VSL#3 ® probiotic therapy does not reduce portal pressures in patients...

C IRRHOS IS AND L IVER FAILURE

VSL#3� probiotic therapy does not reduce portal pressures in patientswith decompensated cirrhosisSaumya Jayakumar1,2, Michelle Carbonneau2, Naomi Hotte2, A. Dean Befus2, Chris St. Laurent, Richard Owen3,Mairin McCarthy3, Karen Madsen2, Robert J. Bailey2, Mang Ma2, Vince Bain2, Kevin Rioux1 and Puneeta Tandon2

1 Department of Medicine, University of Calgary, Calgary, Alberta, Canada

2 Department of Medicine, University of Alberta, Edmonton, Alberta, Canada

3 Department of Radiology, University of Alberta, Edmonton, Alberta, Canada

Keywords

double blind – hepatic venous portal gradient

– placebo controlled – portal hypertension –

randomized

Correspondence

Dr. Puneeta Tandon, 130 University Campus

NW, Zeidler Ledcor Centre, Edmonton,

Alberta T6G2X8, Canada

Tel: 780 492 9844

Fax: 780 492 9873

e-mail: [email protected]

Received 8 March 2013

Accepted 20 June 2013

DOI:10.1111/liv.12280

Liver Int. 2013: 33: 1470–1477

AbstractBackground & Aims: In patients with decompensated cirrhosis, bacterialtranslocation can contribute to splanchnic vasodilatation, decreased effectivecirculating volume, and portal hypertension. The primary objective of thisrandomized, double blind placebo controlled trial was to evaluate theeffect of the probiotic VSL#3� on the hepatic venous pressure gradient(HVPG). Methods: Seventeen patients with decompensated cirrhosis and anHVPG of ≥10 mmHg were randomized to receive 2 months of VSL#3� oran identical placebo. HVPG, endotoxin, interleukin (IL)-6, IL-8, IL-10, renin,aldosterone, nitric oxide and stool microbiota were measured at baseline andstudy end. Results: Two of the 17 patients were taken off the trial beforecompletion (one for alcohol abuse and the second for SBP – both in placeboarm). Data were analysed on the remaining 15 patients. The median modelfor end-stage liver disease score was 12, and 80% of patients had Child PughB disease. The treatment arm had a greater decrease in HVPG from baselineto study end than the placebo arm (median change from baseline �11.6% vs+2.8%), although this reduction was not statistically significant in eithergroup. There was a significant reduction in the plasma aldosterone level inthe VSL#3� group, but no significant changes in the other measured parame-ters, including the stool microflora analysis. Conclusions: Within the limita-tions of our sample size, VSL#3� therapy does not appear to have asignificant impact on portal pressure reduction in patients with decompen-sated cirrhosis.

Bacterial translocation (BT) is defined as the migrationof bacteria and bacterial products from the intestinallumen to extra-intestinal sites such as the liver and mes-enteric lymph nodes. BT results in the release of pro-inflammatory cytokines [interleukin-6 (IL-6), IL-8],tumour necrosis factor-a (TNF-a), and increased nitricoxide (NO) synthesis (1–10). By contributing to theexacerbation of the hyperdynamic circulatory state(HCS) and increase in hepatic vascular resistance, bacte-ria and their products play a role in increasing portalpressure. Portal hypertension accounts for the majorityof the morbidity and mortality associated with cirrhosis.

Antibiotics are known to modulate BT, their useresulting in a reduction in pro-inflammatory cytokinesand NO, and attenuation of systemic vasodilation (11–13). Unfortunately, the chronic use of antibiotics is lim-ited by antibiotic-associated side effects and by thedevelopment of antibiotic resistant bacterial infections.Given these disadvantages, modulation of the intestinal

microbiota and BT by probiotics has gained interest asan alternative therapeutic approach.

Probiotics are orally administered livemicroorganismsthat “when administered in adequate amounts confer ahealth benefit on the host” (14). Their mechanism ofaction is likely multifactorial, including effects on thecomposition and function of the intestinal microbiota,immune modulation, cytokine production and mainte-nance of the epithelial barrier (5, 15–25). Although resultshave been discordant, in some liver disease studies,VSL#3� (a combination of eight species belonging to Lac-tobacillus, Bifidobacterium and Streptococcus) and otherprobioticmixtures have been associatedwith an improve-ment in intestinal permeability, BT and pro-inflamma-tory cytokine release (26–29). In select clinical studies, theuse of probiotics has resulted in reduced bacterial infec-tions after liver transplantation (30, 31), reduced hepaticencephalopathy (32–34) and in certain groups, animprovement in the Child Pugh (CP) class (29, 34).

