Modification of intestinal flora withmultispecies probiotics reducesbacterial translocation and improvesclinical course in a rat model ofacute pancreatitisL. Paul van Minnen, MD, PhD,a Harro M. Timmerman, PhD,a,d Femke Lutgendorff, MD,a

André Verheem, BSc,a Wil Harmsen, BSc,b Sergey R. Konstantinov, PhD,c Hauke Smidt, PhD,c

Maarten R. Visser, MD, PhD,b Ger T. Rijkers, PhD,d,e Hein G. Gooszen, MD, PhD,a andLouis M. A. Akkermans, PhD,a Utrecht, Wageningen, Nieuwegein, The Netherlands

Background. Infection of pancreatic necrosis by gut bacteria is a major cause of morbidity andmortality in patients with severe acute pancreatitis. Use of prophylactic antibiotics remainscontroversial. The aim of this experiment was assess if modification of intestinal flora with specificallydesigned multispecies probiotics reduces bacterial translocation or improves outcome in a rat model ofacute pancreatitis.Methods. Male Sprague-Dawley rats were allocated into 3 groups: (1) controls (sham-operated, notreatment), (2) pancreatitis and placebo, and (3) pancreatitis and probiotics. Acute pancreatitis wasinduced by intraductal glycodeoxycholate and intravenous cerulein infusion. Daily probiotics or placebo wasadministered intragastrically from 5 days prior until 7 days after induction of pancreatitis. Tissue andfluid samples were collected for microbiologic and quantitative real-time PCR analysis of bacterialtranslocation.Results. Probiotics reduced duodenal bacterial overgrowth of potential pathogens (Log10 colony-formingunits [CFU]/g 5.0 � 0.7 [placebo] vs 3.5 � 0.3 CFU/g [probiotics], P � .05), resulting in reducedbacterial translocation to extraintestinal sites, including the pancreas (5.38 � 1.0 CFU/g [placebo] vs3.1 � 0.5 CFU/g [probiotics], P � .05). Accordingly, health scores were better and late phasemortality was reduced: 27% (4/15, placebo) versus 0% (0/13, probiotics), respectively, P � .05.Conclusions. This experiment supports the hypothesis that modification of intestinal flora withmultispecies probiotics results in reduced bacterial translocation, morbidity, and mortality in the courseof experimental acute pancreatitis. (Surgery 2007;141:470-80.)

From the Gastrointestinal Research Unit, Departments of Gastroenterology and Surgery,a Department ofMicrobiology,b University Medical Center, Utrecht; Laboratory of Microbiology, Agrotechnology and Food SciencesGroup, Wageningen University, Wageningenc; Department of Pediatric Immunology, University Medical Center,Utrechtd; Department of Medical Microbiology and Immunology, St. Antonius Hospital, Nieuwegeine;

The Netherlands

Severe acute pancreatitis follows a biphasicclinical course. The early proinflammatory phase isassociated with systemic inflammatory response

Part of this work was presented at the 13th United EuropeanGastroenterology Week (Copenhagen, Denmark) and was pub-lished as an abstract (Gut 2005; suppl VII (54):A24).

Supported in part by Astra Zeneca, Research & Development,Mölndal, Sweden (L.P.M.); Winclove Bio Industries B.V., Am-sterdam, The Netherlands (H.M.T.); and Senter, an agency ofthe Dutch Ministry of Economic Affairs (grant number:TSGE3109). Supporting institutions were not involved in de-

sign, performance or publication of this study.

470 SURGERY

syndrome (SIRS), (multiple) organ damage andearly mortality (�1 week). The late phase is char-acterized by infectious complications following bac-

Accepted for publication October 7, 2006.

Reprint requests: Louis M. A. Akkermans, PhD, Departmentof Surgery G.04.228, University Medical Center Utrecht, POBox 85500, 3508 GA Utrecht, The Netherlands. E-mail:L.M.A.Akkermans@umcutrecht.nl

0039-6060/$ - see front matter

© 2007 Mosby, Inc. All rights reserved.

doi:10.1016/j.surg.2006.10.007

Surgery van Minnen et al 471Volume 141, Number 4

terial translocation of intestinal bacteria and latemortality (�3 weeks).1-4 Infectious complicationsare frequently the cause of mortality in patientswith acute pancreatitis.1,5

The use of prophylactic antibiotics to preventinfectious complications remains a topic of debate.A recent metaanalysis of 6 randomized, controlledtrials concluded that prophylactic antibiotics donot prevent infection of pancreatic necrosis ormortality in severe, acute pancreatitis.6 Further-more, increased concerns regarding the wide-spread use of prophylactic antibiotics associatedcomplications (ie, fungal infections or antibioticsresistance) have been reported.7-11 Prophylacticprobiotics have been suggested as an alternative tothe use of prophylactic antibiotics.12,13 Beneficialeffects of prophylactic probiotics for acute pancre-atitis have been reported in animal experimentsand clinical trails.12,14 Most studies on prophylacticprobiotics however, have focused on a single pro-biotic strain for a variety of medical disorders. Arecent report however, has advocated the use ofspecifically selected multiple probiotic strains.15

For optimal results, probiotic strains should be se-lected to target known pathophysiologic aspects ofthe disorder addressed.16

Experimental and clinical studies have greatlyincreased the knowledge of the pathophysiology ofbacterial translocation during acute pancreatitis.Three major steps in the sequence of bacterialtranslocation have been identified: (1) small-bowel bacterial overgrowth, (2) mucosal barrierfailure, and (3) proinflammatory responses.17-23

All these phenomena occur early after the start ofacute pancreatitis and cumulate into bacterialtranslocation and infectious complications. Basedon these considerations, we have designed a mix-ture of 6 probiotic strains, especially selected toreduce small-bowel bacterial overgrowth and re-duce proinflammatory immune responses.24

It is unknown if modification of the intestinalflora with such a multispecies probiotic mixturereduces bacterial translocation and, consequently,alters the course of disease. Therefore, the aim ofthe present study was to assess if modification ofintestinal flora by a specifically designed, multispe-cies probiotic mixture changes disease course usinga well-established rat model of acute pancreatitis.

