] were able to endure two or three semi-soft stools per day. However, two patients in

6 Lactulose for minimal hepatic encephalopathy Luo et al the study by Nie et al. [12] and two patients in the study by withdrew from the treatment because they could not endure the frequent diarrhea. Diarrhea was determined to be the most common sideeffect of lactulose. Four studies [8,9,12,14] reported the number of patients with diarrhea in the lactulose therapy group. No significant heterogeneity was observed among the four studies (w 2 = 0.17, P = 0.98, I 2 = 0%). Compared with placebo or no intervention, lactulose significantly increased the incidence of diarrhea in patients with MHE (RR: 4.38, 95% CI: , P = 0.01) (Fig. 2). Mortality of patients with minimal hepatic encephalopathy The presence of HE was independently related to increased mortality, and there also seemed to be a correlation between MHE and mortality of patients [19,]. Three studies [11,15,16] reported the mortality of patients with MHE. Dhiman et al. [11] reported that two patients died of liver failure in the lactulose therapy group, and one died of massive esophageal varices bleeding. In the study by Mittal et al. [16], one patient died because of acute variceal bleeding with hepatorenal syndrome. Prasad et al. [15] reported that three patients died of diseases related to liver cirrhosis, but did not give the explicit reasons for death. The data were homogenous (w 2 = 2.53, P = 0.28, I 2 = 21%). The meta-analysis also showed no significantly different effect in the mortality of patients with MHE between lactulose and placebo or no intervention (RR: 0.75, 95% CI: , P = 0.66) (Fig. 2). Publication bias assessment Publication bias was analyzed using an inverted funnel plot analysis. Figure 5 showed the inverted funnel plot of RCT for the risk of no improvement in neuropsychological tests with lactulose or placebo therapy. It was approximately symmetrical, which indicated no significant publication bias. Sensitivity analysis Sensitivity analysis was performed because ethnical differences might have influenced the efficacy of lactulose. In the sensitivity analysis, the studies performed in Belgium and in Japan were excluded, and the clinical efficacy of lactulose was evaluated in studies from China and India. Moreover, higher quality studies, with Jadad quality scores of greater than 4, were included to verify the results of the meta-analysis. As shown in Table 2, the result of the sensitivity analysis remained consistent and stable. Discussion Lactulose is a prebiotic that promotes the growth of probiotic bacteria, alters the gut flora, and lowers colonic Fig. 5 SE[log(RR)] RR The inverted funnel plot of randomized controlled trial for the risk of no improvement in neuropsychological tests with lactulose or placebo therapy. RR, relative risk. Table 2 Results of sensitivity analysis Variable Patients (lactulose/control) Pooled RR P All in China 217 (114/103) 0.54 (0.42,0.70) < All in India 167 (85/82) 0.47 (0.36,0.62) < High-quality studies 177 (93/84) 0.52 (0.,0.67) < RR, relative risk. ph, resulting in reduced production and intestinal adsorption of ammonia and consequently, reduced blood ammonia concentrations [21]. According to a large amount of study data and the AASLD survey, lactulose was promulgated as the first and standard treatment for MHE [4]. The meta-analysis also shows that lactulose is significantly beneficial to patients with MHE based on high-quality studies. Lactulose reduces the risk of no improvement in neuropsychological tests and the time required for the completion of NCT-A, lowers the mean number of abnormal neuropsychological tests and blood ammonia levels, improves HRQoL, and prevents the progression to OHE in patients with MHE. However, lactulose increases the incidence of diarrhea and does not lower the mortality of patients with MHE. Significant heterogeneity was not observed among studies, and the fixed effects model was used in the meta-analysis to strengthen the results. The result of sensitivity analysis remained consistent and stable, and no significant publication bias was observed. Therefore, the findings have important clinical implications. The results of the meta-analysis were limited because of different diagnostic criteria for MHE and the composition of tested population in the included studies, especially to the etiology and the severity of liver cirrhosis [22,23].