Liver International (2013)© 2013 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd1470

Liver International ISSN 1478-3223

The published evidence evaluating the impact ofprobiotic therapy on portal hypertension in cirrhosis islimited. In 2008, our group performed an uncontrolledpilot study of the impact of 2 months of VSL#3� inpatients with predominantly compensated cirrhosis(35). We demonstrated a significant reduction in theplasma aldosterone level and confirmed the safety andtolerability of VSL#3�. The use of VSL#3� in that studydid not, however, impact the portal pressure as mea-sured by the hepatic venous pressure gradient (HVPG).We hypothesized that the negligible effect on portalpressure may have been attributable to the predomi-nantly compensated status of the patients (median CPscore of 6) and therefore the lack of significant baselineperturbation in the HCS. It is known that BT occursmost often in the setting of advanced liver dysfunction(<5% in CP-A cirrhosis and up to 30% in CP-C cirrho-sis) (4, 5, 36). Accordingly, this study is a randomizedcontrolled trial to evaluate the effect of VSL#3� on theHVPG and associated physiological parameters indecompensated patients with CPB or CPC cirrhosis.

Materials and methods

This was an investigator-initiated study of patients fromtwo hospitals (University of Alberta and Royal Alexan-dra) in Edmonton, Alberta, Canada. Study approval wasobtained from the local research ethics board andHealth Canada. This study was registered at ClinicalTri-als.gov (Study ID: NCT01032941). Written consent wasobtained from all patients.

Patients

Inclusion criteria were as follows: age ≥18 years, cirrho-sis diagnosed using either radiological or histologicalcriteria, CP class B or C and clinically significant portalhypertension (HVPG ≥10 mm Hg).

Exclusion criteria were hepatocellular carcinomabeyond Milan criteria, transjugular intrahepatic porto-systemic shunt or surgical portosystemic shunt, portalvein thrombosis, ongoing bacterial infection, use of anti-biotics within the 2 weeks prior to enrolment, GradeIII–IV hepatic encephalopathy, gastrointestinal bleedingwithin the previous 2 weeks, a history of myocardialinfarction, cerebrovascular accident, or arrhythmia,active drug or alcohol use and failure to consent.

Intervention

Patients who met all of the above criteria were random-ized to receive two sachets twice a day of either placeboor VSL#3� for 8 weeks. Each sachet of VSL#3� contains900 billion viable bacteria (totalling 3600 billion bacte-ria daily), with four strains of Lactobacillus, three strainsof Bifidobacterium and Streptococcus salivarus sp thermo-philus. The placebo agent contains cornstarch and wasidentical in appearance and packaging as the active

agent. Both agents were supplied in kind by VSL Phar-maceuticals Inc. (Fort Lauderdale, FL, USA), but thesupplier had no input into study design, data analysis,or manuscript preparation. Both the patients and thestudy investigators were blinded to the interventiondrug. Randomization was carried out by the hospitalpharmacy using a computer randomization planobtained from http:///www.randomization.com.

Outcome measures

The primary outcome measure was reduction in HVPGfrom baseline reading. Secondary outcome measureswere changes in plasma renin, plasma aldosterone,serum IL-6, IL-8, bacterial endotoxin and NO levels.Hepatic venous pressure gradient measurements wereobtained in the Interventional Radiology departmentusing established methodology (37, 38). After ultra-sound-guided puncture of the internal jugular vein andfluoroscopy-guided access to the hepatic vein, a bal-loon-tipped catheter was exchanged over a wire andattached to a MLT0670 Disposable BP Transducer andML117 BP Amp (ADInstruments, Inc., ColoradoSpring, Colorado, USA). Pressure tracings were recordedusing PowerLab 4/25 computer software (ADInstru-ments, Inc.). At least three wedged hepatic vein pressureand three free hepatic vein pressure measurements wereobtained in each patient. The HVPG was calculated bysubtracting the hepatic venous pressure from thewedged pressure measurement and the average of thethree measurements was taken for this study outcome.HVPG measurements were obtained both at the start ofthis study, as well as at the week 8 endpoint, the dayafter the patient had taken the last sachet of VSL#3� orplacebo. To decrease operator variability, two interven-tional radiologists (R. O., M. M.) performed the major-ity of the measurements.

Stool samples were collected for analysis of microbiotaby terminal restriction fragment polymorphism(TRFLP), as previously described (39). Briefly, totalDNA was extracted from 200 to 250 mg aliquots of stoolusing the QIAamp DNA Stool Mini Kit (Qiagen, Missis-sauga, ON, Canada) with a beat-beating step (40). Opti-mal quality TRFLP profiles were obtained using 2–30 ngof template DNA. PCR was performed with primers 8f(5′-AGAGTTTGATCCTGGCTCAG-3′) labelled with 6-FAM and 926r (5′-CCGTCAATTCCTTTRAGT TT-3′),which target part of the 16s rRNA gene (41), with pooledand purified product being digested with 10 U of HpaII(Invitrogen, Burlington, ON, Canada). Terminal restric-tion fragment analysis was performed using an ABI3730XL Genetic Analyser (University of Calgary, CoreDNA Services) and GeneMapper software (ABI Biosys-tems, Streetsville, ON, Canada). Restriction fragmentswere tabulated and analysed by principal componentanalysis using SIMCA-P (v. 11.5) multivariate statisticssoftware (Umetrics, Ume�a, Sweden) to identify intrinsicclusters within the data set and the key bacterial phylo-

Liver International (2013)© 2013 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd 1471

Jayakumar et al. Probiotic therapy for portal hypertension

types that determine such clustering. Orthogonal-projec-tion to latent structure-discriminant analysis (O-PLSDA) was carried on TRFLP binary data (i.e. peakpresence or absence) using unit variance scaling. OPL-Scoefficients were used to visualize the major variablescontributing to the discrimination in the model. Themodel was subject to seven-fold cross-validation, an iter-ative process whereby one-seventh of the samples wereexcluded and used to predict fit into the modified model.The Q2 value represents how accurately the model pre-dicts the data set.