MATERIAL AND METHODSAnimals. Male, pathogen-free Sprague-Dawley

rats, 250 to 350 grams, (Harlan, Horst, The Neth-erlands) were kept under constant housing condi-tions (temperature, 22°C; relative humidity, 60%; and

a 12-hour light/dark cycle) and had free access to

water and food (RMH 1110, Hope Farms, Woerden,The Netherlands) throughout the experiment. Ratswere allowed to adjust to these conditions for 1week prior to surgery. The experimental designshown in Fig 1 was approved by the institutionalanimal care committee of the University MedicalCenter, Utrecht, The Netherlands (reference no.0303027). Rats were randomized (1:2:2 relativegroup size computer-generated randomization forthe respective groups) among 3 experimentalgroups: (1) control animals (gastric cannula, shampancreatitis, no treatment), (2) acute pancreatitisand placebo, and (3) acute pancreatitis and probi-otics. In total, 10 rats were included in the controlgroup, 21 rats in the placebo group, and 17 rats inthe group given daily probiotics.

Probiotics and placebo. The study product (Eco-logic® 641, Winclove Bio Industries, Amsterdam,The Netherlands) consisted of viable and freeze-dried probiotic strains, including 4 lactobacilli(Lactobacillus acidophilus [W70], Lactobacillus casei[W56], Lactobacillus salivarius [W24], and Lactococ-cus lactis [W58], and 2 bifidobacteria (Bifidobacte-rium bifidum [W23] and Bifidobacterium infantis[W52]). The placebo product consisted of carriersubstance only (cornstarch). Probiotics and pla-cebo were packed in identical sachets and coded bythe producer to guarantee blinding during theexperiment. Directly before administration ofthe doses, the products were reconstituted in sterilewater, for 15 min at 37°C. Single probiotics dosevolume of 1.0 ml contained 5 � 109 colony-formingunits (CFU) bacteria. Probiotics or placebo wereadministered intragastrically once daily through apermanent gastric cannula beginning 5 days priorto induction of acute pancreatitis, and twice dailyfor 6 days after induction of acute pancreatitis.

Surgical procedures. All surgical procedures

Fig 1. Experimental design. Eight days prior to induc-tion of acute pancreatitis, a permanent gastric cannulawas fitted. Probiotics or placebo were administered intra-gastrically through a permanent gastric cannula oncedaily, starting 5 days prior to induction of acute pancre-atitis, and twice daily from days 1 to 7 after inductionof acute pancreatitis. Seven days after induction of pan-creatitis, surviving rats were anesthetized to allow sterileremoval of organ and blood samples. Control animalsdid not receive administrations through the gastric can-nula and underwent a sham pancreatitis procedure only.

were performed on a heated operating table under

472 van Minnen et al SurgeryApril 2007

general anesthesia using a combination of 2%isoflurane gas (flow: 0.5 l/min O2, 1.5 l/m air)through a snout-mask and intramuscular 0.3 ml10% buprenonorphine (Temgesic, Reckitt Beck-iser Healthcare Ltd, Hull, UK). All surgical proce-dures were performed with sterile instrumentsunder strict aseptic conditions. Throughout theexperiment, random control swabs of the abdomenand used surgical material remained negative onbacterial cultures, ensuring external contamina-tion did not occur.

Gastric cannulation. At the start of the experi-ment, a permanent gastric cannula was fitted in allrats. Under general anesthesia, a 20-cm siliconecannula (outer diameter, 1.65 mm; inner diameter,0.76 mm; Rubber, Amsterdam, The Netherlands)was tunneled subcutaneously from the abdominalwall to the back, penetrating the skin between thescapulae. A 1.5-cm midline laparotomy was made toinsert the gastric end of the cannula into the stom-ach through a puncture within a purse-string su-ture on the greater curvature. The cannula wassecurely fixed and the abdomen was closed in 2layers. The dorsal end of the cannula was kept inplace between the scapulae with a rodent infusionjacket (Uno Zevenaar BV, Zevenaar, The Nether-lands). Animals in the probiotics and placebogroups were allowed to recover for 3 days priorto the start of daily probiotics or placebo ad-ministrations.

Induction of acute pancreatitis. Five days afterstarting daily administration of placebo or probiot-ics, acute pancreatitis was induced using the inter-nationally accepted model described by Schmidtet al.25 Briefly, during midline relaparotomy, thepapilla of Vater was cannulated transduodenallyusing a 24G Abbocath®-T intravenous (i.v.) infu-sion cannula (Abbott, Sligo, Republic of Ireland).Before pressure-monitored infusion (MMS, En-schede, The Netherlands) of 0.5 ml sterilizedglycodeoxycholic acid in 10 mmol/l glycylglycine-NaOH-buffered solution (Sigma-Aldrich ChemieBV, Zwijndrecht, The Netherlands), pH 8.0, 37°C,the common bile duct was clamped, and bile andpancreatic fluid were allowed to drain through thecannula. No animals needed to be excluded forinfusion pressures exceeding 35 mm Hg. Directlyafter infusion, hepatoduodenal bile flow was re-stored by removal of the clamp. The puncture holein the duodenum was carefully closed using an 8.0polyprolene serosal suture. After closure of theabdomen in 2 layers, the right jugular vein wascannulated for continuous postoperative intrave-nous infusion of cerulein (5 �g/kg/h for 6 hours).

The jugular vein cannula was fixed to the rodent

infusion jacket and attached to a swivel system toprovide unrestricted mobility of the rat during in-fusion. During the sham procedure in the controlrats, the papilla of Vater was cannulated, the com-mon bile duct was clamped, but no glycodeoxy-cholic acid was infused. A jugular vein cannula wasfitted for 6 hours of intravenous saline infusion.After acute pancreatitis induction or sham proce-dure, pain relief was provided for 48 hours bysubcutaneous injections of 0.3 ml 10% buprenon-orphine twice daily.

The clinical response of the rats after induction ofacute pancreatitis was assessed using a 0-to-6 pointscoring system: Grooming: normal � 2 points; de-creased � 1 point; none � 0 points; Mobility: nor-mal � 2 points; decreased � 1 point; immobile � 0points; Pain posture: none � 2 points; arching (con-vex back and retraction of abdomen from floor) �1 point; stretching (whole body is stretched out onfloor, spine is straight and horizontal) � 0 points.Aspects of this scoring system are well-recognizedbehavioral parameters expressing health or mor-bidity (including abdominal/visceral pain).26,27 Ac-cording to Dutch animal welfare laws and localprotocols of the animal ethics committee, dailyassessments of these aspects are mandatory to mon-itor animal welfare throughout the experimentalprotocol. Indeed, 2 rats in the placebo group dem-onstrating signs of severe suffering and poor clini-cal prognosis (low health scores) were killed on day6 and added to the Kaplan-Meier statistic the sameday.