7 1256 European Journal of Gastroenterology & Hepatology 11, Vol 23 No 12 Firstly, no ideal and uniform standard is currently available for the diagnosis of MHE [1]. All nine studies used different neuropsychological tests to diagnose the disease, but the results of these tests might be influenced by the age of patients, the educational level, and the cultural background, as well as subject to the effect of repeated learning [24,25]. Some included studies [12,14] used precise neurophysiological tests, such as electroencephalogram and evoked potentials, to diagnose MHE. However, their sensitivity was unsatisfactory compared with that of neuropsychological tests, and the specificity could not be fully clarified [26,27]. Therefore, different diagnostic criteria for MHE may induce a little influence on the results of our meta-analysis. Secondly, neuropsychological tests were unsuitable for patients with MHE that result from alcoholic liver cirrhosis because some neuropsychological defects, such as reduced cognitive function and memory deficits, could be aggravated by chronic alcoholism, which might result in bias in the results of neuropsychological tests. Chronic alcoholism also could increase the time required for the completion of NCT-A, thus affecting the diagnosis of MHE [28]. Physicians found difficulty in determining whether neuropsychological defects were directly linked to MHE rather than to chronic alcoholism. However, in all nine studies, patients with MHE attributable to different etiology of liver cirrhosis were randomly assigned to either the lactulose or the control group. Therefore, the inadequate and problematic assignment methods that neglected different etiology of liver cirrhosis could affect the outcomes of the meta-analysis. Thirdly, five studies [9,11,14 16] used the Child Turcotte Pugh (CTP) class or model for end-stage liver disease (MELD) scores to assess the severity of liver cirrhosis, but neither the CTP class nor MELD score precisely evaluated the neuropsychological symptoms of patients with MHE [29]. No judging criterion were presented for HE or MHE in the MELD score system, and the CTP class system only contains the judging criteria for OHE, not MHE. Therefore, the MELD score and CTP class are inappropriate for patients with MHE, and the severity of liver cirrhosis might have influenced the results of the meta-analysis. Prasad et al. [15] and evaluated the improvement of the HRQoL of patients with MHE treated with lactulose. The improvement, which was measured using SIP scores, was significant in terms of emotional behavior, ambulation, mobility, and sleep. However, the result was also affected by the ages of patients, the severity and etiology of liver cirrhosis, and the educational background of patients. Similar to neuropsychological tests, SIP scores are particularly affected by chronic alcoholism, and some neuropsychological symptoms, such as slow reaction and forgetfulness, could not be directly attributed to either MHE or chronic alcoholism. A significant correlation existed between an improved SIP score and the improvement in neuropsychological tests, thus, it could be confirmed that the improvement in HRQoL was related to the improvement in cognitive functions [15]. Moreover, no significant correlation was observed between improved SIP score and improvement in MELD score or CTP class [15,30]. This lack of correlation can be attributed to the fact that MELD scores and CTP class are disease-specific measures for liver cirrhosis, but the SIP questionnaire is non-diseasespecific and it could be used for assessing the effect of different diseases and treatments on daily function [15,30,31]. Therefore, future studies with a larger number of patients should use a new HRQoL questionnaire that should be specific for liver disease. None of the patients in all studies reported adverse events related to the lactulose therapy, but a few patients withdrew from the lactulose therapy group because they could not endure the frequent diarrhea. Thus, long-term treatment with lactulose might increase the severity and frequency of diarrhea and result in hypertonic dehydration with hypernatremia, which could aggravate the mental state of the patient [32]. Probiotics, synbiotics, and rifaximin also