On the morning of the pressure measurements (bothbaseline and end of study), patient vital signs, plasmarenin, aldosterone, cytokine levels, endotoxin levels,liver enzymes and model for end-stage liver disease(MELD) labs were obtained. The plasma renin andaldosterone were obtained after the patient had been ina supine position for at least 2 h, and were measuredusing commercial assays available in the hospital lab,namely the Cis Bio-coated tube radioimmunoassay andCoat-a-Count no extraction radioimmunoassay respec-tively (Diagnostic Products Corp., Los Angeles, CA,USA). Levels of IL-6, IL-8 and IL-10 were measuredusing enzyme-linked immunosorbent assay kits (Medi-corp, Montreal, Quebec). Plasma endotoxin levels weremeasured using a Limulus amebocyte lysate-basedchromatographic assay as previously described (35).Nitrite and nitrate levels were measured using theParameterTM Total NO/Nitrite/Nitrate kit (R&D Sys-tems, Minneapolis, MN, USA). Plasma samples werefirst deproteinized using 10 KDa molecular weightcut-off filters (Millipore, Billerica, MA, USA). Nitrateconcentrations were obtained by subtracting themeasured nitrite concentration from the totalnitrate + nitrite concentrations.

Patients had routine blood work to assess both liverfunction as well as liver enzymes at the 2, 4 and 6 weekmark. At the 4-week mark, they were seen in clinic andassessed for complications, drug tolerance and compli-ance. At 8 weeks, they returned for a final pressure mea-surement, and the blood work drawn at baseline wasrepeated.

Statistical analysis and sample size calculation

Statistical analysis was carried out using SPSS version 19(SPSS Inc., Chicago, IL, USA). Variables were describedusing the number and proportion or using median andinterquartile range (IQR). Given the small sample size,the non-parametric Wilcoxon signed rank test was usedto assess whether there were significant changes frombaseline to study end within each group. Significancewas established at a two-tailed P-value of less than 0.05.

The sample size calculation was performed usingHVPG data from our trial in compensated cirrhosis(35). We assumed that there would be a 0% reductionin the HVPG in the placebo group. A conservativestandard deviation of change of 14% was estimated

using standard deviations from other published HVPGintervention studies (42–44) (standard deviations ran-ged from 3 to 12%) and from our trial in compensatedcirrhosis (18%) (35). Therefore, to detect a 20% reduc-tion in HVPG in the intervention group with an alphaof 0.05 and a power of 80%, eight patients were neededin each group (placebo and VSL#3�). To account for a10% patient dropout rate, the sample size was inflatedto 20 patients.

Results

Baseline characteristics

Seventeen patients with decompensated cirrhosis wereenrolled between October 2009 and May 2011, and ran-domly assigned to either placebo or VSL#3�. Of the 17,two patients were removed from this study prior tocompletion (one owing to heavy ongoing alcohol abuse,and the second because of an episode of spontaneousbacterial peritonitis and progressive renal dysfunction).Both these patients were on placebo. The data wereanalysed for the remaining 15 patients (7 on VSL#3�

and 8 on placebo).The majority of patients were men (n = 12; 80%),

and had alcoholic cirrhosis (n = 11; 73%). The medianage was 53 (IQR: 48–55) and a median MELD score of12. Thirteen patients had CPB disease and two had CPCdisease (both patients of the latter group were in the pla-cebo arm). All 17 patients were on diuretics for controlof ascites, the median dose of which was not signifi-cantly altered for either group during the study period.Five of the seven patients on VSL#3� and four of theeight patients on placebo were on lactulose for the treat-ment of hepatic encephalopathy.

Baseline clinical characteristics of the patientsbetween the two groups were the same with respect toage, gender, aetiology of cirrhosis, MELD scores and CPclassification. Lactulose, b-blocker and diuretic use werealso similar between the groups (Table 1).

Primary outcome measure

The median baseline HVPG was lower in the VSL#3�

group [18.0 (IQR: 14.4–18.8) mmHg] than in the pla-cebo group [22.3 (IQR: 20.9–24.2) mmHg], P = 0.001.Comparisons made within each group (Table 2) did notshow significant changes in the HVPG from baseline tostudy end in patients receiving placebo [median changeof 2.8% (IQR: �11.8 to 8.8%] or VSL#3� [medianchange of �11.6% (IQR: �20.2 to 5.4%). Figure 1 dem-onstrates individual patient HVPG data.