Collection of tissue and fluid samples. On day 7,surviving rats were anesthetized to allow sterile re-moval of organ and fluid samples. To avoid cross-contamination, samples were taken under strictaseptic conditions in the following order: perito-neal fluid, blood (inferior vena cava), mesentericlymph nodes (MLN), liver, spleen, pancreas, andduodenum. After sample collection, rats werekilled by blood loss. Samples were collected formicrobiologic analysis, and a portion of each sam-ple was snap-frozen in liquid nitrogen and stored at�80°C for future analysis. Another portion of pan-creatic samples was analyzed histopathologically,using standard hematoxylin and eosin (H&E) stain-ing. Histopathologic severity of acute pancreatitiswas assessed based on the acute pancreatitis scoringsystem, as previously described.28

Culture-based microbiologic analysis for bacte-rial identification and quantification. All organsamples were weighed and processed immediatelyfor quantitative and qualitative cultures of aerobicand anaerobic organisms. All organs were homog-

enized in cysteine broth with a sterile blender and

Surgery van Minnen et al 473Volume 141, Number 4

cultured in 10-fold dilution series. The sampleswere cultured on blood agar (Becton Dickinson,Leiden, The Netherlands); MacConkey agar (Ox-oid, Haarlem, The Netherlands) for gram-negativestrains; Columbia colistin nalidixic acid agar (CNA;Becton Dickinson) for staphylococci and strepto-cocci; de Man, Rogosa and Sharpe agar (Oxoid)for lactobacilli; and Brucella Blood agar (BectonDickinson) for anaerobic bacteria. The microor-ganisms were identified using standard microbio-logic techniques. For analysis of organ samples,cultured bacteria were subdivided into 3 groups:gram-positive cocci (GPC), gram-positive rods(GPR) and gram-negative rods/anaerobes (GNR�anear). Also, beta-hemolytic streptococcus groupB, Enterococcus spp., Staphylococcus aureus, and Enter-obacteriaceae, such as Escherichia coli (E. coli),Proteus mirabilis, and Morganella morganii, werecategorized as potential pathogens. Bacterialcounts are expressed as Log10 CFU per gram tissue(CFU/g) � standard error of the mean. Thresholddetection level of bacterial growth was �102

CFU/g.DNA isolation and real-time polymerase chain

reaction assay for total bacterial quantification.DNA was isolated from mesenteric lymph nodes andpancreas homogenates using Fast DNA Spin Kit(Qbiogene, Inc, Carlsbad, Calif) as previously de-scribed.29 Subsequently, total bacterial quantificationwas performed employing 16S rRNA gene-targetedprimers, 968F (5=- AAC GCG AAG AAC CTT AC -3=)and R1401 (5=-CGG TGT GTA CAA GAC CC-3=).Real-time polymerase chain reaction (PCR) wasdone on an iCycler IQ real-time detection systemcoupled to the iCycler optical system interface soft-ware ver 2.3 (Bio-Rad, Veenendaal, The Nether-lands). The reaction mixture (25 �l) consisted of12.5 �l of IQ SYBR Green Supermix (Bio-Rad), 0.2�M of each primer set, and 5 �l of the templateDNA. The PCR conditions for total bacterial quan-tification were 94°C for 5 min, and 35 cycles of94°C for 30 s, 56°C for 20 s, 68°C for 40 s.30 Seriallydiluted genomic DNA of selected bacterial isolateswas used as real-time PCR control for total bacteriaquantification. PCR bacterial counts are expressedas Log10 cells per gram tissue (Cells/g) � standarderror of the mean.

Statistical analysis. Survival rates were analyzedwith the Kaplan-Meier method. Health scores andincidence of positive bacterial cultures were com-pared between groups using the nonparametricMann-Whitney U test. Bacterial counts (cultures)and cell counts (PCR) were analyzed using t testsfor relevant subgroups (SPSS 12.0 statistical soft-

ware; SPSS Benelux, Gorinchem, The Nether-

lands). Spearman rank correlation coefficientswere computed for linear correlation analyses. Re-sults are presented as mean � standard error of themean. Culture results are presented as mean Log10

CFU/gram tissue and quantitative real-time PCRresults as Log10 cells/gram tissue. Statistical signif-icance was accepted when 2-tailed P values wereless than .05.

RESULTSMorbidity and mortality. After the start of daily

placebo or probiotic administrations, physical be-havior of all rats remained normal, resulting inmaximal health scores from day -5 until day 0. Theclinical response of the rats after induction of ex-perimental pancreatitis followed a biphasic course.During the first 72 hours, the animals exhibiteddecreased grooming or motility and, to some ex-tent, behavior associated with pain, despite analge-sic administration during the first 48 hours. Fromdays 3 to 5, surviving animals apparently recovered,evidenced by near-to-normal physical behavior. Af-ter day 5, the rats’ status deteriorated, resulting ina second decrease of health scores. Throughoutdays 1 to 7, median health scores of surviving ratswere higher for rats in the probiotics group com-pared with those in the placebo group, with signif-icant contrasts on days 1, 2, and 3: median � 5(range, 3 to 6) vs median � 4 (range, 1 to 6), P �.020; median � 5 (range, 4 to 6) vs median � 4.5(range, 2 to 6), P � .034; and median � 6 (range,3 to 6) vs median � 3 (range, 1 to 5), P � .001,respectively. An interpretation of the biphasiccourse of acute pancreatitis is illustrated in Fig 2 bythe curves superimposed on the median healthscores.

In the pancreatitis groups, histologic examina-tion of the pancreatic samples revealed late se-quelae of severe necrotizing acute pancreatitis(Fig 3). The extent of necrosis, hemorrhage, in-flammatory infiltrate, or fibrosis was comparablefor the probiotics and placebo groups, suggestingrats of both pancreatitis groups were subject toacute pancreatitis of equal severity.

Overall mortality due to acute pancreatitis was37% (14/38). Mortality in the probiotics group was24% (4/17) and 48% (10/21) in the placebogroup (P � .16). Mortality within the first 24 and48 h was comparable between both groups (�24 h:12% (2/17) vs 14% (3/21); �48 h: 24% (4/17) vs29% (6/21) for probiotics and placebo groups,respectively). However, late mortality (�48 h) didnot occur in the probiotics group, resulting in asignificant reduction of mortality compared to the

placebo group (�48 h: 0% (0/13) vs 27% (4/15),

474 van Minnen et al SurgeryApril 2007

respectively, P � .049). The Kaplan-Meier survivalcurve for both pancreatitis groups is shown in Fig 4.Rats that died before the scheduled 7 days afterinduction of acute pancreatitis were not analyzedfurther, leaving 34 rats for bacteriologic analysis (con-trols: n � 10, placebo: n � 11, probiotics: n � 13).

Duodenal bacterial overgrowth and bacterialtranslocation. Total bacterial counts and numbersof lactobacilli in the duodenum were not signifi-cantly affected by induction of acute pancreatitis oradministration of probiotics (Fig 5). In contrast,acute pancreatitis caused a significant increase in to-tal counts of potential pathogens in the duodenum(beta-hemolytic streptococcus group B, Enterococcusspp., Staphylococcus aureus, and enterobacteriaceaesuch as E. coli, Proteus mirabilis, and Morganella morga-nii), compared with sham-operated controls (5.0 �0.7 CFU/g vs 2.9 � 0.2 CFU/g, P � .010). Mostinterestingly, probiotics prevented this duodenalbacterial overgrowth of potential pathogens (3.5 �0.3 CFU/g vs 5.0 � 0.7 CFU/g, for probiotics vsplacebo, respectively, P � .048). The most predom-inant effects of probiotics on duodenal flora werethe reduction of enterococci and E. coli (1.9 � 0.2CFU/g vs 2.8 � 0.6 CFU/g, P � .159 and 2.0 �0.2 CFU/g vs 3.8 � 0.8 CFU/g, P � .032, respec-tively; Fig 5).

Fig 2. Median health scores after induction of acutepancreatitis, of placebo rats (□) and rats of the probiot-ics group (�). Bars represent 25% to 75% interquartilerange. Median health scores were improved by probioticsthroughout days 1 to 7, with significant differences ondays 1, 2 and 3 (*P � .05). Health scores of all rats wereinvariably 6 (maximum score) from day -5 until induc-tion of acute pancreatitis (data not shown). The placebo(dashed line) and the probiotics (solid line) and are fittedto demonstrate an interpretation of the biphasic courseof acute pancreatitis.

In rats of the placebo group, duodenal bacterial

overgrowth correlated positively and significantlywith bacterial translocation to the pancreas (r �0.89, P � .001, Fig 6, A). Strikingly, no correlationbetween duodenal bacterial overgrowth and infec-tion of pancreatic necrosis was present in the pro-biotics group (r � 0.13, P � .829, Fig 6, B).

Multispecies probiotics reduce bacterialtranslocation. The occurrence of bacterial infec-tion of gram-positive cocci, gram-positive rods, andgram-negative rods/anaerobes in peritoneal fluidand blood are shown in Table I. Gram-negativerods/anaerobes were cultured less frequently inblood of rats of the probiotics group (0/13 vs 4/11,P � .05). Induction of acute pancreatitis resulted insignificant bacterial translocation to the mesentericlymph nodes, spleen, liver, and pancreas. (Figs 7,A-E). Rats in the probiotics group demonstratedsignificantly lower counts of microorganisms cul-tured in spleen, liver and pancreas compared withrats of the placebo group (2.4 � 0.3 CFU/g vs3.7 � 0.5 CFU/g, P � .039; 2.8 � 0.8 CFU/gvs 4.6 � 0.8 CFU/g, P � .049; 3.1 � 0.5 CFU/g vs5.4 � 1.0 CFU/g, P � .042, respectively; Fig 7, A).Bacterial counts of gram-positive cocci, gram-posi-tive rods, and gram-negative rods/anaerobes in thevarious abdominal organs are shown in Figs 7, B-E.Absolute bacterial counts were lower in all tissuesof the probiotics group compared to placebo,and reached significance for gram-positive cocciin the spleen and for gram-negative rods/anaer-obes in the spleen and liver (2.3 � 0.3 CFU/g vs3.5 � 0.5 CFU/g, P � .026; 2.0 � 0.1 CFU/gvs 3.0 � 0.5 CFU/g, P � .032; 2.0 � 0.1 CFU/gvs 3.1 � 0.6 CFU/g, P � .022, respectively;Figs 7, C-D).

E. coli and Enterococcus spp. were the most pre-dominant bacteria found in tissues of rats withacute pancreatitis. Administration of probiotics re-sulted in significantly reduced bacterial growth ofboth E. coli and enterococci in the mesentericlymph nodes (1.9 � 0.1 CFU/g vs 2.6 � 0.4 CFU/g,P � .045; 1.8 � 0.05 CFU/g vs 2.5 � 0.3 CFU/g,P � .015, respectively). Pancreatic counts of E. coliand enterococci were numerically reduced by pro-biotics but failed to reach significant differences(1.7 � 0.01 CFU/g vs 3.0 � 0.7 CFU/g, P � .067;1.7 � 0.01 CFU/g vs 3.0 � 0.7 CFU/g, P � .060,respectively).

In accordance with the data obtained by bacte-rial culture, results of quantitative real-time PCRdemonstrated lower bacterial translocation in themesenteric lymph nodes and pancreas of rats in theprobiotics group (5.9 � 0.4 Cells/g vs7.0 � 0.1Cells/g, P � .043, and 6.7 � 0.3 Cells/g vs 7.7 � 0.3

Cells/g, P � .013, respectively; Fig 8).

inal m

Surgery van Minnen et al 475Volume 141, Number 4

DISCUSSIONThis is the first study to assess the potential of a

specifically designed multispecies probiotic mix-ture to reduce bacterial translocation during acutepancreatitis. In this paper, we demonstrate thatmodification of intestinal flora with this probioticmixture alters the course of experimental acutepancreatitis. Administration of the selected probi-otic mixture resulted in the following: (1) reducedduodenal overgrowth of pathogens, such as E. coli;(2) reduced bacterial translocation to distant or-gans, including the pancreas, (3) improved clinicalcourse, and (4) reduced late mortality.

These results confirm our earlier reports thatsmall-bowel bacterial overgrowth during acute pan-creatitis correlates with infection of pancreatic ne-crosis.17 Daily administration of probiotics did notincrease the duodenal total count of lactobacilli.

Fig 3. A, Normal pancreatic histology of controlincluding destruction of acinar structures, fibrosinduction of acute pancreatitis. H&E staining; orig

Fig 4. Kaplan-Meier survival plot of probiotics groups(solid line) and placebo groups (dashed line). Overallmortality placebo vs probiotics: P � .16. Late mortality(�2 days): P � .05.

Functionality of the administered probiotics is

demonstrated by the significantly reduced num-bers of potential pathogens in the duodenum, par-ticularly E. coli. Reduction of acute pancreatitisinduced pathogen overgrowth in the small bowelresulted in reduced bacterial infection of pancre-atic necrosis in rats of the probiotics group. Re-duced presence of luminal pathogens may havehad favorable effects on mucosal barrier functionof the proximal small bowel, reducing bacterialtranslocation. However, effects of the selectedprobiotic mixture on mucosal barrier functionneeds to be investigated in additional experimentalstudies.

Microbiologic cultures demonstrated a signifi-cant decrease of bacterial growth in the spleen,liver, and pancreas in rats of the probiotics group.Confirming these results, quantitative real-time

, Histopathologic sequelae of acute pancreatitisd a massive inflammatory infiltrate 7 days afteragnification �100.

Fig 5. Counts of total duodenal bacterial; lactobacilli;1potential pathogens (beta-hemolytic Streptococcus groupB, Enterococcus spp., Staphylococcus aureus, and enterobac-teriaceae, such as Escherichia coli [E. coli], Proteus mirabilis,and Morganella morganii, and enterococci); E. coli; andenterococci. Acute pancreatitis resulted in bacterial over-growth of potential pathogens. †Controls vs placebo; P �.05. Probiotics (black bars) reduced bacterial counts ofpotential pathogens, mainly E. coli, compared with pla-cebo (gray bars). *Placebo vs probiotics; P � .05.

rats. Bis, an

PCR detection of bacterial DNA revealed a signifi-

476 van Minnen et al SurgeryApril 2007

cant decrease of bacterial translocation to mesen-teric lymph nodes and pancreas. Bacterial DNA isstable, and PCR-based methods are highly sensitiveand specific to detect minimal amounts of bacterialDNA in serum of patients with acute pancreatitis.31

This method detects viable and nonviable translo-cated bacteria, probably killed by the host immunesystem. Therefore, total bacterial load estimated byreal-time PCR is higher than counts of viable mi-croorganisms only. Moreover, average reduction ofpancreatic bacterial load by probiotic prophylaxiswas greater when analyzed by culture (�2 Log)than by PCR (�1 Log). Thus, reduction of viablebacteria cultured from pancreatic necrosis cannotbe completely explained by an absolute reductionof bacterial translocation, as indicated by quantita-tive real-time PCR. It could be suggested that pro-biotic prophylaxis renders the immune system

Fig 6. Correlation between duodenal bacterial countsand bacterial counts in the pancreas for rats in theplacebo (A) and probiotics (B) group.

more capable of killing translocated bacteria in

distant organs. In follow-up studies, we are cur-rently addressing effects of enteral probiotics on awide panel of plasma cytokines to assess immunemodulatory potential in the early and late phases ofexperimental acute pancreatitis.

Rats in the probiotics group showed less stress- orpain-associated behavior, demonstrated by objec-tive improvement in the clinical course of experi-mental pancreatitis. Albeit a biphasic course inclinical presentation could still be identified in ratsof the probiotics group, health scores were clearlyimproved compared with placebo rats. Moreover,probiotic prophylaxis numerically reduced overallmortality of acute pancreatitis, and a significantreduction was observed in late-phase mortality. Inhumans, infectious complications are thought tobe responsible for late-phase mortality.2,5 In accor-dance with these reports, reduced infectious com-plications in probiotic-treated rats were associatedwith reduced late-phase mortality in the presentexperiment. Probiotics did not affect histologic se-verity, which was assessed 7 days after induction ofacute pancreatitis. Early-phase histologic changeswere not assessed.

The experimental design of the present studyaimed to assess the effect of gut flora modulationby probiotics on the course of experimental acutepancreatitis, using bacterial translocation as a ma-jor outcome parameter. Therefore, rats were pre-treated with the selected probiotics or placebo. Inexperimental acute pancreatitis, timing of the startof treatment remains a challenging issue. Thecourse of acute pancreatitis in rats is approximately3 to 6 times faster than in humans.32-35 This leavesonly a small treatment window between onset ofdisease and occurrence of complications, poten-tially leading to false-negative results if treatment isstarted after induction of pancreatitis. For this rea-son, treatment was started before induction ofacute pancreatitis in many other studies.34,36,37

Also, for probiotics in particular, assessment oftheir efficacy by pretreatment is an accepted exper-imental method to provide proof of principle.38 Weemphasize that results of the present experimentdo not necessarily reflect potential results if treat-ment is started after the onset of acute pancreatitisin general, and potential clinical success or validityin particular.

For a prophylactic strategy to be effective, itshould intervene with the pathophysiology of bac-terial translocation during acute pancreatitis asearly as possible. The exact pathophysiology ofbacterial translocation, infection of pancreatic ne-crosis, and the ensuing systemic effects is still not

fully understood. Yet, the sequence of some major

Surgery van Minnen et al 477Volume 141, Number 4

pathophysiologic aspects has been clarified. Earlyafter the onset of acute pancreatitis, neurohor-monal effects result in reduced small-bowel mo-tility,17 which causes stasis of luminal contents andsmall-bowel bacterial overgrowth with potentialpathogens, including E. coli and Enterococcus spp.

The abundant presence of luminal pathogens

Table. Incidence of bacterial infection of peritonewith pancreatitis treated with placebo or probiotic

Bacterial group

Peritoneal fluid % (n)

Controls(n � 10)

Placebo(n � 11)

GPC 10.0 (1) 36.4 (4)GPR 0 (0) 0 (0)GNR�anaerobes 0 (0) 36.4 (4)

GPC, Gram-positive cocci; GPR, gram-positive rods; GNR�anaerobes, gramNumbers represent percentage and absolute numbers (n) of positive cu*Probiotics vs placebo: P � .05 (Mann-Whitney U test).

Fig 7. A, Total bacterial counts in mesenteric lymrats (white bars), placebo rats (gray bars), and rats oof gram-positive cocci (GPC), gram-positive rods (Gin mesenteric lymph nodes (MLN; B), spleen (Cprobiotics; †P � .05 controls vs placebo.

forms a challenge for the mucosal barrier. Further-

more, pancreatitis-associated reduced intestinalblood flow results in mucosal ischemia and reper-fusion damage.39-41 Luminal bacteria, normallyheld at bay by the mucosal barrier, now have op-portunity to penetrate into the intestinal epithe-lium. Local intestinal inflammation follows, furthercompromising mucosal barrier function. Pancreati-

id and blood in sham-operated controls and rats

Blood % (n)

biotics13)

Controls(n � 10)

Placebo(n � 11)

Probiotics(n � 13)

1 (3) 0 (0) 18.2 (2) 7.7 (1)0 (0) 0 (0) 0 (0) 0 (0)7 (1) 0 (0) 36.4 (4) 0 (0*)

ve rods � anaerobes.mples.

des (MLN), spleen, liver, and pancreas in controlrobiotics group (black bars). B-E, Bacterial counts

and gram-negative rods/anaerobes (GNR�anaer)er (D), and pancreas (E). *P � .05: placebo vs

al flus

Pro(n �

23.

7.

-negatilture sa

ph nof the pPR),), liv

tis and ensuing intestinal inflammation both contrib-

478 van Minnen et al SurgeryApril 2007

ute to a systemic proinflammatory response (systemicinflammatory response syndrome, [SIRS]), with dam-aging effects on distant organs.42,43 If the systemicresponse is severe, multiple organ dysfunction syn-drome (MODS) might follow.44,45 If the patientsurvives the early phase, counter-regulatory immuno-logic pathways releasing antiinflammatory cytokinesresult in a refractory state characterized by immuno-suppression.46,47 Persistent immunosuppression willrender the patient liable for infection of pancreaticnecrosis. MODS caused by infectious complicationsis considered accountable for so-called late mortal-ity or “late septic death.”46,48

With this pathophysiology of local and systemicevents during severe acute pancreatitis in mind,6 probiotic strains were selected for this study.Selection of strains was based on their in vitroantibacterial and immunomodulatory properties.24

Lactobacillus acidophilus and Lactobacillus salivariuswere selected for their ability to suppress growth of

Fig 8. Quantification of total bacterial load by real-timePCR in placebo rats (gray bars) and rats of the probioticsgroup (black bars) in mesenteric lymph nodes (A) andpancreas (B). Data of microbiologic cultures is shown forcomparison and are also presented in Fig 7. *P � .05

E. coli and enterococci. Bifidobacterium infantis also

demonstrated antimicrobial effects. Lactococcus lac-tis and Bifidobacterium bifidum demonstrated im-mune-modulating properties, including decreasingproinflammatory and increasing antiinflammatoryimmune responses. Finally, Lactobacillus casei dem-onstrated both antimicrobial and immune-modu-lating properties.

A thorough review on the use of animal modelsof acute pancreatitis demonstrated that the modelused in this study was preferred to examine patho-physiology of bacterial translocation and for testingtreatment strategies.32,33 The major advantages ofthis model include resemblance to human acutepancreatitis with regard to bacteriologic results,reaction to treatment, and disease course.25,32 Thetransduodenal approach to the biliopancreaticduct for bile salt infusion is often thought to be amajor drawback of the model because of its poten-tial to introduce bacteria in pancreatic tissue How-ever, results of the control group in the presentstudy once more confirm that transduodenal can-nulation of the biliopancreatic duct does not resultin bacterial contamination of any concern to studyoutcome. Because of its demonstrated value, thismodel has been applied in many experiments test-ing the value of antibiotics during acute pancreati-tis.34,35,49,50

Experimentally, prophylactic antibiotics re-duced overgrowth of E. coli and enterococci in thesmall bowel, resulted in significantly reduced bac-terial translocation to distant organs, including thepancreas, and reduced mortality.4,34 Unfortu-nately, clinical results of prophylactic antibioticswere not as successful. In a recent placebo-con-trolled, double-blinded clinical trial, Isenmannet al51 demonstrated that prophylactic antibiotics(ciprofloxacin/metronidazole) showed no effecton bacterial infection of pancreatic necrosis or clin-ical outcome. A recent metaanalysis confirmedthese findings.6 Furthermore, various concernsabout the use of prophylactic use of broad spec-trum antibiotics have been expressed, includingthe possibility that they could increase the inci-dence of nosocomial infections with resistant bac-teria or fungi.7-9,52-57 The specifically selectedmultispecies probiotics presented in this study maybe a novel and potentially effective alternative.However, as was demonstrated by the contrast be-tween experimental and clinical results of antibiot-ics in acute pancreatitis, the clinical value ofspecifically selected multispecies probiotics re-mains to be proven. For this reason, the DutchAcute Pancreatitis Study Group embarked on arandomized, double-blinded, placebo-controlled,

multicenter trial on prophylactic multispecies pro-

Surgery van Minnen et al 479Volume 141, Number 4

biotics in patients with predicted severe acute pan-creatitis.58

In summary; modification of intestinal flora withmultispecies probiotics that were especially de-signed to address pathophysiology of bacterialtranslocation resulted in reduced small-bowel bac-terial overgrowth, bacterial translocation to distantorgans, and associated morbidity and late mortalityin experimental acute pancreatitis.

The authors are grateful to Prof. dr. J.E. van Dijk of theDepartment of Pathobiology, Division of Pathology, Fac-ulty of Veterinary Science, Utrecht University, The Neth-erlands, for processing pancreatic samples for histologicreviewing.

REFERENCES1. Buchler MW, Gloor B, Muller CA, Friess H, Seiler CA, Uhl

W. Acute necrotizing pancreatitis: treatment strategy ac-cording to the status of infection. Ann Surg 2000;232:619-26.

2. Widdison AL, Karanjia ND. Pancreatic infection complicat-ing acute pancreatitis. Br J Surg 1993;80:148-54.

3. Isenmann R, Rau B, Zoellner U, Beger HG. Management ofpatients with extended pancreatic necrosis. Pancreatology2001;1:63-8.

4. Beger HG, Rau B, Isenmann R, Schwarz M, Gansauge F,Poch B. Antibiotic prophylaxis in severe acute pancreatitis.Pancreatology 2005;5:10-9.

5. Beger HG, Rau B, Mayer J, Pralle U. Natural course of acutepancreatitis. World J Surg 1997;21:130-5.

6. Mazaki T, Ishii Y, Takayama T. Meta-analysis of prophylacticantibiotic use in acute necrotizing pancreatitis. Br J Surg2006;93:674-84.

7. Hoerauf A, Hammer S, Muller-Myhsok B, Rupprecht H.Intra-abdominal Candida infection during acute necrotiz-ing pancreatitis has a high prevalence and is associated withincreased mortality. Crit Care Med 1998;26:2010-5.

8. Grewe M, Tsiotos GG, Luque de-Leon E, Sarr MG. Fungalinfection in acute necrotizing pancreatitis. J Am Coll Surg1999;188:408-14.

9. Isenmann R, Schwarz M, Rau B, Trautmann M, Schober W,Beger HG. Characteristics of infection with Candida speciesin patients with necrotizing pancreatitis. World J Surg2002;26:372-6.

10. De Waele JJ, Vogelaers D, Blot S, Colardyn F. Fungal infec-tions in patients with severe acute pancreatitis and the use ofprophylactic therapy. Clin Infect Dis 2003;37:208-13.

11. De Waele JJ, Vogelaers D, Hoste E, Blot S, Colardyn F.Emergence of antibiotic resistance in infected pancreaticnecrosis. Arch Surg 2004;139:1371-5.

12. Olah A, Belagyi T, Issekutz A, Gamal ME, Bengmark S.Randomized clinical trial of specific lactobacillus and fibresupplement to early enteral nutrition in patients with acutepancreatitis. Br J Surg 2002;89:1103-7.

13. Besselink MG, Timmerman HM, van Minnen LP, Akker-mans LM, Gooszen HG. Prevention of infectious complica-tions in surgical patients: potential role of probiotics. DigSurg 2005;22:234-44.

14. Mangiante G, Colucci G, Canepari P, Bassi C, Nicoli N,Casaril A, et al. Lactobacillus plantarum reduces infection ofpancreatic necrosis in experimental acute pancreatitis. Dig

Surg 2001;18:47-50.

15. Timmerman HM, Koning CJ, Mulder L, Rombouts FM,Beynen AC. Monostrain, multistrain and multispecies pro-biotics—a comparison of functionality and efficacy. Int JFood Microbiol 2004;96:219-33.

16. Isolauri E, Rautava S, Kalliomaki M, Kirjavainen P, SalminenS. Role of probiotics in food hypersensitivity. Curr OpinAllergy Clin Immunol 2002;2:263-71.

17. Van Felius ID, Akkermans LM, Bosscha K, Verheem A,Harmsen W, Visser MR, et al. Interdigestive small bowelmotility and duodenal bacterial overgrowth in experi-mental acute pancreatitis. Neurogastroenterol Motil2003;15:267-76.

18. Gautreaux MD, Deitch EA, Berg RD. T lymphocytes in hostdefense against bacterial translocation from the gastrointes-tinal tract. Infect Immun 1994;62:2874-84.

19. Berg RD. Bacterial translocation from the gastrointestinaltract. Trends Microbiol 1995;3:149-54.

20. Berg RD. Bacterial translocation from the gastrointestinaltract. Adv Exp Med Biol 1999;473:11-30.

21. Ammori BJ, Leeder PC, King RF, Barclay GR, Martin IG,Larvin M, et al. Early increase in intestinal permeability inpatients with severe acute pancreatitis: correlation with en-dotoxemia, organ failure, and mortality. J Gastrointest Surg1999;3:252-62.

22. Gloor B, Todd KE, Lane JS, Rigberg DA, Reber HA. Mech-anism of increased lung injury after acute pancreatitis inIL-10 knockout mice. J Surg Res 1998;80:110-4.

23. Van Laethem JL, Eskinazi R, Louis H, Rickaert F, Robbere-cht P, Deviere J. Multisystemic production of interleukin 10limits the severity of acute pancreatitis in mice. Gut1998;43:408-13.

24. Timmerman HM, Niers LEM, Ridwan BU, Koning CJM,Mulder L, Akkermans LMA, et al. Design of a multispeciesprobiotic mixture (Ecologic 641) to prevent infectious com-plications in critically ill patients. Utrecht, Utrecht Univer-sity Thesis 2006;167-88.

25. Schmidt J, Rattner DW, Lewandrowski K, Compton CC,Mandavilli U, Knoefel WT, et al. A better model of acutepancreatitis for evaluating therapy. Ann Surg 1992;215:44-56.

26. Stam R, van Laar TJ, Wiegant VM. Physiological and behav-ioural responses to duodenal pain in freely moving rats.Physiol Behav 2004;81:163-9.

27. Houghton AK, Kadura S, Westlund KN. Dorsal columnlesions reverse the reduction of home cage activity in ratswith pancreatitis. Neuroreport 1997;8:3795-800.

28. van Minnen LP, Venneman NG, van Dijk JE, Verheem A,Gooszen HG, Akkermans LM, van Erpecum KJ. Cholesterolcrystals enhance and phospholipids protect against pancre-atitis induced by hydrophobic bile salts: a rat model study.Pancreas 2006;32:369-75.

29. Konstantinov SR, Awati A, Smidt H, Williams BA, Akker-mans AD, de Vos WM. Specific response of a novel andabundant Lactobacillus amylovorus-like phylotype to dietaryprebiotics in the guts of weaning piglets. Appl EnvironMicrobiol 2004;70:3821-30.

30. Nubel U, Engelen B, Felske A, Snaidr J, Wieshuber A,Amann RI, et al. Sequence heterogeneities of genes encod-ing 16S rRNAs in Paenibacillus polymyxa detected by tem-perature gradient gel electrophoresis. J Bacteriol 1996;178:5636-43.

31. de Madaria E, Martinez J, Lozano B, Sempere L, Benlloch S,Such J, et al. Detection and identification of bacterial DNAin serum from patients with acute pancreatitis. Gut 2005;

54:1293-7.

480 van Minnen et al SurgeryApril 2007

32. Foitzik T, Hotz HG, Eibl G, Buhr HJ. Experimental modelsof acute pancreatitis: are they suitable for evaluating ther-apy? Int J Colorectal Dis 2000;15:127-35.

33. Van Minnen LP, Blom M, Timmerman HM, Visser MR,Gooszen HG, Akkermans LMA. The use of animal models tostudy bacterial translocation during acute pancreatitis. JGastrointest Surg 2007, in press.

34. Mithofer K, Fernandez-Del Castillo C, Ferraro MJ, Lewand-rowski K, Rattner DW, Warshaw AL. Antibiotic treatmentimproves survival in experimental acute necrotizing pancre-atitis. Gastroenterology 1996;110:232-40.

35. Schwarz M, Thomsen J, Meyer H, Buchler MW, Beger HG.Frequency and time course of pancreatic and extrapancre-atic bacterial infection in experimental acute pancreatitis inrats. Surgery 2000;127:427-32.

36. Lange JF, van Gool J, Tytgat GN. The protective effect of areduction in intestinal flora on mortality of acute haemor-rhagic pancreatitis in the rat. Hepatogastroenterology1987;34:28-30.

37. Park SJ, Seo SW, Choi OS, Park CS. Alpha-lipoic acid pro-tects against cholecystokinin-induced acute pancreatitis inrats. World J Gastroenterol 2005;11:4883-5.

38. Zareie M, Johnson-Henry KC, Jury J, Yang PC, Ngan BY,McKay DM, et al. Probiotics prevent bacterial translocationand improve intestinal barrier function in rats followingchronic psychological stress. Gut 2006; Apr 25; [Epub aheadof print].

39. Inoue K, Hirota M, Kimura Y, Kuwata K, Ohmuraya M,Ogawa M. Further evidence for endothelin as an importantmediator of pancreatic and intestinal ischemia in severeacute pancreatitis. Pancreas 2003;26:218-23.

40. Yasuda T, Takeyama Y, Ueda T, Hori Y, Nishikawa J, KurodaY. Nonocclusive visceral ischemia associated with severeacute pancreatitis. Pancreas 2003;26:95-7.

41. Rahman SH, Ammori BJ, Holmfield J, Larvin M, McMahonMJ. Intestinal hypoperfusion contributes to gut barrier fail-ure in severe acute pancreatitis. J Gastrointest Surg 2003;7:26-35.

42. McKay CJ, Imrie CW. The continuing challenge of earlymortality in acute pancreatitis. Br J Surg 2004;91:1243-4.

43. Tran DD, Cuesta MA, Schneider AJ, Wesdorp RI. Prevalenceand prediction of multiple organ system failure and mortal-ity in acute pancreatitis. J Crit Care 1993;8:145-53.

44. Bhatia M, Wong FL, Cao Y, Lau HY, Huang J, Puneet P,et al. Pathophysiology of acute pancreatitis. Pancreatol-ogy 2005;5:132-44.

45. Bhatia M. Inflammatory response on the pancreatic acinar

cell injury. Scand J Surg 2005;94:97-102.

46. Gloor B, Muller CA, Worni M, Martignoni ME, Uhl W,Buchler MW. Late mortality in patients with severe acutepancreatitis. Br J Surg 2001;88:975-9.

47. Dugernier TL, Laterre PF, Wittebole X, Roeseler J, LatinneD, Reynaert MS, et al. Compartmentalization of the inflam-matory response during acute pancreatitis: correlation withlocal and systemic complications. Am J Respir Crit Care Med2003;168:148-57.

48. Wilson PG, Manji M, Neoptolemos JP. Acute pancreatitis as amodel of sepsis. J Antimicrob Chemother 1998;41(Suppl A):51-63.

49. Foitzik T, Fernandez-Del Castillo C, Ferraro MJ, Mithofer K,Rattner DW, Warshaw AL. Pathogenesis and prevention ofearly pancreatic infection in experimental acute necrotizingpancreatitis. Ann Surg 1995;222:179-85.

50. Gloor B, Worni M, Strobel O, Uhl W, Tcholakov O, MullerCA, et al. Cefepime tissue penetration in experimentalacute pancreatitis. Pancreas 2003;26:117-21.

51. Isenmann R, Runzi M, Kron M, Kahl S, Kraus D, Jung N,et al. Prophylactic antibiotic treatment in patients withpredicted severe acute pancreatitis: a placebo-controlled,double-blind trial. Gastroenterology 2004;126:997-1004.

52. Clark NM, Patterson J, Lynch JP, III.Antimicrobial resis-tance among gram-negative organisms in the intensive careunit. Curr Opin Crit Care 2003;9:413-23.

53. Sorberg M, Farra A, Ransjo U, Gardlund B, Rylander M,Settergren B, et al. Different trends in antibiotic resistancerates at a university teaching hospital. Clin Microbiol Infect2003;9:388-96.

54. Murray BE. Vancomycin-resistant enterococcal infections.N Engl J Med 2000;342:710-21.

55. Leavis HL, Willems RJ, Mascini EM, Vandenbroucke-GraulsCM, Bonten MJ. [Vancomycin resistant enterococci in theNetherlands] Vancomycineresistente enterokokken in Ned-erland. Ned Tijdschr Geneeskd 2004;148:878-82.

56. Albrich WC, Angstwurm M, Bader L, Gartner R. Drug resis-tance in intensive care units. Infection 1999;27(Suppl 2):S19-S23.

57. Berrouane YF, Herwaldt LA, Pfaller MA. Trends in antifun-gal use and epidemiology of nosocomial yeast infections ina university hospital. J Clin Microbiol 1999;37:531-7.

58. Besselink MGH, Timmerman HM, Buskens E, NieuwenhuijsVB, Akkermans LMA, Gooszen HG. Probiotic prophylaxis inpatients with predicted severe acute pancreatitis (PROPATRIA):design and rationale of a blinded, placebo- controlled random-ised controlled trial. Dutch Acute Pancreatitis Study Group.

BMC Surg 2004;29:4:12.