showed significant therapeutic effects in the management of MHE and presented higher tolerability and safety along with less gastrointestinal side-effects [33 35]. In the study by Sharma et al. [36], lactulose or probiotics or combinations of both were equally effective in the treatment of MHE. These drugs can be used as an alternative for patients who are lactulose-intolerant, but more RCT are required to confirm their effectiveness. Our meta-analysis showed no significant difference in the mortality of patients with MHE between lactulose and control therapies because most of the expired patients died of esophageal varices bleeding and liver failure, which were not directly associated with MHE. Furthermore, the treatment durations were very short, and none of the included studies followed up patients after the end of treatment [37]. Thus, the correlation between lactulose and the mortality of patients with MHE should be interpreted with caution. The findings of our meta-analysis provide available evidence on the use of lactulose in patients with MHE and have important clinical implications in the management of MHE. However, whether lactulose affects the overall mortality of patients with MHE remains to be investigated in large-scale clinical studies. Treatment options for MHE are evolving, and probiotics, synbiotics, and rifaximin are currently used in its management. Therefore, comparison of the new agents with lactulose is

8 Lactulose for minimal hepatic encephalopathy Luo et al required in further studies to evaluate their clinical efficacy and safety in the management of MHE. Acknowledgements W.-K.C. and C.-Z.L. designed and organized the study protocol. M.L. and L.L. reviewed the papers, extracted and analyzed the data, and wrote the manuscripts. All authors had access to the statistical analyses, approved the final article, and attested to the validity of the results. Conflicts of interest There are no conflicts of interest. References 1 Ferenci P, Lockwood A, Mullen K, Tarter R, Weissenborn K, Blei AT. Hepatic encephalopathy-definition, nomenclature, diagnosis, and quantification: final report of the working party at the 11th World Congresses of Gastroenterology, Vienna, Hepatology 02; 35: Das A, Dhiman RK, Saraswat VA, Verma M, Naik SR. Prevalence and natural history of subclinical hepatic encephalopathy in cirrhosis. J Gastroenterol Hepatol 01; 16: Lockwood AH, Yap EW, Wong WH. Cerebral ammonia metabolism in patients with severe liver disease and minimal hepatic encephalopathy. J Cereb Blood Flow Metab 1991; 11: Bajaj JS, Etemadian A, Hafeezullah M, Saeian K. Testing for minimal hepatic encephalopathy in the United Statess: an AASLD survey. Hepatology 07; 45: Als-Nielsen B, Gluud LL, Gluud C. Non-absorbable disaccharides for hepatic encephalopathy: systematic review of randomized trials. BMJ 04; 328: Moher D, Cook DJ, Eastwood S, Olkin I, Rennie D, Stroup DF, et al. Improving the quality of reports of meta-analyses of randomised controlled trials: the QUOROM statement. Quality of Reporting of Meta-analyses Lancet 1999; 354: Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds DJ, Gavaghan DJ, McQuay HJ. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials 1996; 17: Watanabe A, Sakai T, Sato S, Imai F, Ohto M, Arakawa Y, et al. Clinical efficacy of lactulose in cirrhotic patients with and without subclinical hepatic encephalopathy. Hepatology 1997; 26: Horsmans Y, Solbreux PM, Daenens C, Desager JP, Geubel AP. Lactulose improves psychometric testing in cirrhotic patients with subclinical encephalopathy. Aliment Pharmacol Ther 1997; 11: Li Z, Zhang H, Hong Y. Clinical effect of lactulose in the treatment of subclinical hepatic encephalopathy. Zhongguo Zhong Xi Yi Jie He Za Zhi 1999; 9: Dhiman RK, Sawhney MS, Chawla YK, Das G, Ram S, Dilawari JB. Efficacy of lactulose in cirrhotic patients with subclinical hepatic encephalopathy. Dig Dis Sci 00; 45: Nie YQ, Zheng Z, Yu-yuan L, Sha WH, Ping L, Dai SJ. Long-term efficacy of lactulose in patients with subclinical hepatic encephalopathy. Zhonghua Nei Ke Za Zhi 03; 42: Xing Q, Liu L. Research of lactulose in the treatment of minimal hepatic encephalopathy. World Chinese Journal of Digestion 03; 11: Zeng Z, Li YY. Effect of lactulose treatment on the course