Secondary outcome measures

The probiotic group demonstrated a small but statis-tically significant, reduction in plasma aldosteronelevels from study baseline to study end (1545.0–


Probiotic therapy for portal hypertension Jayakumar et al.

1525.0 pmol/L; P = 0.046) compared to the placebogroup (2205.5–2033.5 pmol/L; P = 0.67). There wasno difference in levels of plasma renin, serum cytokines(IL-6, IL-8, or IL-10), or in clinical parameters such asthe systolic blood pressure or body weight. Neithergroup demonstrated a change in liver function, eitherclinically or based on blood work.

Supervised OPL-S discriminant analysis of TRFLPprofiles from VSL#3� -treated control patients showedno significant clustering, which indicates that the faecalmicrobiota of each study group at baseline was similar.Despite 8 weeks of treatment with VSL#3�, OPL-S dis-criminant analysis did not show any significant cluster-ing of the faecal microbiota of probiotic-treated patientsfrom controls. When discriminant analysis was per-formed within the VSL#3� group, the faecal microbiotaalso could not reliably distinguish pre- and post-treat-ment states. The Q2 value of the latter model was �0.08,indicating that faecal TRFLP profiles could not reliablyclassify pre- from post-VSL#3� treatment states in thesame patients. Interestingly, OPL-S coefficients in thismodel showed that TRFs corresponding to Lactobacilliand Streptococcus species (main components ofVSL#3�) were increased after probiotic treatment.

Overall, VSL#3� and placebo were well tolerated. Theonly adverse event was in a patient in the placebo groupwho developed spontaneous bacterial peritonitis andrenal dysfunction. This patient was removed from theanalysis, as he was unable to complete this study. Com-

pliance with therapy was >80% as determined by patientreport (phone calls to patients every 2 weeks), clinic vis-its at weeks 4 and 8 and return of unused product atstudy end.

Discussion

This randomized double blind placebo-controlled trialof patients with cirrhosis is the second to evaluate theimpact of probiotic therapy on the HVPG and the firstto do so in patients with decompensated disease. Inaddition to performing the gold-standard HVPG mea-surements, this study is unique in that it involves a com-prehensive evaluation of secondary outcomes includingliver function, plasma cytokines, NO, plasma aldoste-rone, plasma renin and stool microbiota. These second-ary outcome parameters have been evaluated in otherclinical studies, with discordant results (27, 29, 34, 45,46). Although there was a mild but significant reductionin the plasma aldosterone in the active treatment group,there were no significant changes in the other measuredoutcomes, including the HVPG. Moreover, althoughthere have been concerns about the safety of probioticsin patients who are immunocompromised (47), overthe 8-week study period, there were no significantadverse events noted.

We did not identify a significant reduction in theHVPG after 2 months of VSL#3� therapy. We are una-ware of any other published studies that have evaluatedprobiotic-related portal pressure reduction in decom-pensated cirrhosis. As antibiotics and probiotics sharesimilar pathways for the modulation of intestinal flora,however, we can extrapolate the findings of studies eval-uating the impact of antibiotic therapy on the HVPG. Astudy by Albillos et al. (13) found that modulation ofthe intestinal flora with antibiotics improved the HCSand reduced pro-inflammatory cytokines only in thesubgroup of patients with a baseline elevation of theirserum lipopolysaccharide binding protein (LBP). Inaddition, despite the improvement in the HCS inpatients with high baseline LBP levels, antibiotic therapyin this subgroup did not impact the HVPG. The authorshypothesized that the antibiotic mediated reduction inNO levels had not only led to a reduction in portalinflow but also a concurrent increase in hepatic vascularresistance negating the portal pressure benefit. Twoother antibiotic studies support the lack of change inthe HVPG despite variable levels of improvement in theHCS (12, 13, 48). This lack of HVPG reduction withantibiotics would lend support to our findings that theprobiotic VSL#3� has minimal to no effect on portalpressures in cirrhosis.

The statistically significant reduction in the plasmaaldosterone identified in our active treatment groupmay reflect an improvement in the HCS. This wouldfit with the results of selected probiotic, synbiotic andantibiotic studies in the setting of cirrhosis (11, 12, 22,26–29, 35, 48) and with the findings of this study in

Table 1. Baseline clinical characteristics*

Placebo (n = 8) VSL#3� (n = 7)

Median age(Interquartilerange = IQR)

53.5 (50.0–55.0) 50.0 (46.0–58.0)

Gender, male/female 7/1 5/2Aetiology, alcohol/HCV 6/2 5/2History of ascites, n 8 7Patients receivingß-adrenergic blockers, n

1 1

Patients receivingdiuretics, n

8 7

Patients receivinglactulose, n

4 5

Child Pugh class, B/C 6/2 7/0Median model forend-stageliver disease score (IQR)

13.5 (10–14.8) 11 (6–13)

Median albumin (IQR) (g/L) 34 (28–35) 34 (33–35)Median creatinine(IQR) (lmol/L)

76 (67.3–93.8) 67 (63–82)

Median plateletcount (IQR) (9106)

96 (76–52) 94 (70–109)

Median INR (IQR) 1.5 (1.3–1.5) 1.2 (1.1–1.4)Median bilirubin(IQR) (lmol/L)

52 (34.3–70.3) 29 (19.3–40)

*There were no statistically significant differences between the two

groups, all P > 0.05.



compensated cirrhosis (35). Probiotics are hypothesizedto stabilize the intestinal epithelial barrier, decrease BTand thereby reduce systemic endotoxemia, the release ofpro-inflammatory cytokines and NO, all contributors tothe HCS (28, 29, 34). Importantly, despite being statisti-cally significant, the reduction in plasma aldosteroneseen in this study was small and was not associatedwith a clinically meaningful change in ascites, diureticrequirements or a significant reduction in the plasmarenin level. Moreover, it was not associated with animprovement in liver function as measured by the CPscore.

Several study limitations must be acknowledged.Most notably, the sample size in this study was smalland therefore the potential for Type 2 error exists notonly for the primary but also for the secondary out-comes. For the primary outcome, carrying forward datafrom this study (control group mean HVPG reduction

1%, VSL#3 group mean HVPG reduction 7%, commonstandard deviation 14.2%), it would take an estimatedsample size of 178 patients to detect a statistically signif-icant difference between the two therapies. This largenumber lends support to the conclusion that the use ofVSL#3� is unlikely to lead to a clinically significantreduction in portal pressure. Secondly, despite a dose ofVSL#3� therapy that has previously shown changes instool microflora (49, 50), using the methodology avail-able to us, we did not identify any widespread or signifi-cant alterations in the stool microbiota that enableddistinction of the treatment and control groups at theend of this study by discriminant analysis. It is notable,however, that Lactobacillus and Streptococcus specieswere more common in patients after administration ofVSL#3� although not statistically significant in our mul-tivariate analyses. These findings may either be relatedto a true lack of change in the microflora (reflecting

Table 2. Blood tests and hepatic venous pressure gradient (HVPG) changes before and after placebo and VSL#3� (expressed as mediansand interquartile ranges)

Placebo (n = 8) VSL#3� (n = 7)

Baseline Study end P-value Baseline Study end P-value

Child Pughscore

8.5 (8–10) 8.5 (8–9.8) 1.0 8 (7–9) 7 (7–8) 0.32

Model forend-stageliver disease(MELD)

13.5 (10–14.8) 11.5 (9.5–12.8) 0.27 11 (6–13) 10 (8–11) 0.86

Change inMELD

�8.6 (�34.2 to 15.0) �10 (�15.4 to 33.3)

HVPG (mmHg) 22.3 (20.9–24.2) 22.7 (18.8–26.4) 1.0 18.0 (14.4–18.8) 15.0 (13.0–19.6) 0.27Change inHVPG (%)

2.8 (�11.8 to 8.8) �11.6 (�20.2 to 5.4)

Plasmaaldosterone(pmol/L)

2205.5 (946.5–3111.3) 2033.5 (773.5–5775.0) 0.67 1545.0 (810.0–4503.0) 1525.0 (898.0–3815.0) 0.046

Plasma renin(ng/L)

212.5 (144.0–468.8) 232.0 (139.5–482.5) 0.40 63.5 (46.3–1343.3) 98.0 (50.0–2380.0) 0.12

Endotoxin (OD) 0.11 (0.0–2.21) 1.6 (0.03–7.3) 0.24 0.0 (0.0–2.4) 0.5 (0.0–2.0) 0.59(n = 5)

Nitrate (lM/L) 21.7 (11.8–28.4) 33.6 (24.1–79.7) 0.13 26.0 (15.4–38.4) 44.9 (23.9–99.3) 0.12IL6 (pg/ml) 29.4 (8.5–82.4) 32.9 (8.0–55.5) 0.89 7.2 (2.0–59.7) 6.3 (1.7–212.9) 0.69

(n = 5)IL8 (pg/ml) 66.8 (22.1–372.6) 161.9 (33.9–406.0) 0.89 33.7 (22.8–331.0) 34.1 (22.8–737.2) 0.23

(n = 5)IL10 (pg/ml) 20.9 (3.4–275.4) 50.7 (12.0–172.7) 0.58 5.7 (0.0–149.0) 13.4 (0.9–270.8) 0.47

(n = 5)Systolic bloodpressure

110 (104–116) 104 (94–111) 0.11 105.5 (92.5–118.5) 100.0 (95.5–110.5) 0.47

Weight (kg) 75.0 (58.0–94.0) 75 (62.1–90.9) 0.60 64.5 (53.2–74.3) 73.1 (71.9–75.1) 0.59Albumin (g/L) 34 (28–35) 34 (29–34) 0.67 34 (33–35) 36 (32–38) 0.09Bilirubin(lmol/L)

52.0 (34.3–70.3) 39.5 (32.5–44.5) 0.05 29 (18–43) 25 (22–82) 0.45

Creatinine(lmol/L)

76.0 (67.3–93.8) 80.0 (49.0–97.0) 0.74 67 (63–82) 85 (73–85) 0.18

INR 1.5 (1.3–1.5) 1.4 (1.3–1.5) 0.59 1.2 (1.2–1.4) 1.2 (1.2–1.3) 0.66Plateletcount (9106)

96.5 (47.3–118.5) 98 (53.5–149.3) 0.12 94 (70–109) 82 (68–107) 0.24



inherent resistance of the distal gut microbiome ofpatients with advanced cirrhosis) or to limitations withthe chosen TRFLP technology. TRFLP is known todetect larger compositional shifts in the dominantmicrobiota, but may be insensitive to changes in rarebacterial taxa that comprise less than 1% of the stoolbacterial community (51). Moreover, the exogenousLactobacillus, Bifidobacterium and Streptococcus spe-cies present in VSL#3� may have been obscured fromdetection by the presence of indigenous members ofthese genera (49). Analyses of the microbiota that arequantitative and/or provide data at a deeper taxonomiclevel may resolve these matters in future studies. A thirdstudy limitation is that because other probiotic mixturescontain unique combinations of microorganisms, ourfindings cannot be generalized to all probiotics. More-over, it is not clear whether the optimal dose and dura-tion of therapy were tested. In this study, the 2-monthduration was based on favourable preliminary observa-tions in our original 2-month pilot study and in a studyby Liu et al. of only 1 month in duration (34, 35). Sixtyper cent of our patients were on Lactulose, a recognizedprebiotic. It is thereby possible that patients did getsome synbiotic effect. As a result of small numbers andvariable dosing of lactulose, the implications of this pre-biotic on stool microflora or portal pressure cannot,however, be meaningfully analysed. Future studies willbe needed to evaluate this prebiotic probiotic combina-tion. Lastly, the lack of portal pressure reduction in thisstudy cannot be generalized to mean that probiotic ther-apy has no benefit on other chronic liver failure compli-cations. As evidenced by other studies, probiotic

therapy has been associated with some promisingresults, particularly in the setting of hepatic encephalop-athy (46, 52).

In conclusion, the use of VSL#3� probiotic therapy iswell tolerated in a group of patients with decompensat-ed cirrhosis. Within the limitations of our small samplesize, the findings from this study and those from ourearlier study do not support the use of VSL#3� probiot-ic therapy for the reduction in portal pressure inpatients with compensated or decompensated cirrhosis.

Acknowledgements

Financial support: Dr. Tandon received funding fromthe University of Alberta Hospital Foundation to carryout this research. The authors are also thankful toVSL#3� Pharmaceuticals Inc. (Fort Lauderdale, Florida)for supplying both active drug and placebo in kind.

Conflict of interest: The authors do not have anydisclosures to report.

References

1. Moller S, Hobolth L, Winkler C, Bendtsen F, ChristensenE. Determinants of the hyperdynamic circulation and cen-tral hypovolaemia in cirrhosis. Gut 2011; 60: 1254–9.

2. Wiest R, Das S, Cadelina G, et al. Bacterial translocationin cirrhotic rats stimulates eNOS-derived NO productionand impairs mesenteric vascular contractility. J Clin Invest1999; 104: 1223–33.

3. Guarner C, Soriano G, Tomas A, et al. Increased serumnitrite and nitrate levels in patients with cirrhosis: rela-tionship to endotoxemia. Hepatology 1993; 18: 1139–43.

4. Garcia-Tsao G, Lee FY, Barden GE, Cartun R, West AB.Bacterial translocation to mesenteric lymph nodes isincreased in cirrhotic rats with ascites. Gastroenterology1995; 108: 1835–41.

5. Cirera I, Bauer TM, Navasa M, et al. Bacterial translo-cation of enteric organisms in patients with cirrhosis.J Hepatol 2001; 34: 32–7.

6. Runyon BA, Squier S, Borzio M. Translocation of gut bac-teria in rats with cirrhosis to mesenteric lymph nodes par-tially explains the pathogenesis of spontaneous bacterialperitonitis. Hepatology 1994; 26: 1372–8.

7. Garcia-Tsao G. Bacterial translocation: cause or conse-quence of decompensation in cirrhosis? J Hepatol 2001;34: 150–5.

8. Wiest R, Cadelina G, Milstien S, et al. Bacterial transloca-tion up-regulates GTP-cyclohydrolase I in mesenteric vas-culature of cirrhotic rats. Hepatology 2003; 38: 1508–15.

9. Petros A, Bennett D, Vallance P. Effect of nitric oxide syn-thase inhibitors on hypotension in patients with septicshock. Lancet 1991; 338: 1557–8.

10. Vallance P, Moncada S. Hyperdynamic circulation in cir-rhosis: a role for nitric oxide? Lancet 1991; 337: 776–8.

11. Chin-Dusting JP, Rasaratnam B, Jennings JL, Dudley FJ.Effect of fluoroquinolone on the enhanced nitric oxide-induced peripheral vasodilation seen in cirrhosis. AnnIntern Med 1997; 127: 985–8.

12. Rasaratnam B, Kaye D, Jennings G, Dudley F, Chin-Dust-ing J. The effect of selective intestinal decontamination on

HV

PG

(m

mH

g)

5

10

15

20

25

Baseline End of study

No treatment

Baseline End of study

VSL#3

Fig. 1. Individual patient data for hepatic venous pressure gradient(HVPG) results from baseline to end of study. There was no statisti-cally significant change in the HVPG from baseline to study end inpatients receiving placebo (median change of 2.8%) or VSL#3�

(median change of �11.6%).



the hyperdynamic circulatory state in cirrhosis. A random-ized trial. Ann Intern Med 2003; 139: 186–93.

13. Albillos A, de la Hera A, Gonzalez M, et al. Increased lipo-polysaccharide binding protein in cirrhotic patients withmarked immune and hemodynamic derangement. Hepa-tology 2003; 37: 208–17.

14. Berg RD. Inhibition of Escherichia coli translocation fromthe gastrointestinal tract by normal cecal flora in gnotobi-otic or antibiotic-decontaminated mice. Infect Immun1980; 29: 1073–81.

15. Madsen KL, Doyle JS, Jewell LD, Tavernini MM, FedorakRN. Lactobacillus species prevents colitis in interleukin 10gene-deficient mice. Gastroenterology 1999; 116: 1107–14.

16. Madsen K, Cornish A, Soper P, et al. Probiotic bacteriaenhance murine and human intestinal epithelial barrierfunction. Gastroenterology 2001; 121: 580–91.

17. Boirivant M, Strober W. The mechanism of action of pro-biotics. Curr Opin Gastroenterol 2007; 23: 679–92.

18. Qin HL, Shen TY, Gao ZG, et al. Effect of lactobacillus onthe gut microflora and barrier function of the rats withabdominal infection.World J Gastroenterol 2005; 11: 2591–6.

19. Resta-Lenert S, Barrett KE. Probiotics and commensalsreverse TNF-alpha- and IFN-gamma-induced dysfunctionin human intestinal epithelial cells. Gastroenterology 2006;130: 731–46.

20. Yan F, Cao H, Cover TL, et al. Soluble proteins producedby probiotic bacteria regulate intestinal epithelial cell sur-vival and growth. Gastroenterology 2007; 132: 562–75.

21. DA Hernandez IG, Vorackova FV, Uribe M. Abnormalitiesin proximal small bowel motility in patients with cirrhosis.Ann Hepatol 2008; 7: 120.

22. Rasaratnam B, Connelly N, Chin-Dusting J. Nitric oxideand the hyperdynamic circulation in cirrhosis: is there arole for selective intestinal decontamination? Clin Sci2004; 107: 425–34.

23. Yan F, Polk DB. Probiotic bacterium prevents cytokine-induced apoptosis in intestinal epithelial cells. J Biol Chem2002; 277: 50959–65.

24. Tien MT, Girardin SE, Regnault B, et al. Anti-inflamma-tory effect of Lactobacillus casei on Shigella-infectedhuman intestinal epithelial cells. J Immunol 2006; 176:1228–37.

25. Zyrek AA, Cichon C, Helms S, et al. Molecular mecha-nisms underlying the probiotic effects of Escherichia coliNissle 1917 involve ZO-2 and PKCzeta redistributionresulting in tight junction and epithelial barrier repair. CellMicrobiol 2007; 9: 804–16.

26. Chiva M, Soriano G, Rochat I, et al. Effect of Lactobacillusjohnsonii La1 and antioxidants on intestinal flora andbacterial translocation in rats with experimental cirrhosis.J Hepatol 2002; 37: 456–62.

27. Loguercio C, Federico A, Tuccillo C, et al. Beneficialeffects of a probiotic VSL#3 on parameters of liver dys-function in chronic liver diseases. J Clin Gastroenterol2005; 39: 540–3.

28. Ewaschuk J, Endersby R, Thiel D, et al. Probiotic bacteriaprevent hepatic damage and maintain colonic barrierfunction in a mouse model of sepsis. Hepatology 2007; 46:841–50.

29. Lata J, Novotny I, Pribramska V, et al. The effect of probi-otics on gut flora, level of endotoxin and Child-Pugh scorein cirrhotic patients: results of a double-blind randomizedstudy. Eur J Gastroenterol Hepatol 2007; 19: 1111–3.

30. Rayes N, Seehofer D, Hansen S, et al. Early enteral supplyof lactobacillus and fiber versus selective bowel decontami-nation: a controlled trial in liver transplant recipients.Transplantation 2002; 74: 123–7.

31. Rayes N, Seehofer D, Theruvath T, et al. Supply ofpre- and probiotics reduces bacterial infection rates afterliver transplantation–a randomized, double-blind trial.Am J Transplant 2005; 5: 125–30.

32. Malaguarnera M, Greco F, Barone G, et al. Bifidobacteri-um longum with fructo-oligosaccharide (FOS) treatmentin minimal hepatic encephalopathy: a randomized, dou-ble-blind, placebo-controlled study. Dig Dis Sci 2007; 52:3259–65.

33. Boca M, Vyskocil M, Mikulecky M, et al. [Complex ther-apy of chronic hepatic encephalopathy supplemented withprobiotic: comparison of two studies]. Casopis lekaru ces-kych. 2004; 143: 324–8. Komplexna liecba chronickej he-patalnej encefalopatie doplnena probiotikom. Porovnaniedvoch studii.

34. Liu Q, Duan ZP, Ha d, et al. Synbiotic modulation of gutflora: effect on minimal hepatic encephalopathy in patientswith cirrhosis. Hepatology 2004; 39: 1441–9.

35. Tandon P, Moncrief K, Madsen K, et al. Effects of probi-otic therapy on portal pressure in patients with cirrhosis: apilot study. Liver Int 2009; 29: 1110–5.

36. Llovet JM, Bartoli R, March F, et al. Translocated intesti-nal bacteria cause spontaneous bacterial peritonitis in cir-rhotic rats: molecular epidemiologic evidence. J Hepatol1998; 28: 307–13.

37. Bosch J, Abraldes JG, Berzigotti A, Garcia-Pagan JC. Theclinical use of HVPG measurements in chronic liver dis-ease. Nat Rev Gastroenterol Hepatol 2009; 6: 573–82.

38. Groszmann RJ, Wongcharatrawee S. The hepatic venouspressure gradient: anything worth doing should be doneright. Hepatology 2004; 39: 280–3.

39. Andrews CN, Griffiths TA, Kaufman J, et al. Mesalazine(5-aminosalicylic acid) alters faecal bacterial profiles, butnot mucosal proteolytic activity in diarrhoea-predominantirritable bowel syndrome. Aliment Pharmacol Ther 2011;34: 374–83.

40. Amon JJ, Devasia R, Xia G, et al. Molecular epidemiologyof foodborne hepatitis a outbreaks in the United States,2003. J Infect Dis 2005; 192: 1323–30.

41. Nainan OV, Armstrong GL, Han XH, et al. Hepatitis amolecular epidemiology in the United States, 1996-1997:sources of infection and implications of vaccination pol-icy. J Infect Dis 2005; 191: 957–63.

42. Debernardi-Venon W, Martini S, Biasi F, et al. AT1receptor antagonist Candesartan in selected cirrhoticpatients: effect on portal pressure and liver fibrosismarkers. J Hepatol 2007; 46: 1026–33.

43. Gonzalez-Abraldes J, Albillos A, Banares R, et al. Random-ized comparison of long-term losartan versus propranololin lowering portal pressure in cirrhosis. Gastroenterology2001; 121: 382–8.

44. Abraldes JG, Albillos A, Banares R, et al. Simvastatin low-ers portal pressure in patients with cirrhosis and portalhypertension: a randomized controlled trial. Gastroenterol-ogy 2009; 136: 1651–8.

45. Stadlbauer V, Mookerjee RP, Hodges S, et al. Effect ofprobiotic treatment on deranged neutrophil function andcytokine responses in patients with compensated alcoholiccirrhosis. J Hepatol 2008; 48: 945–51.



46. Bajaj JS, Saeian K, Christensen KM, et al. Probiotic yogurtfor the treatment of minimal hepatic encephalopathy. AmJ Gastroenterol 2008; 103: 1707–15.

47. Morrow LE, Gogineni V, Malesker MA. Synbiotics andprobiotics in the critically ill after the PROPATRIA trial.Curr Opin Clin Nutr Metabol Care 2012; 15: 147–50.

48. Kemp W, Colman J, Thompson K, et al. Norfloxacintreatment for clinically significant portal hypertension:results of a randomised double-blind placebo-controlledcrossover trial. Liver Int 2009; 29: 427–33.

49. Bibiloni R, Fedorak RN, Tannock GW, et al. VSL#3 probi-otic-mixture induces remission in patients with activeulcerative colitis. Am J Gastroenterol 2005; 100: 1539–46.

50. Venturi A, Gionchetti P, Rizzello F, et al. Impact onthe composition of the faecal flora by a new probiotic

preparation: preliminary data on maintenance treatmentof patients with ulcerative colitis. Aliment PharmacolTher 1999; 13: 1103–8.

51. Blackwood CB, Hudleston D, Zak DR, Buyer JS. Interpret-ing ecological diversity indices applied to terminal restric-tion fragment length polymorphism data: insights fromsimulated microbial communities. Appl Environ Microbiol2007; 73: 5276–83.

52. Sharma P, Sharma BC, Puri V, Sarin SK. An open-labelrandomized controlled trial of lactulose and probiotics inthe treatment of minimal hepatic encephalopathy. Eur JGastroenterol Hepatol 2008; 20: 506–11.



VSL#3 ® probiotic therapy does not reduce portal pressures in patients...

Documents

Transcript of VSL#3 ® probiotic therapy does not reduce portal pressures in patients...