of subclinical hepatic encephalopathy. Zhonghua Yi Xue Za Zhi 03; 83: Prasad S, Dhiman RK, Duseja A, Chawla YK, Sharma A, Agarwal R. Lactulose improves cognitive functions and health-related quality of life in patients with cirrhosis who have minimal hepatic encephalopathy. Hepatology 07; 45: Mittal VV, Sharma BC, Sharma P, Sarin SK A. randomized controlled trial comparing lactulose, probiotics and L-ornithine L-aspartate in treatment of minimal hepatic encephalopathy. Eur J Gastroenterol Hepatol 11; 23: Romero-Gómez M, Boza F, García-Valdecasas MS, García E, Aguilar-Reina J. Subclinical hepatic encephalopathy predicts the development of overt hepatic encephalopathy. Am J Gastroenterol 01; 96: Hartmann IJ, Groeneweg M, Quero JC, Beijeman SJ, de Man RA, Hop WC, Schalm SW. The prognostic significance of subclinical hepatic encephalopathy. Am J Gastroenterol 00; 95: Stewart CA, Malinchoc M, Kim WR, Kamath PS. Hepatic encephalopathy as a predictor of survival in patients with end-stage liver disease. Liver Transpl 07; 13: Bustamante J, Rimola A, Ventura PJ, Navasa M, Cirera I, Reggiardo V, Rodés J. Prognostic significance of hepatic encephalopathy in patients with cirrhosis. J Hepatol 1999; 30: Clausen MR, Mortensen PB. Lactulose, disaccharides and colonic flora. Clinical consequences. Drugs 1997; 53: Quero JC, Hartmann IJ, Meulstee J, Hop WC, Schalm SW. The diagnosis of subclinical hepatic encephalopathy in patients with cirrhosis using neuropsychological tests and automated electroencephalogram analysis. Hepatology 1996; 24: Sood GK, Sarin SK, Mahaptra J, Broor SL. Comparative efficacy of psychometric tests in detection of subclinical hepatic encephalopathy in nonalcoholic cirrhotics: search for a rational approach. Am J Gastroenterol 1989; 84: Kato A, Kato M, Ishii H, Ichimiya Y, Suzuki K, Kawasaki H, et al. Development of quantitative neuropsychological tests for diagnosis of subclinical hepatic encephalopathy in liver cirrhosis patients and establishment of diagnostic criteria-multicenter collaborative study in Japanese. Hepatol Res 04; 30: Kircheis G, Fleig WE, Görtelmeyer R, Grafe S, Haussinger D. Assessment of low-grade hepatic encephalopathy: a critical analysis. J Hepatol 07; 47: Saxena N, Bhatia M, Joshi YK, Garg PK, Dwivedi SN, Tandon RK. Electrophysiological and neuropsychological tests for the diagnosis of subclinical hepatic encephalopathy and prediction of overt encephalopathy. Liver 02; 22: Kharbanda PS, Saraswat VA, Dhiman RK. Minimal hepatic encephalopathy: diagnosis by neuropsychological and neurophysiologic methods. Indian J Gastroenterol 03; 22: Edwing D, Flynn L, Klein A, Thuluvath PJ. Cognitive impairment in alcoholic and nonalcoholic cirrhotic patients. Hepatology 1999; 30: Ham J, Gish RG, Mullen K. Model for end-stage liver disease (MELD) exception for hepatic encephalopathy. Liver Transpl 06; 12: Marchesini G, Bianchi G, Amodio P, Salerno F, Merli M, Panella C, et al. Factors associated with poor health-related quality of life of patients with cirrhosis. Gastroenterology 01; 1: Wunsch E, Szymanik B, Post M, Marlicz W, Mydłowska M, Milkiewicz P. Minimal hepatic encephalopathy does not impair health-related quality of life in patients with cirrhosis: a prospective study. Liver Int 11; 31: Warren SE, Mitas JA II, Swerdlin AH. Hypernatremia in hepatic failure. JAMA 1980; 243: Bajaj JS, Saeian K, Christensen KM, Hafeezullah M, Varma RR, Franco J, et al. Probiotic yogurt for the treatment of minimal hepatic encephalopathy. Am J Gastroenterol 08; 103: Liu Q, Duan ZP, Ha DK, Bengmark S, Kurtovic J, Riordan SM. Synbiotic modulation of gut flora: effect on minimal hepatic encephalopathy in patients with cirrhosis. Hepatology 04; 39: Bass NM, Ahmed A, Johnson L, Gardner JD. Rifaximin treatment is beneficial mild hepatic encephalopathy. Hepatology 04; : Sharma P, Sharma BC, Puri V, Sarin SK. An open-label randomized controlled trial of lactulose and probiotics in the treatment of minimal hepatic encephalopathy. Eur J Gastroenterol Hepatol 08; : Amodio P, Del Piccolo F, Marchetti P, Angeli P, Iemmolo R, Caregaro L. Clinical features and survival of cirrhotic patients with subclinical cognitive alterations detected by the number connection test and computerized psychometric tests. Hepatology 1999; 29: