UvA-DARE is a service provided by the library of the University of Amsterdam (http://dare.uva.nl)

UvA-DARE (Digital Academic Repository)

Neuroimmune mechanisms in postoperative ileus

Boeckxstaens, G.E.; de Jonge, W.J.

Published in:Gut

DOI:10.1136/gut.2008.169250

Link to publication

Citation for published version (APA):Boeckxstaens, G. E., & de Jonge, W. J. (2009). Neuroimmune mechanisms in postoperative ileus. Gut, 58(9),1300-1311. https://doi.org/10.1136/gut.2008.169250

General rightsIt is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s),other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons).

Disclaimer/Complaints regulationsIf you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, statingyour reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Askthe Library: https://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam,The Netherlands. You will be contacted as soon as possible.

Download date: 23 Aug 2019

doi: 10.1136/gut.2008.169250 2009 58: 1300-1311Gut

 G E Boeckxstaens and W J de Jonge ileusNeuroimmune mechanisms in postoperative

http://gut.bmj.com/content/58/9/1300.full.htmlUpdated information and services can be found at:

These include:

References

http://gut.bmj.com/content/58/9/1300.full.html#related-urlsArticle cited in:  

http://gut.bmj.com/content/58/9/1300.full.html#ref-list-1This article cites 88 articles, 17 of which can be accessed free at:

serviceEmail alerting

box at the top right corner of the online article.Receive free email alerts when new articles cite this article. Sign up in the

Topic collections

(32 articles)GUT Recent advances in basic science   � Articles on similar topics can be found in the following collections

Notes

http://group.bmj.com/group/rights-licensing/permissionsTo request permissions go to:

http://journals.bmj.com/cgi/reprintformTo order reprints go to:

http://journals.bmj.com/cgi/epTo subscribe to BMJ go to:

group.bmj.com on October 5, 2010 - Published by gut.bmj.comDownloaded from

Neuroimmune mechanisms inpostoperative ileusG E Boeckxstaens,1,2 W J de Jonge2

1 Department ofGastroenterology, UniversityHospitals Leuven, CatholicUniversity of Leuven, Leuven,Belgium; 2 Department ofGastroenterology andHepatology, Academic MedicalCenter, Amsterdam, TheNetherlands

Correspondence to:Professor G E Boeckxstaens,Department of Gastroenterology,University Hospitals Leuven,Catholic University of Leuven,Herestraat 49, 3000 Leuven,Belgium; guy.boeckxstaens@med.kuleuven.be

ABSTRACTPostoperative ileus (POI) is a common clinical conditionarising after almost every abdominal surgical procedure,leading to increased patient morbidity and prolongedhospitalisation. Recent advances in insight into theunderlying pathophysiology have identified intestinalinflammation triggered by handling of the intestine as themain mechanism. Not only does the local inflammatoryprocess compromise the contractile activity of thehandled intestine, but it also activates inhibitory neuralpathways and possibly triggers inflammation at distantuntouched areas, leading to a generalised impairment ofgastrointestinal motility. Macrophages residing in themuscularis externa and mast cells are the key players inthis inflammatory cascade. Pharmacological interventionspreventing the activation of these immune cells reducethe influx of leucocytes into the intestine, an effectassociated with a reduction of the duration of POI. Newpotential therapeutic strategies to shorten POI based onthese new insights will undoubtedly enter the clinicalarena soon.

In the past century, the introduction of antibioticsand sterile surgical conditions has contributed to arevolutionary improvement in postoperative mor-bidity and survival. Yet, each patient undergoingan abdominal surgical procedure, even if minimallyinvasive techniques are applied, will develop atransient episode of impaired gastrointestinalmotility or postoperative ileus (POI). Althoughsome argue that uncomplicated POI should beconsidered as a ‘‘normal’’ or ‘‘physiological’’response of the intestine to a traumatic eventand thus should be disregarded, it clearly has asignificant impact on patient morbidity, withprolonged hospitalisation and thus increased costs.The annual costs related to POI have beenestimated to be as much as US$1.47 billionannually in the USA, illustrating its large socio-economic impact.1 Over the past decade, ourinsight into its pathophysiology has increasedexponentially. The role of inflammation triggeredby handling of the intestine is now generallyaccepted as the key event in POI. In particular,insight into the bidirectional interaction betweenthe immune system (mast cells, macrophages andother leucocytes) and the autonomic nervoussystem (afferents and efferents), also referred toas neuroimmune interaction, has significantlycontributed to a better understanding of thepathophysiology of POI. Moreover, it has becomeclear that inflammatory mediators released byleucocytes within the gut wall also directlyimpair smooth muscle contractility. The present

manuscript will mainly describe these new devel-opments with the emphasis on their potentialclinical relevance.

DEFINITION—CLINICAL ASPECTSAlthough POI may occur following extra-abdom-inal operations, it is most pronounced and inevi-table after every abdominal surgical procedure. Itpresents clinically as the inability to tolerate food,with abdominal distension, absence of bowelsounds and lack of flatus and defecation. Nauseaand vomiting, pain and postoperative fatiguefurther contribute to the morbidity and prolongedhospitalisation of patients. On average, this periodlasts 2–4 days for conventional abdominal proce-dures, but decreases to as little as (2 days in thecase of laparoscopic surgery.2 Some surgeonsconsider the inability to tolerate food and absenceof bowel sounds during the first few postoperativedays as a normal phenomenon, and only consider‘‘prolonged’’ or ‘‘pathological paralytic ileus’’,which lasts .3 days after surgery, as clinicallyrelevant.3 Others propose to prolong this period to.6 days.4 The incorrect usage of prolonged orparalytic ileus to define POI has introduced someconfusion in the literature regarding the exactdefinition of POI. During a recent consensusmeeting, however, POI was defined as the timefrom surgery until passage of flatus or stooltogether with the time to adequate oral intakemaintained during 24 h. Secondary POI wasdefined by the same symptoms but precipitatedby a complication of surgery, such as an anastomicleak, abscess or peritonitis.5 6 Such prolonged orcomplicated ileus develops in ,10% of abdominalsurgeries, but may increase to up to 25% ofpatients undergoing a hemicolectomy1 (table 1).Obviously, the pathophysiological mechanismsinvolved and the treatment of the latter conditionare completely different and will not be discussedhere.

Transient inhibition of gastrointestinal motilityis well documented as the underlying mechanismand involves the entire gastrointestinal tract. Itwas first described at the turn of the century byBayliss and Starling.7 By now, we know that not allsegments are equally affected; small intestinalmotility is on average disturbed for approximately24 h, gastric motility for between 24 and 48 h,whereas colonic motility is impaired for between48 and 72 h (reviewed in Benson and Wingate8). Itshould be emphasised, however, that normalisa-tion of motility, for example the return of themigrating motor complex in the small intestine,

Recent advances in basic science

1300 Gut 2009;58:1300–1311. doi:10.1136/gut.2008.169250

group.bmj.com on October 5, 2010 - Published by gut.bmj.comDownloaded from

does not necessarily imply that normal functionand transit have returned. Nevertheless, thesestudies underscore that colonic motility is themain determinant of clinical recovery. Thisexplains why first defecation and flatus are oftenused as primary outcome parameters in clinicaltrials. These parameters are, however, non-specific.Passage of flatus strongly depends on patientreporting, whereas passage of stool may simplyreflect rectal emptying and not necessarily there-fore inform of recovery of purposeful gastrointest-inal contractile activity. With the possibility thatnew treatments for POI may arise in the near future,there will be a definite need for better outcomevariables or other biological markers in order toevaluate new treatments better and objectively.

PATHOPHYSIOLOGYAlthough it is now clear that opening of theperitoneal cavity and manipulation of the intestineis the main mechanism underlying POI, otherfactors impairing neuromuscular function, such asanaesthetics and postoperative pain medication,certainly contribute to the postponement ofrecovery of normal transit. The effect of anaes-thetics, especially that of volatile drugs, is, how-ever, short lasting and therefore of only marginalimportance. The use of postoperative opioids tocontrol postoperative pain on the other hand has amuch greater impact on postoperative motility. Itshould be acknowledged that the introduction ofpostoperative opioids has significantly improvedthe patient’s comfort in the very early post-operative phase. Nevertheless, these drugs potentlyinhibit gastrointestinal transit, even when patient-controlled epidural morphine is administered.Efforts to reduce the dose of opioids or toantagonise their effects by peripherally actingopioid m-antagonists such as methylnaltrexone1 oralvimopan9 10 are therefore important to minimisethe detrimental effect of opioids on gastrointest-inal motility.

The main cause of POI, however, clearly relatesto the surgical procedure itself. In the last decade,evidence has accumulated that abdominal surgerytriggers two different phases, each with its owndynamics and underlying pathophysiologicalmechanism (fig 1). The first or early phase isneurally mediated and involves neural reflexesactivated during and immediately following sur-gery. In the late 1990s, the concept was introducedthat the manipulation of the intestine triggers theinflux of leucocytes in manipulated intestinalsegments, impairing the contractile properties ofthe inflamed intestine.11–13 This second phase starts3–4 h after surgery and is responsible for the

sustained and thus clinically more relevant inhibi-tion of gastrointestinal motility. From a clinicalperspective, interference with or prevention of thissecond phase is clearly expected to be mostrelevant and most effective in the treatment ofPOI. It should be emphasised, however, that dataobtained in rodents may not necessarily translateto the human situation for obvious reasons of‘‘species’’ differences, for example due to differentexpression of receptors, different mediatorsreleased or absence of vomiting in rodents.Moreover, mainly for simplicity and to increasereproducibility, most animal models only involvemanipulation of the intestine/colon but do notinclude intestinal resection. The latter could notonly be important from a pathophysiological pointof view, but may also be extremely relevant toexclude potential side effects on healing of anasto-moses. Therefore, animal models dealing withthese shortcomings are certainly warranted.

The early neurogenic phase of POIEarly physiological studies revealed that theintensity and the nature of the nociceptive stimuliapplied strongly determine the duration of theperiod of ileus. Incision of the skin or a simplelaparotomy briefly interrupts gastrointestinalmotility. This brief inhibition is neurally mediatedand is adrenergic in nature since depletion ofadrenergic nerves prevents interruption of normalmotility.14–17 Most probably, the activated neuralpathway involves a spinal loop with afferentsplanchnic nerves synapsing in the spinal cordand efferents travelling back to the gut (fig 2A).3 Incontrast, more intense activation by intestinal

Table 1 Differences between postoperative ileus and prolonged/paralytic ileus

Definition Duration Frequency

Postoperative ileus Time until first flatus or stool +adequate oral intake during 24 h

2–4 days Almost every abdominalsurgical procedure

Prolonged orparalytic ileus

Precipitated by complication ofsurgery

.6 days 10–25%

Figure 1 Schematic representation of the two phasesinvolved in postoperative ileus. The first phase startsduring abdominal surgery and ends soon after it. Thesecond inflammatory phase starts approximately 3–4 hafter surgery, lasts much longer and is therefore clinicallymore relevant.

Abdominal surgery triggers two different phases—that is an early neurogenic and a late inflammatoryphase—each with its own dynamics and underlyingpathophysiological mechanism.

Recent advances in basic science

Gut 2009;58:1300–1311. doi:10.1136/gut.2008.169250 1301

group.bmj.com on October 5, 2010 - Published by gut.bmj.comDownloaded from

manipulation results in prolonged inhibition ofmotility which can only be partially blocked byadrenergic antagonists. Acute studies focusing onthe first 30–90 min after surgery identified theinvolvement of high-threshold supraspinal path-ways activating specific hypothalamic and pon-tine–medullary nuclei such as the nucleus tractussolitarii and the paraventricular and supraopticnucleus of the hypothalamus18–20 (fig 2B). Withinthis pathway, corticotrophin-releasing factor (CRF)seems to play a central role. It is hypothesised that

CRF release stimulates neurons in the supraopticnucleus of the hypothalamus, which send projec-tions to the spinal cord, including the intermedio-lateral column of the thoracic cord, wheresympathetic preganglionic neurons are located.18 21

Activation of these nerves subsequently inhibitsmotility of the entire gastrointestinal tract. Inaddition to this adrenergic inhibitory pathway,intense stimulation of splanchnic afferents triggersan inhibitory non-adrenergic, vagally mediatedpathway22(fig 2B).

Although insight into the neural pathways andneurotransmitters involved in this early short-lasting phase is important, it should be emphasisedthat activation of nociceptors and/or mechanor-eceptors by mechanical stimuli per se duringabdominal surgery will cease once the abdomen isclosed. Other factors such as mediators released bytissue damage or subsequent inflammation there-fore must come into play, explaining the moreprolonged nature of POI.

The late inflammatory phase of POIIn 1978, Bueno et al observed two phases ofinhibition after abdominal surgery in dogsimplanted with intestinal electrodes.23 The first or‘‘primary’’ phase consisted of complete inhibitionof electrical spiking activity which transientlyceased at the end of surgery. After 3–4 h, a secondperiod of inhibition was observed, its durationbeing dependent on the nature of surgery. Inparticular, resection of the small bowel resulted ina long-lasting reduction of spiking activity withrecovery of the first myoelectric complex after94 h.23 Although the authors demonstrated thatthe first phase was mediated by an inhibitoryneural pathway, the exact origin of the secondphase remained unclear.

Almost 20 years later, Kalff et al11–13 demon-strated that this second, long-lasting phase of POIwas mainly due to inflammation of the intestinalmuscularis (fig 1). It was hypothessed thatintestinal manipulation activates resident macro-phages present in the intestinal muscularis externa.These normally quiescent macrophages are orga-nised into a layer or ‘‘network’’ at the level of themyenteric plexus and at the serosal side of theintestine.24–26 Activation of these phagocytes sub-sequently resulted in cytokine and chemokinerelease, followed by an influx of leucocytes startingapproximately 3–4 h after surgery. Most interest-ingly, the spontaneous and stimulated contractileactivity of muscle strips obtained from theinflamed intestine was significantly impaired, con-sistent with observations in POI. Moreover, pre-treatment of animals with antibodies or antisenseoligonucleotides against intercellular adhesion mole-cule-1 (ICAM-1) not only prevented the influx ofleucocytes, but also preserved normal neuromuscu-lar function of muscle strips, providing the proof ofconcept that inflammation induced by manipulationindeed largely contributes to POI.13 27 28

Figure 2 Schematic representation of the neural pathways involved in the inhibition ofgastrointestinal motility induced by laparotomy (A) and more intense nociceptivestimulation during intestinal manipulation. After laparotomy, spinal afferents areactivated, synapsing in the spinal cord where they activate an inhibitory pathwayinvolving prevertebral adrenergic neurons, abolishing the motility of the entiregastrointestinal tract (A). During intestinal manipulation, additional pathways areactivated mediated by the brainstem. Afferent signals are transmitted to the brainstemwhere they trigger an increased autonomic output to the neurons of the intermediolateralcolumn of the thoracic cord, where sympathetic preganglionic neurons (releasingnoradrenaline (NA)) are located. In addition, the motor nucleus of the vagus nerve isactivated, synapsing to inhibitory nitrergic (NO) and vipergic (VIP) neurons.CRF, corticotrophin-releasing factor; ggl, ganglion.

Recent advances in basic science

1302 Gut 2009;58:1300–1311. doi:10.1136/gut.2008.169250

group.bmj.com on October 5, 2010 - Published by gut.bmj.comDownloaded from

Mechanisms triggering the local inflammatoryresponseIn general, the innate immune system plays a keyrole in host defence and in initiating an inflamma-tory response. The latter is due to recognition of avariety of macromolecules through pattern recog-nition rather than by reacting to specific antigens.The innate immune system recognises two largeclasses of macromolecules: first, those related topathogens or pathogen-associated molecular pat-terns (PAMPs), and secondly, molecules released inresponse to cell damage or damage-associatedmolecular patterns (DAMPs).29 30 The prototypeof PAMPs is lipopolysaccharide (LPS), a constituentof the Gram-negative bacterial cell wall, but manyother types exist such as peptidoglycan, bacterialflagellin, or lipoteichoic acid from Gram-positivebacteria. Intracellular molecules such as ATP, uricacid, heat-shock proteins or S100 proteins areexamples of DAMPs, also referred to as ‘‘alarmins’’or danger signals. Both DAMPs and PAMPs arerecognised by pattern recognition receptors such asToll-like receptors (TLRs) and RAGE (receptor foradvanced glycation end-products). For example,LPS is recognised by TLR4 and its binding results inactivation of the immune cells expressing the TLR4receptor. DAMPs activate a variety of (intracellu-lar) receptors and lead, for instance, to theactivation of a multiprotein complex, the ‘‘inflam-masome’’, a structure required for the synthesis ofbiologically active interleukin 1 (IL1) and IL18.31

In POI, most evidence so far identifies mast cells,most probably peritoneal mast cells, and residentmacrophages as the main players of the innateimmune system involved in the inflammatoryresponse to intestinal handling. As shown in fig 3,one of the earliest observations in rodent modelsand even in human is the activation of peritonealmast cells and the subsequent release of mediatorssuch as histamine and mMCP-1 (murine monocytechemoattractant protein)in the peritoneal cav-ity.32 33 In patients undergoing surgery, even gentleinspection of the intestine at the very beginning ofthe surgical procedure triggers the release of mastcell mediators. This may be an important step inthe inflammatory cascade as it will lead to a

transient increase in intestinal permeability withtranslocation of intraluminal bacteria and bacterialproducts (see below). Subsequently, residentintestinal macrophages will be activated, followedby phosphorylation of transcription factors, upre-gulation of inflammatory genes and secretion ofcytokines and chemokines. The latter induce theupregulation of endothelial adhesion molecules andthe subsequent influx of leucocytes at a later stage(fig 3). In addition, the release of DAMPs inresponse to tissue damage evoked by handling canactivate the resident macrophages. In the followingparagraphs, this inflammatory cascade will bediscussed in more detail.

Activation of peritoneal mast cellsWithin the serosa and mesentery, mast cells arefound close to blood vessels before entering the gutwall, often in twos or threes, and particularlyclosely associated with afferent nerve fibres(,25 mm).34 Mast cells are not only involved inadaptive immunity, they are also vital for therecruitment of neutrophils and the elimination ofbacteria from the peritoneal cavity. Mast cell-deficient mice indeed show a significantlyincreased mortality and impaired bacterial clear-ance in a model of acute septic peritonitis.35 Mastcells therefore should be considered as sentinels ofthe peritoneal cavity providing protection againstpotential threats.

The importance of mast cells in the inflamma-tory cascade triggered by intestinal manipulationwas demonstrated in experiments using mast cellstabilisers.32 Both ketotifen and doxantrazolereduced the inflammatory response and delayedgastric emptying 24 h after abdominal surgery.Conversely, incubation of intestinal loops insolution containing the mast cell activator 48/80induces an inflammatory response and POI.Finally, W/WV mutant mice that lack mast cellsfail to develop an intestinal infiltrate followingintestinal manipulation. Reconstitution with wild-type mast cells on the other hand restores thecapacity of mutant animals to recruit leucocytes tothe intestine after surgery.

To date, the exact trigger(s) activating theperitoneal mast cells are still unclear, but neuro-peptides such as substance P or calcitonin gene-related peptide (CGRP) released from activatedafferent nerves could be involved.36 Mast cells inthe mesentery are indeed in close proximity toafferent nerves.34 Once activated, vasoactive andproinflammatory substances such as histamine andproteases are released in the peritoneal cavity. Bothin rodents and in human, these agents can indeedbe detected in the peritoneal fluid immediatelyafter intestinal manipulation.32 33 Given the anato-mical location of mesenteric mast cells—that is,

Figure 3 Schematic representation of the timing of theinflammatory events triggered by abdominal surgery.COX-2, cyclooxygenase 2.

Mast cells and macrophages are the key players inthe pathophysiology of postoperative ileus.

Recent advances in basic science

Gut 2009;58:1300–1311. doi:10.1136/gut.2008.169250 1303

group.bmj.com on October 5, 2010 - Published by gut.bmj.comDownloaded from

adjacent to the mesenteric blood vessels wherethese enter the intestinal wall 34—mast cellmediators will easily diffuse into the mesentericblood vessels (fig 4). We hypothesise that thiscould explain the diffuse increase in mucosalpermeability observed after intestinal manipula-tion.37 38 When fluorescent LPS and fluorescentmicrobeads are introduced into the intestine priorto surgery, intestinal handling results in transloca-tion of fluorescent material through the mucosainto the intestinal wall. This period of increasedpermeability occurs only during a short timewindow within 3–4 h after manipulation. Oncethe beads enter the intestinal wall, they arephagocytosed by the resident macrophages ortransported to the lymph nodes via thelymphatics37(fig 4). As such, mast cell activationcould represent a key event that triggers the nextstage of the inflammatory cascade—that is, activa-tion of the resident macrophages.

Activation of resident macrophagesThe intestinal mucosa, submucosa and muscularisexterna are densely populated with several subsetsof resident phagocytes and antigen-presenting cells(APCs) of haematopoetic origin.39 Under healthyconditions, such resident macrophages are orga-nised into a layer at the level of the myentericplexus (between the longitudinal and circularmuscle layer) and in the intestinal serosa.24–26

Most of these cells possess phagocytic properties,express LPS-binding receptor CD1439 and areactivated by LPS.26 40 41 Moreover, muscularismacrophages stain for macrophage scavengerreceptor CD163, which has been shown to possess

bacteria binding and sensing capacities.42 Thisphagocyte population in the muscularis externahas an interesting nature and most probablyconsists of different subsets of APCs, includingmacrophage-like cells expressing F4/80, and den-dritic cell (DC)-like cells expressing most commonDC markers such as CD11c and DEC205.39

However, in mouse bowel wall, major histocom-patibility (MHC) II+ cells outnumber F4/80+ cellsindicating that the majority of these residentmuscularis macrophages function as phagocytesrather than APCs.

Hence, the exact cellular constituents of thephagocyte population are yet to be defined, buttheir importance in the development of intestinalinflammation following intestinal manipulationwas first demonstrated by Kalff et al12 13 (fig 4).Surgical manipulation caused an increase in resi-dent phagocytes that stained for the activationmarker lymphocyte function-associated antigen-1(LFA-1). Moreover, pharmacological or geneticdepletion (op/op mice) of resident macrophagesresulted in a decrease of inflammatory mediatorsand diminished the recruitment of leucocytes intothe muscularis.43 Moreover, macrophage-alteredanimals had near normal in vitro jejunal circularmuscle function and gastrointestinal transitdespite surgical manipulation, clearly illustratingthe importance of these phagocytes in POI.

Several potential mechanisms may trigger theactivation of these resident macrophages (fig 4).First, damage of the intestine due to manipulationwill release DAMPs,44 such as ATP, which has beenshown to be a potent activator of muscularphagocytes.45 Secondly, cytokines or LPS entering

Figure 4 Proposed mechanism involved in the inflammatory response following intestinal handling. Mast cells residingaround the mesenteric vessels are activated by intestinal handling and release vasoactive substances diffusing into theblood vessels. These substances increase the mucosal permeability, allowing the entrance of luminal bacteria orbacterial products (lipopolysaccharides) to enter the lymphatics or to interact with the Toll-like receptor (TLR) on theresident macrophages. Another route of resident macrophage activation is by binding of damage-associated molecularpatterns (DAMPs) released by damaged tissue. Binding of the TLRs or RAGE (receptor for advanced glycation end-products) activates intracellular signalling pathways (see inset) with transcription of proinflammatory genes in thenucleus. JNK, c-Jun N-terminal kinase; NF-kB, nuclear factor-kB; STATS, signal transducers and activators oftranscription.

Recent advances in basic science

1304 Gut 2009;58:1300–1311. doi:10.1136/gut.2008.169250

group.bmj.com on October 5, 2010 - Published by gut.bmj.comDownloaded from

the systemic circulation may activate the networkof resident macrophages.40 46 47 For example, injec-tion of LPS results in nuclear factor-kB (NF-kB)activation and subsequent upregulation of induci-ble nitric oxide synthase (iNOS) and cyclooxygen-ase 2 (COX-2) in resident macrophages. Finally,evidence points towards translocation of bacteriaor bacterial products such as LPS (see above). Pre-treatment with antibiotics indeed improved theinflammatory response to intestinal handling and,moreover, postoperative small intestinal smoothmuscle contractility was significantly less impaired2 h after surgery in TLR4 knock-out animals.38

Phagocytosis of the incoming bacteria and/oractivation of TLRs and RAGE by PAMPs andDAMPs stimulates macrophages to secrete proin-flammatory cytokines and chemokines. Regulationof this process is largely determined and controlledby intracellular signalling pathways involving theactivation of a series of kinases that finally lead tophosphorylation and activation of transcriptionfactors.48 The latter migrate to the nucleus to startthe transcription of proinflammatory genes (fig 4).Activation of the intracellular signalling pathwaysp38, JNK/SAP (stress-activated protein) and extra-cellular signal-regulated kinase (ERK)1/2 has beendemonstrated within 1 h after intestinal manip-ulation.49 50 After translocation of the phosphory-lated transcription factors (such as NF-kB) to thenucleus, proinflammatory cytokines (tumournecrosis factor a (TNFa), IL1b and IL6) andchemokines (MCP-1 and macrophage inflamma-tory protein-1a (MIP-1a)) are secreted by theresident macrophages leading to the upregulationof adhesion molecules (ICAM-1) in the endothe-lium and the progressive influx of leuco-cytes.12 13 28 33 38 51 The influx of leucocytes intothe muscularis starts approximately 3 h aftermanipulation, gradually increasing until 24 h post-operatively (fig 5), with monocytes, neutrophilsand mast cells as predominantly infiltratingleucocytes.11 Finally, intestinal manipulationinduces the synthesis of enzymes such as iNOSand COX-2 in the resident macrophages,52–55 aphenomenon that greatly contributes to theimpaired gastrointestinal motility which charac-terises POI.

How does surgery-induced local intestinalinflammation lead to generalised POI?Local inflammation and impaired neuromuscularfunction in POIImpaired neuromuscular function due to inflam-mation has been extensively demonstrated in avariety of inflammatory models.56 57 In POI, evi-dence has been reported that upregulation of iNOS

and COX-2 in the resident macrophages andinfiltrating leucocytes to a large extent mediatesthe blunted contractile response of inflamedtissue52–54 58 59 (fig 5). Blockade of iNOS and COX-2 indeed normalises spontaneous contractility ofmuscle strips and restores transit 24 h aftersurgery. Similar findings were obtained in micegenetically deficient in iNOS and COX-2, clearlydemonstrating that increased production of NOand prostagladins (PGs), including PGE2, is largelyresponsible for the compromised neuromuscularfunction of the manipulated and inflamed intest-inal segments.

Mechanisms leading to generalised hypomotility in POIImpaired contractility of the manipulated areasdue to local inflammation does not, however,explain the complete clinical picture. POI is mainlya condition of generalised hypomotility, whichincludes areas that have been left untouched by thesurgeon. We therefore proposed that othermechanisms must be involved and anticipated thatthe inflammatory process triggers neural pathwaysinhibiting gastrointestinal motility of distantuntouched areas. Manipulation of the mouse smallintestine indeed delayed gastric emptying eventhough no gastric inflammation was observed.Neural blockade with the ganglion blockers gua-nethidine and hexamethonium to prevent theinhibitory input to the stomach normalised gastricemptying, illustrating that the local inflammatoryresponse activates an adrenergic inhibitory path-way impairing motility of distant areas.27 This wasfurther corroborated by increased c-fos expression(a marker of neural activation) in the spinal cordand brainstem, and increased nerve activity ofspinal afferent nerves triggered by the intestinalinfiltrate 24 h after surgery.27 53 55 These findingsthus indicate that neural pathways activated bythe local infiltrate indeed play a role in thegeneralised paralysis of the gastrointestinal tract(fig 5). Interference with this mechanism by, forexample, epidural blockade with local anaestheticssuch as bupivacaine may explain the beneficialeffect observed in clinical trials, provided theblockade is positioned at the level of visceral inputto the spinal cord—that is, at the thoracic level.Lumbar or low thoracic epidural administration oflocal anaesthetics on the other hand will beclinically ineffective.60

Finally, the Bauer group demonstrated thatespecially colonic manipulation leads to pan-enteric molecular expression of proinflammatorymediators such as IL6, MCP-1, iNOS and COX-2.This is most probably due to bacterial transloca-tion into the circulation and systemic circulation ofcytokines.37 38 44 As discussed earlier, mast cellmediators diffused throughout the peritonealcavity may contribute to promote translocationof intestinal bacteria and/or bacterial productsalong the largest part of the intestine and triggera more general inflammatory response.

Taken together, it is becoming increasingly clearthat manipulation of the intestine during surgery

Inflammation of the intestinal muscularis is the keypathophysiological mechanism underlying the sec-ond long-lasting phase of postoperative ileus.

Recent advances in basic science

Gut 2009;58:1300–1311. doi:10.1136/gut.2008.169250 1305

group.bmj.com on October 5, 2010 - Published by gut.bmj.comDownloaded from

triggers a cascade of events orchestrated by theimmune system compromising neuromuscular func-tion of the entire gastrointestinal tract by recruitingboth neural and inflammatory mechanisms.

In summary, intestinal handling during surgerytriggers a cascade of events leading to intestinalinflammation. In addition to local damage of theintestine, mast cell activation triggers a brief periodof increased mucosal permeability allowing passageof intraluminal bacteria or bacterial productswhich will activate resident macrophages in themuscularis externa. The subsequent release ofcytokines and chemokines will initiate the recruit-ment of leucocytes into the intestine, impairing thecontractility of the inflamed intestine by increasedproduction of NO and COX-2 metabolites.Moreover, activation of inhibitory adrenergicneural pathways by the infiltrate will impairneuromuscular function of distant areas, explain-ing the generalised character of POI. Impairedtransit at distant areas can also result fromactivation of the resident macrophages by circulat-ing cytokines and bacterial products or evenbacteraemia induced by manipulation of theintestine further contributing to the generalisedimpairment of gastrointestinal motility.

Intestinal inflammation and POI: evidence inhumansComparably with the rodent models, several linesof evidence support that the pathophysiologicalmechanisms described above also apply to thehuman situation. Mast cell mediators are detectedin peritoneal lavage fluid very early during surgery.Even very gentle inspection of the intestines at thebeginning of the abdominal procedure increasedthe level of peritoneal tryptase.33 In contrast,patients undergoing a laparoscopic or a vaginalhysterectomy hardly showed an increase in tryp-tase. Inflammation induced by handling of theintestine is also demonstrated in human tissue.Intestinal tissue removed during surgery shows

activation of resident macrophages and time-dependent induction of IL6, IL1b, TNFa, iNOS,COX-2, ICAM-1 and LFA-1.33 61 In line with this,increased levels of the cytokines TNFa, IL6, IL8 andIL10 have been documented in the peritoneal fluidand blood of patients undergoing abdominalsurgery.33 62 Next, influx of leucocytes was clearlydemonstrated in intestinal tissue removed at theend of the surgical procedure and in tissue obtainedfrom re-operated patients.33 61 Finally, in vivorecruitment of radiolabelled leucocytes to theintestine was demonstrated in patients undergoingconventional abdominal surgery, but not inpatients undergoing a laparoscopic procedure.33

As in rodents, inflammation significantlyimpairs the neuromuscular function of the intes-tine. Spontaneous contractile activity and theresponse to bethanechol of muscle strips frompatients re-operated on 24 or 48 h after an initialabdominal procedure was significantly impaired,and was restored by incubation with specificpharmacological blockers of iNOS and COX-2.61

Moreover, we found that delay of clinical recoverywas correlated with the influx of radiolabelledleucocytes into the intestine.33 Although certainlymore evidence confirming the role of intestinalinflammation in man is awaited, the data are verysuggestive that drugs targeting the inflammatorycascade described here may be effective in reducingPOI.

IMPLICATIONS FOR THE TREATMENT OF POIFor an extensive review on the preventive techni-ques and treatment of POI, the reader is referred toexcellent reviews specifically dealing with thisissue.5 60 It is important to stress, however, thefact that for new drugs to enter the clinical arena,they will have to prove their clinical benefit againstor in combination with the current new andexciting initiatives in perioperative patient care.In particular the fast track programme, a multi-modal approach for patients undergoing colonicsurgery, has proven to reduce significantly the rateof perioperative morbidity, hospital stay and costs.5

In this programme, several perioperative mea-sures—that is, restricted fluid management, opti-mised analgesia, forced patient mobilisation andearly oral feeding—are introduced into patientmanagement with impressive results. Most prob-ably fluid restriction and an effective epiduralanalgesia are the key factors determining theoutcome.63 To what extent a similar improvementis achieved in other types of surgery and whetherthe fast track programme can easily be implemen-ted in a general surgical ward remains to bedetermined. Nevertheless, these studies clearly

Figure 5 Mechanisms underlying the impaired contractility of the intestine followingabdominal surgery. Activated resident macrophages release inflammatory cytokines andchemokines. This leads to upregulation of endothelial adhesion molecules such asintercellular adhesion molecule-1 (ICAM-1), attracting leucocytes to invade into theintestinal muscularis externa. These leucocytes and the resident macrophages produceslarge amounts of nitric oxide (NO) and prostaglandins (PGs) (by upregulation of induciblenitric oxide synthase (iNOS) and cyclooxygenase 2 (Cox-2)), impairing the contractileactivity of the smooth muscle cells. In addition, PGs activate and increase the sensitivityof spinal afferents contributing to the generalised picture of postoperative ileus.LFA-1, lymphocyte function-associated antigen-1 (LFA-1).

Also in humans, intestinal handling induces influx ofleucocytes underlying postoperative ileus.

Recent advances in basic science

1306 Gut 2009;58:1300–1311. doi:10.1136/gut.2008.169250

group.bmj.com on October 5, 2010 - Published by gut.bmj.comDownloaded from

illustrate that postsurgical recovery can be signifi-cantly improved with relatively simple and cheapmeasures. The focus of the following paragraphswill be on new therapeutic approaches based onnew insights into the pathophysiology of POI.

Accepting that the ‘‘inflammatory’’ or ‘‘secondprolonged’’ phase of POI is clinically most relevant,treatment should preferentially aim to prevent orreduce the inflammatory response to intestinalhandling. It is important to stress, however, thatinterference with the immune response may havedevastating effects on the first defence againstbacterial infection and perhaps even more impor-tantly on wound healing. The latter is of greatclinical importance as increased risk of anastomoticleak is the most feared consequence of anyimmune-modulating therapy. Even if drugs proveto be safe, ideally handling of the intestine shouldbe prevented or minimised, most probably explain-ing the shortened POI reported after minimallyinvasive or laparoscopic procedures. Moreover, onewould prefer to prevent rather than to treatinflammation, again provided that treatment doesnot interfere with the healing process or does notlead to an increased risk of infectious complica-tions. Interference early in the inflammatorycascade may also be more effective compared withdrugs administered at a later stage when inflam-mation is well established and a variety ofinflammatory mediators are released. Therefore,given the fact that mast cells and macrophagesinitiate and to a large extent orchestrate thecascade of events, these immune cells seem to bethe most interesting targets for treatment.

Mast cells as a target for treatmentStabilisation of mast cells has been proven success-ful in our mouse model of POI.32 Pretreatment withthe mast cell stabilisers ketotifen and doxantrazolesignificantly reduced the release of mast cellmediators in the peritoneal cavity, impaired theinflammatory response to intestinal handling andprevented POI. Based on these findings, weconducted a dose-finding pilot study on 60 patientsundergoing major abdominal surgery for gynaeco-logical malignancy with standardised anaesthesia,randomised to oral treatment with ketotifen (4 or12 mg) or placebo.64 Gastric retention 1 h afterliquid intake was significantly reduced by thehighest dose compared with placebo. Abdominalcramps improved significantly in patients treatedwith 12 mg of ketotifen, whereas other clinicalparameters were unaffected. Although this study ispromising, there is much room for improvement.As peritoneal mast cells should preferentially betargeted, peritoneal lavage with mast cell stabilisersmay be a more effective approach. These experi-ments are currently underway.

Resident macrophages as a target for treatmentIn rodents, pharmacological depletion and inacti-vation of the resident macrophages by chlodronateliposomes reduces the upregulation of inflammatory

mediators and adhesion molecules, restoring normaltransit in the postoperative period.43 These dataunderscore that reducing the activity of the residentmacrophages may indeed represent an interestingnew approach to treat POI. This can be achieved byseveral different strategies (fig 6).

The haem oxygenase/CO pathwayThe Bauer group has reported abundant evidencethat activation of the haem oxygenase-1 (HO-1)pathway or carbon monoxide (CO) potentlyprevents POI. HO-1 is the rate-limiting enzymein the degradation of haem, converting haem intoiron ions, biliverdin and CO. It is highly inducibleunder conditions of oxidative stress, as in ischae-mia/reperfusion injury and inflammatory condi-tions, and is one of the cell’s first lines of defenceagainst oxidative stress.65 Inhalation66 67 or intra-peritoneal lavage with CO68 or intraperitonealinjection of the water-soluble CO-releasing mole-cule CORM-349 are all effective in reducingintestinal inflammation and preventing POI.CORM-3 is most probably active via reduction ofoxidative stress induced by intestinal manipulationand further induction of HO-1 in a p38-dependentmanner.49 As CORM-3 did not result in increasedlevels of carboxyhaemoglobin and can easily beadministered prior to or during surgery, this type ofdrug is very attractive to prevent POI.

The cholinergic anti-inflammatory pathwayAnother interesting approach is to downregulateresident macrophages and thereby prevent theonset of the inflammatory cascade via activationof the cholinergic anti-inflammatory pathway(CAIP) (fig 6). The latter was recently describedby Tracey and co-workers and proposed as apowerful endogenous anti-inflammatory sys-tem.69 70 In sepsis models, electrical stimulation ofthe vagus nerve potently prevented LPS-inducedsepsis, reduced splenic TNF production andincreased survival.71 72 This effect is mediated byacetylcholine interacting with a7 nicotinic acet-ylcholine receptors (nAChRs) located on macro-phages and can be mimicked by specific receptoragonists.73 74 Recently, we extended the concept ofthe CAIP to the gastrointestinal tract.51 75 Vagusnerve stimulation dampened the levels of themacrophage inflammatory mediators TNF, IL6,MIP-2 and MIP-1a in the peritoneal fluid 3 h afterabdominal surgery. In addition, the inflammatoryresponse of the muscularis in the manipulatedintestine was reduced and gastric emptying 24 hafter surgery was normalised. This beneficial effectof vagus nerve stimulation was mediated by activa-tion of nicotinic receptors on macrophages presentin the intestinal muscularis.51 Most importantly, weshowed that these macrophages are in close contactwith cholinergic nerve fibres, providing anatomicalevidence supporting the concept of cholinergicinnervation of the resident macrophages in theintestinal muscularis. Given its anti-inflammatorypotency, activation of the CAIP or mimicry of its

Recent advances in basic science

Gut 2009;58:1300–1311. doi:10.1136/gut.2008.169250 1307

group.bmj.com on October 5, 2010 - Published by gut.bmj.comDownloaded from

effect by a7nAChR agonists may represent a newapproach for the treatment of POI.

Activation of the CAIP resulting in dampeningof inflammation can be achieved by activation ofthe motor neurons of the vagus nerve in thebrainstem by intracerebroventricular administra-tion of agents such as acetylcholinesterase inhibi-tors,76 muscarinic agonists,77 ghrelin78 or the p38inhibitor CNI-1493.69 79 Intracerebroventricularinjection of CNI-1493 reduced POI and intestinalinflammation in our POI model, an effect abolishedby vagotomy, illustrating that the anti-inflamma-tory effect of CNI-1493 is mediated via the vagusnerve.80 Interestingly, peripheral administration ofthese agents results in intracerebral concentrationshigh enough to activate the CAIP,79 making thesedrugs very interesting tools to treat inflammatoryconditions. Most probably, although the prokineticeffect of ghrelin will certainly contribute, thebeneficial effect of ghrelin in a recent phase II trialfor POI could also be explained by this mechanism.

An alternative approach to activate the CAIP isto stimulate the vagus nerve via a more physiolo-gical approach, for example via enteral feeding.High fat enteral nutrition was reported to exert ananti-inflammatory effect in a model of haemor-rhagic shock mediated via CCK-mediated activa-tion of the vagus nerve.81 To what extent thebeneficial effect of gum chewing82 on POI ismediated via activation of the vagus nerve isspeculative, but certainly deserves further study.Similarly, early introduction of feeding aftersurgery, an important aspect of fast track surgery,5

may contribute to the faster clinical recovery ofPOI after fast track surgery.

Finally, specific agonists of the a7nAChR,mimicking the effect of CAIP activation, could bean interesting new class of anti-inflammatory

drugs.74 Recently, we demonstrated that thespecific a7nAChR agonist AR-R1779 indeedreduced POI in mice via a peripheral effect.83 Itshould be emphasised though that nicotine morepotently dampened cytokine production in vitro,as compared with a7 agonists, suggesting that thatother nAChRs or cells may mediate the in vivoeffects of AR-R1779.83 The latter is supported byrecent experiments failing to demonstrate a7nAChRtranscripts in mouse peritoneal macrophages.84

Intracellular signalling pathwaysGiven the fact that intestinal handling activatesmacrophages with concomitant activation ofmitogen-activated protein kinase (MAPK) signal-ling p38, ERK1/2 and JNK MAPK49 50 and NF-kBsignalling,50 interference with these pathways maybe an interesting alternative approach to treat POI(fig 6). Semapimod, a p38 MAPK inhibitor indeedreduced POI by dampening the expression of theproinflammatory genes MIP-1a, IL6, MCP-1 andICAM-1. Moreover, it inhibited NO production inthe muscularis externa by its NOS-inhibitingproperties,50 contributing to a further improvementof gastrointestinal transit. Lastly, we and othersshowed that semapimod also activates the choli-nergic anti-inflammatory pathway,79 80 providing athird mechanism of action, making this drug a veryattractive tool to prevent POI.

Activation of cytosolic transcription factors byphosphorylation is a crucial step in the initiation ofthe synthesis of proinflammatory mediators.Interference at this level therefore will also dampenthe release of a variety of mediators and willtheoretically be more efficient than blockade of onecytokine or chemokine. A single bolus injection oftyrphostin AG 126, a protein tyrosine kinaseinhibitor, before surgery reduced the transcriptionof genes encoding the proinflammatory mediatorsIL1b, MCP-1, iNOS and COX-2, and significantlyinhibited the activation of the transcription factorNF-kB.85 In line with these findings, mice lackingthe early growth response gene-1 (Egr-1) had adownregulation in the release of NO, prostanoids,MCP-1, MIP-1a, IL6, IL1 and granulocyte colony-stimulating factor, as well as a decrease in therecruitment of leucocytes into the manipulatedmuscle wall of the intestine compared with wild-type mice.86 To what extent interference with theactivation of transcription factors carries a risk ofseriously compromising the immune response ingeneral, leading to serious side effects, certainlyrequires further study. Nevertheless, these datacontribute further to the insight that prevention of

Figure 6 Potential treatment routes to inhibit the activation of resident macrophages(see text). Ach, acetylcholine; AchR, acetylcholine receptor; ICAM-1, intercellularadhesion molecule-1; IL10; interleukin 10; LFA-1, lymphocyte function-associatedantigen; TGFb, transforming growth factor b.

Activation or mimicry of the cholinergic anti-inflammatory pathway may be an interesting newapproach to treat postoperative ileus.

Recent advances in basic science

1308 Gut 2009;58:1300–1311. doi:10.1136/gut.2008.169250

group.bmj.com on October 5, 2010 - Published by gut.bmj.comDownloaded from

the inflammatory response to intestinal handling isindeed an effective strategy to treat POI.

Other new treatments for POIOnce the inflammatory cascade is initiated, adhe-sion molecules such ICAM-1 will be upregulated toattract leucocytes from the circulation. Antibodiesor antisense molecules to, for example, ICAM-1may therefore be an elegant approach to down-regulate leucocyte trafficking and prevent theinflux of leucocytes into the intestine. In rodents,this approach indeed improved smooth musclecontractility in vitro and normaliced gastric emp-tying and intestinal transit in vivo.13 27 28 Antisenseinhibition of ICAM-1 expression has proven safeand well tolerated in several trials performed inpatients with inflammatory bowel disease, unfor-tunately with varying success.87

Given the impact of the metabolites of COX-2on smooth muscle function and their stimulatoryeffect on intestinal afferents, selective inhibitorsadministered prior to surgery should be clinicallyeffective. In rodents, the COX-2 inhibitor DFU andthe genetic absence of COX-2 indeed preventedPOI and diminished the leucocytic infiltrate by 40–50%.53 In current clinical practice, non-steroidalanti-inflammatory drugs are usually given in thepostoperative phase to reduce the use of opioidsand thereby contribute to shortening of POI,5 butbased on the current knowledge earlier adminis-tration of COX-2 inhibitors (before surgery) mayperhaps be even more efficient.

Prokinetics have been advocated as potentialtherapy for several decades; however, no clinicalstudies are available providing solid evidenceunderscoring this.60 In fact, one can seriouslyquestion whether prokinetics can overcome the‘‘brake’’ exerted by the intestinal infiltrate. Drugswith combined prokinetic and anti-inflammatoryproperties, such as ghrelin, on the other hand, willtheoretically have much greater potency to shortenPOI effectively. Most studies with ghrelin or theghrelin agonist RC-113988 89 in rodents, however,have studied the effect of these agents on the acutephase of POI, which is clinically less important.Very recently, another ghrelin agonist TZP-101was shown to enhance recovery of transit up to48 h after surgery,90 but unfortunately no data oninflammation were provided. A phase IIb clinicaltrial with TZP-101 has recently been completed.Over 200 patients were randomised in a multi-national double blind fashion receiving eitherplacebo or TZP-101 within the first hour aftersurgery for up to 7 days. The median time to firstbowel movement was reduced from 89.6 h to 68 h

for the highest dose tested (480 mg/kg) (http://www.drugs.com). Although confirmation is cer-tainly awaited, these data look very promising.

Finally, peripherally acting m-opioid receptorantagonists such as methylnaltrexone1 and alvimo-pan9 10 are extensively studied as treatment of POI,of which alvimopan has been approved in the USAto accelerate the time to upper and lower gastro-intestinal recovery after bowel resection. It appearsthough that these agents only block the detrimentaleffects of opioids on motility but do not interactwith the underlying mechanisms described above.91

The fact that alvimopan shortened the time todischarge compared with placebo in several phase IIItrials10 therefore most probably results from antag-onism of the opioid-induced delay in recovery.

SUMMARYPOI is an inevitable consequence of abdominalsurgery caused by a combination of several factorssuch as the use of pharmacological agents (anaes-thetics, opioids) in the perioperative period, neuralmechanisms and intestinal inflammation. From aclinical and therapeutic point of view, the inflam-matory response of the intestine following manip-ulation during surgery is without any doubt themost important pathophysiological mechanism.During the past decade, insight into the underlyingmechanisms has grown exponentially and will leadto new pharmacological treatments in the nearfuture. It should be emphasised though that themechanisms described here mainly if not solelyapply to uncomplicated POI following electiveabdominal surgery. In prolonged POI due to leakageor other complications, additional mechanisms willbe activated, requiring a different therapeuticapproach. The same applies for surgical proceduresperformed under acute emergency conditions on, forexample, polytrauma patients or patients enteringthe operating room in hypovolaemic shock. In thesecases, the inflammatory cascade of events hasalready started and other strategies such as admin-istration of anti-inflammatory cytokines such asIL1044 92 may perhaps offer better results.

As discussed earlier, drugs interfering with theinflammatory response carry great potency toshorten POI and thus hospitalisation. Futureclinical trials will have to prove this conceptfurther and will have to evaluate the effect ofadministration of these agents before and shortlyafter surgery on clinical recovery. If the fast trackprogramme, however, proves to be effective for anytype of abdominal surgery and can be introduced inany clinical ward, clearly these new drugs will haveto prove their superiority against this new approachof perioperative patient care. Nevertheless, the nearfuture will be bright with several potential newdrugs and targets in the pipeline.

Funding: GEB (VICI) and WJdJ (VIDI) are supported by grants from theNetherlands Organisation for Scientific Research (NWO), and GEB by agrant (Odysseus programme, G.0905.07) of the Flemish ‘‘FondsWetenschappelijk Onderzoek’’ (FWO).

Competing interests: None.

It remains questionable whether prokinetics canovercome the ‘‘brake’’ on intestinal motility exertedby the intestinal infiltrate.

Recent advances in basic science

Gut 2009;58:1300–1311. doi:10.1136/gut.2008.169250 1309

group.bmj.com on October 5, 2010 - Published by gut.bmj.comDownloaded from

REFERENCES1. Kraft MD. Methylnaltrexone, a new peripherally acting mu-opioid

receptor antagonist being evaluated for the treatment ofpostoperative ileus. Expert Opin Investig Drugs 2008;17:1365–77.

2. Delaney CP. Clinical perspective on postoperative ileus and theeffect of opiates. Neurogastroenterol Motil 2004;16(Suppl 2):61–6.

3. Livingston EH, Passaro EP Jr. Postoperative ileus. Dig Dis Sci1990;35:121–32.

4. Artinyan A, Nunoo-Mensah JW, Balasubramaniam S, et al.Prolonged postoperative ileus—definition, risk factors, andpredictors after surgery. World J Surg 2008;32:1495–500.

5. Kehlet H. Postoperative ileus—an update on preventivetechniques. Nat Clin Pract Gastroenterol Hepatol 2008;5:552–8.

6. Delaney CP, Senagore AJ, Viscusi ER, et al. Postoperative upperand lower gastrointestinal recovery and gastrointestinal morbidityin patients undergoing bowel resection: pooled analysis of placebodata from 3 randomized controlled trials. Am J Surg2006;191:315–9.

7. Bayliss WM, Starling EH. The movements and innervation of thesmall intestine. J Physiol 1899;24:99–143.

8. Benson MJ, Wingate DL. Ileus and mechanical obstruction. In:Kumar D, Wingate D, eds. An illustrated guide to gastrointestinalmotility. London: Churchill Livingston, 1993:547–66.

9. Delaney CP, Wolff BG, Viscusi ER, et al. Alvimopan, forpostoperative ileus following bowel resection: a pooled analysis ofphase III studies. Ann Surg 2007;245:355–63.

10. Curran MP, Robins GW, Scott LJ, et al. Alvimopan. Drugs2008;68:2011–9.

11. Kalff JC, Buchholz BM, Eskandari MK, et al. Biphasic response togut manipulation and temporal correlation of cellular infiltrates andmuscle dysfunction in rat. Surgery 1999;126:498–509.

12. Kalff JC, Schraut WH, Simmons RL, et al. Surgical manipulation ofthe gut elicits an intestinal muscularis inflammatory responseresulting in postsurgical ileus. Ann Surg 1998;228:652–63.

13. Kalff JC, Carlos TM, Schraut WH, et al. Surgically inducedleukocytic infiltrates within the rat intestinal muscularis mediatepostoperative ileus. Gastroenterology 1999;117:378–87.

14. Plourde V, Wong HC, Walsh JH, et al. CGRP antagonists andcapsaicin on celiac ganglia partly prevent postoperative gastricileus. Peptides 1993;14:1225–9.

15. Zittel TT, Rothenhofer I, Meyer JH, et al. Small intestinalcapsaicin-sensitive afferents mediate feedback inhibition of gastricemptying in rats. Am J Physiol 1994;267:G1142–5.

16. Zittel TT, Reddy SN, Plourde V, et al. Role of spinal afferents andcalcitonin gene-related peptide in the postoperative gastric ileus inanesthetized rats. Ann Surg 1994;219:79–87.

17. Zittel TT, Lloyd KC, Rothenhofer I, et al. Calcitonin gene-relatedpeptide and spinal afferents partly mediate postoperative colonicileus in the rat. Surgery 1998;123:518–27.

18. Barquist E, Bonaz B, Martinez V, et al. Neuronal pathwaysinvolved in abdominal surgery-induced gastric ileus in rats.Am J Physiol 1996;270:R888–94.

19. Bonaz B, Plourde V, Tache Y. Abdominal surgery induces Fosimmunoreactivity in the rat brain. J Comp Neurol1994;349:212–22.

20. Zittel TT, De GR, Brecha NC, et al. Abdominal surgery induces c-fos expression in the nucleus of the solitary tract in the rat.Neurosci Lett 1993;159:79–82.

21. Tache Y, Monnikes H, Bonaz B, et al. Role of CRF in stress-relatedalterations of gastric and colonic motor function. Ann NY Acad Sci1993;697:233–43.

22. Boeckxstaens GE, Hirsch DP, Kodde A, et al. Activation of anadrenergic and vagally-mediated NANC pathway in surgery-induced fundic relaxation in the rat. Neurogastroenterol Motil1999;11:467–74.

23. Bueno L, Fioramonti J, Ruckebusch Y. Postoperative intestinalmotility in dogs and sheep. Am J Dig Dis 1978;23:682–9.

24. Mikkelsen HB, Mirsky R, Jessen KR, et al. Macrophage-like cellsin muscularis externa of mouse small intestine:immunohistochemical localization of F4/80, M1/70, and Ia-antigen.Cell Tissue Res 1988;252:301–6.

25. Mikkelsen HB. Macrophages in the external muscle layers ofmammalian intestines. Histol Histopathol 1995;10:719–36.

26. Mikkelsen HB, Larsen JO, Hadberg H. The macrophage system inthe intestinal muscularis externa during inflammation: animmunohistochemical and quantitative study of osteopetrotic mice.Histochem Cell Biol 2008;130:363–73.

27. de Jonge WJ, van den Wijngaard RM, The FO, et al.Postoperative ileus is maintained by intestinal immune infiltratesthat activate inhibitory neural pathways in mice. Gastroenterology2003;125:1137–47.

28. The FO, de Jonge WJ, Bennink RJ, et al. The ICAM-1 antisenseoligonucleotide ISIS-3082 prevents the development ofpostoperative ileus in mice. Br J Pharmacol 2005;146:252–8.

29. Lotze MT, Zeh HJ, Rubartelli A, et al. The grateful dead: damage-associated molecular pattern molecules and reduction/oxidationregulate immunity. Immunol Rev 2007;220:60–81.

30. Foell D, Wittkowski H, Roth J. Mechanisms of disease: a ‘‘DAMP’’view of inflammatory arthritis. Nat Clin Pract Rheumatol2007;3:382–90.

31. Petrilli V, Dostert C, Muruve DA, et al. The inflammasome: adanger sensing complex triggering innate immunity. Curr OpinImmunol 2007;19:615–22.

32. de Jonge WJ, The FO, van der CD, et al. Mast cell degranulationduring abdominal surgery initiates postoperative ileus in mice.Gastroenterology 2004;127:535–45.

33. The FO, Bennink RJ, Ankum WM, et al. Intestinal handling-inducedmast cell activation and inflammation in human postoperative ileus.Gut 2008;57:33–40.

34. Coldwell JR, Phillis BD, Sutherland K, et al. Increasedresponsiveness of rat colonic splanchnic afferents to 5-HT afterinflammation and recovery. J Physiol 2007;579:203–13.

35. Echtenacher B, Mannel DN, Hultner L. Critical protective role ofmast cells in a model of acute septic peritonitis. Nature1996;381:75–7.

36. Bueno L, Fioramonti J, Delvaux M, et al. Mediators andpharmacology of visceral sensitivity: from basic to clinicalinvestigations. Gastroenterology 1997;112:1714–43.

37. Schwarz NT, Beer-Stolz D, Simmons RL, et al. Pathogenesis ofparalytic ileus: intestinal manipulation opens a transient pathwaybetween the intestinal lumen and the leukocytic infiltrate of thejejunal muscularis. Ann Surg 2002;235:31–40.

38. Turler A, Schnurr C, Nakao A, et al. Endogenous endotoxinparticipates in causing a panenteric inflammatory ileus after colonicsurgery. Ann Surg 2007;245:734–44.

39. Flores-Langarica A, Meza-Perez S, Calderon-Amador J, et al.Network of dendritic cells within the muscular layer of the mouseintestine. Proc Natl Acad Sci USA 2005;102:19039–44.

40. De Winter BY, van NL, De Man JG, et al. Role of oxidative stressin the pathogenesis of septic ileus in mice. NeurogastroenterolMotil 2005;17:251–61.

41. Eskandari MK, Kalff JC, Billiar TR, et al. Lipopolysaccharideactivates the muscularis macrophage network and suppressescircular smooth muscle activity. Am J Physiol 1997;273:G727–34.

42. Fabriek BO, van BR, Deng DM, et al. The macrophage scavengerreceptor CD163 functions as an innate immune sensor for bacteria.Blood 2009;113:887–92.

43. Wehner S, Behrendt FF, Lyutenski BN, et al. Inhibition ofmacrophage function prevents intestinal inflammation andpostoperative ileus in rodents. Gut 2007;56:176–85.

44. Bauer AJ. Mentation on the immunological modulation ofgastrointestinal motility. Neurogastroenterol Motil 2008;20(Suppl1):81–90.

45. Ozaki H, Kawai T, Shuttleworth CW, et al. Isolation andcharacterization of resident macrophages from the smooth musclelayers of murine small intestine. Neurogastroenterol Motil2004;16:39–51.

46. Hori M, Kita M, Torihashi S, et al. Upregulation of iNOS by COX-2in muscularis resident macrophage of rat intestine stimulated withLPS. Am J Physiol Gastrointest Liver Physiol 2001;280:G930–8.

47. Torihashi S, Ozaki H, Hori M, et al. Resident macrophagesactivated by lipopolysaccharide suppress muscle tension andinitiate inflammatory response in the gastrointestinal muscle layer.Histochem Cell Biol 2000;113:73–80.

48. Hommes DW, Peppelenbosch MP, van Deventer SJ. Mitogenactivated protein (MAP) kinase signal transduction pathways andnovel anti-inflammatory targets. Gut 2003;52:144–51.

49. De Backer O, Elinck E, Blanckaert B, et al. Water-soluble CO-releasing molecules reduce the development of postoperative ileusvia modulation of MAPK/HO-1 signalling and reduction of oxidativestress. Gut 2009;58:347–56.

50. Wehner S, Straesser S, Vilz TO, et al. Inhibition of p38 mitogen-activated protein kinase pathway as prophylaxis of postoperativeileus in mice. Gastroenterology 2009;136:619–29.

51. de Jonge WJ, van der Zanden EP, The FO, et al. Stimulation ofthe vagus nerve attenuates macrophage activation by activatingthe Jak2–STAT3 signaling pathway. Nat Immunol 2005;6:844–51.

52. Kalff JC, Schraut WH, Billiar TR, et al. Role of inducible nitric oxidesynthase in postoperative intestinal smooth muscle dysfunction inrodents. Gastroenterology 2000;118:316–27.

53. Kreiss C, Birder LA, Kiss S, et al. COX-2 dependent inflammationincreases spinal Fos expression during rodent postoperative ileus.Gut 2003;52:527–34.

Recent advances in basic science

1310 Gut 2009;58:1300–1311. doi:10.1136/gut.2008.169250

group.bmj.com on October 5, 2010 - Published by gut.bmj.comDownloaded from

54. Schwarz NT, Kalff JC, Turler A, et al. Prostanoid production viaCOX-2 as a causative mechanism of rodent postoperative ileus.Gastroenterology 2001;121:1354–71.

55. Mueller MH, Glatzle J, Kampitoglou D, et al. Differentialsensitization of afferent neuronal pathways during postoperativeileus in the mouse jejunum. Ann Surg 2008;247:791–802.

56. Ohama T, Hori M, Ozaki H. Mechanism of abnormal intestinalmotility in inflammatory bowel disease: how smooth musclecontraction is reduced? J Smooth Muscle Res 2007;43:43–54.

57. Grossi L, McHugh K, Collins SM. On the specificity of alteredmuscle function in experimental colitis in rats. Gastroenterology1993;104:1049–56.

58. Bauer AJ, Boeckxstaens GE. Mechanisms of postoperative ileus.Neurogastroenterol Motil 2004;16(Suppl 2):54–60.

59. Kreiss C, Toegel S, Bauer AJ. Alpha2-adrenergic regulation of NOproduction alters postoperative intestinal smooth muscledysfunction in rodents. Am J Physiol Gastrointest Liver Physiol2004;287:G658–66.

60. Holte K, Kehlet H. Postoperative ileus: a preventable event.Br J Surg 2000;87:1480–93.

61. Kalff JC, Turler A, Schwarz NT, et al. Intra-abdominal activation ofa local inflammatory response within the human muscularisexterna during laparotomy. Ann Surg 2003;237:301–15.

62. van Berge Henegouwen MI, van der PT, van Deventer SJ, et al.Peritoneal cytokine release after elective gastrointestinal surgeryand postoperative complications. Am J Surg 1998;175:311–6.

63. Muller S, Zalunardo MP, Hubner M, et al. A fast-track programreduces complications and length of hospital stay after opencolonic surgery. Gastroenterology 2009;136:842–7.

64. The FO, Buist MR, Lei A, et al. The role of mast cell stabilization inthe treatment of postoperatve ileus: a pilot study.Am J Gastroenterol 2009 Epub ahead of print.

65. Bauer M, Huse K, Settmacher U, et al. The heme oxygenase–carbon monoxide system: regulation and role in stress responseand organ failure. Intensive Care Med 2008;34:640–8.

66. Moore BA, Overhaus M, Whitcomb J, et al. Brief inhalation oflow-dose carbon monoxide protects rodents and swine frompostoperative ileus. Crit Care Med 2005;33:1317–26.

67. Moore BA, Otterbein LE, Turler A, et al. Inhaled carbon monoxidesuppresses the development of postoperative ileus in the murinesmall intestine. Gastroenterology 2003;124:377–91.

68. Nakao A, Schmidt J, Harada T, et al. A single intraperitonealdose of carbon monoxide-saturated Ringer’s lactate solutionameliorates postoperative ileus in mice. J Pharmacol Exp Ther2006;319:1265–75.

69. Oke SL, Tracey KJ. From CNI-1493 to the immunologicalhomunculus: physiology of the inflammatory reflex. J Leukoc Biol2008;83:512–7.

70. Gallowitsch-Puerta M, Tracey KJ. Immunologic role of thecholinergic anti-inflammatory pathway and the nicotinicacetylcholine alpha 7 receptor. Ann NY Acad Sci2005;1062:209–19.

71. Borovikova LV, Ivanova S, Zhang M, et al. Vagus nervestimulation attenuates the systemic inflammatory response toendotoxin. Nature 2000;405:458–62.

72. Wang H, Yu M, Ochani M, et al. Nicotinic acetylcholine receptoralpha7 subunit is an essential regulator of inflammation. Nature2003;421:384–8.

73. Borovikova LV, Ivanova S, Nardi D, et al. Role of vagus nervesignaling in CNI-1493-mediated suppression of acute inflammation.Auton Neurosci 2000;85:141–7.

74. Ulloa L. The vagus nerve and the nicotinic anti-inflammatorypathway. Nat Rev Drug Discov 2005;4:673–84.

75. van der Zanden EP, Boeckxstaens GE, de Jonge WJ. The vagusnerve as a modulator of intestinal inflammation. NeurogastroenterolMotil 2009;21:6–17.

76. Pavlov VA, Parrish WR, Rosas-Ballina M, et al. Brainacetylcholinesterase activity controls systemic cytokine levelsthrough the cholinergic anti-inflammatory pathway. Brain BehavImmun 2009;23:41–5.

77. Pavlov VA, Ochani M, Gallowitsch-Puerta M, et al. Centralmuscarinic cholinergic regulation of the systemic inflammatoryresponse during endotoxemia. Proc Natl Acad Sci USA2006;103:5219–23.

78. Wu R, Dong W, Ji Y, et al. Orexigenic hormone ghrelin attenuateslocal and remote organ injury after intestinal ischemia–reperfusion.PLoS ONE 2008;3:e2026.

79. Bernik TR, Friedman SG, Ochani M, et al. Pharmacologicalstimulation of the cholinergic antiinflammatory pathway. J ExpMed 2002;195:781–8.

80. The FO, van der Vliet J, de Jonge WJ, et al. Central activation ofthe cholinergic anti-inflammatory pathway shortens postoperativeileus in mice. Br J Pharmacol. In press.

81. Luyer MD, Jacobs JA, Vreugdenhil AC, et al. Enteraladministration of high-fat nutrition before and directly afterhemorrhagic shock reduces endotoxemia and bacterialtranslocation. Ann Surg 2004;239:257–64.

82. Vasquez W, Hernandez AV, Garcia-Sabrido JL. Is gum chewinguseful for ileus after elective colorectal surgery? A systematicreview and meta-analysis of randomized clinical trials.J Gastrointest Surg 2009;13:649–56.

83. The FO, Boeckxstaens GE, Snoek SA, et al. Activation of thecholinergic anti-inflammatory pathway ameliorates postoperativeileus in mice. Gastroenterology 2007;133:1219–28.

84. Van der Zanden EP, Snoek SA, Heinsbroek SE, et al. Vagus nerveactivity augments intestinal macrophage phagocytosis via nicotinicacetylcholine receptor a4b2. Gastroenterology 2009 [Epub aheadof print].

85. Moore BA, Turler A, Pezzone MA, et al. Tyrphostin AG 126inhibits development of postoperative ileus induced by surgicalmanipulation of murine colon. Am J Physiol Gastrointest LiverPhysiol 2004;286:G214–24.

86. Schmidt J, Stoffels B, Moore BA, et al. Proinflammatory role ofleukocyte-derived Egr-1 in the development of murinepostoperative ileus. Gastroenterology 2008;135:926–36.

87. Philpott JR, Miner PB Jr. Antisense inhibition of ICAM-1expression as therapy provides insight into basic inflammatorypathways through early experiences in IBD. Expert Opin Biol Ther2008;8:1627–32.

88. Trudel L, Tomasetto C, Rio MC, et al. Ghrelin/motilin-relatedpeptide is a potent prokinetic to reverse gastric postoperative ileusin rat. Am J Physiol Gastrointest Liver Physiol 2002;282:G948–52.

89. Poitras P, Polvino WJ, Rocheleau B. Gastrokinetic effect of ghrelinanalog RC-1139 in the rat. Effect on post-operative and onmorphine induced ileus. Peptides 2005;26:1598–601.

90. Fraser GL, Venkova K, Hoveyda HR, et al. Effect of the ghrelinreceptor agonist TZP-101 on colonic transit in a rat model ofpostoperative ileus. Eur J Pharmacol 2009;604:132–7.

91. Schmidt J, Stoffels B, Nazir A, et al. Alvimopan and COX-2inhibition reverse opioid and inflammatory components ofpostoperative ileus. Neurogastroenterol Motil 2008;20:689–99.

92. Kobbe P, Schmidt J, Stoffels B, et al. IL-10 administrationattenuates pulmonary neutrophil infiltration and alters pulmonaryiNOS activation following hemorrhagic shock. Inflamm Res2009;58:170–4.

Recent advances in basic science

Gut 2009;58:1300–1311. doi:10.1136/gut.2008.169250 1311

group.bmj.com on October 5, 2010 - Published by gut.bmj.comDownloaded from

National Cancer Institute, in the USA, with aminimum of 12 nodes, the required number.However, while the association betweensurvival and node number is evident, thetheory of stage migration is lacking insupport.8 Furthermore, statistical analysis ofthe actual number of nodes needed toaccurately determine nodal status (85% prob-ability) puts the figure closer to 30.9 This datacoupled with large-scale figures from the USshowing that less than 50% of institutionsadhere to the guidelines brings the radicallymph node dissection theory into doubt10

Therefore, it would appear that once more wehave provided a flawed explanation for a veryreal finding. The question remains: What is itthat is causing increasing survival?

LYMPH NODES: NUMBER VERSUSFUNCTIONRather than scrutinise surgical techniques it ispossible that the answer lays in the patholo-gist’s laboratory and the processes behindlymph node retrieval. The search for lymphnodes is primarily performed by vision alone.Fat clearing techniques are much vaunted butin reality are rarely used. Accordingly, largernodes are easier to find whether they areinfiltrated or not and patients with biggernodes will have higher lymph node counts. Itis possible that it is the ease by which nodesare found rather than their absolute numberthat has a bearing on prognosis.11 Thisconcept is neither new nor controversial yetis consistently overlooked in the considera-tion of the role of lymphatics in cancercontrol. It is counter-intuitive to think thatcancer causes no immune response and thatlymphatics act only as vehicles of malignantspread, yet this is the role to which they aremost commonly assigned. Innate immuneresponse to colorectal cancer has been shownto be an independent prognostic indicator,12

possibly superior to our current tumour/node/metastasis staging system. Genes asso-ciated with surveillance and immune responseare differentially downregulated in moreadvanced rectal tumours, indicating predic-tion of tumour invasion based on geneticprofiling of the primary cancer.13

Faced with these exciting advances, thesearch for lymph nodes purely to register theirabsolute number represents a missed oppor-tunity to gain real insight into prognosis.

NEOADJUVANT THERAPY AND LYMPHNODES: AN UNHAPPY ALLIANCEA further note of caution, regarding the useof lymph node number as a measure ofquality, must be raised in the era ofneo-adjuvant therapies. Combination radia-tion and chemotherapeutics cause tumourregression in a significant proportion ofrectal cancers.14 This can result in tumourshrinkage allowing for sphincter-saving sur-gery or, in a small cohort, complete clinical

and histological response which may negatethe need for surgery altogether.15 The anti-tumoural effects are not confined to therectum, however, and have been shown tocause inversion of regional lymph nodes.Not surprisingly, TME specimens frompatients post neo-adjuvant therapy consis-tently contain fewer nodes.16 Clearly, thesurgical procedure is not the dependentvariable; however, the current guidelines donot allow for this ever-growing patientcohort. Indeed we have no data to confirmwhether lymph node number post neo-adjuvant therapy still impacts on survival,the premise on which the guidelines arebased. While the need for long-term pro-spective data is implicit the unfortunatedrive toward surgeon assessment based onlymph node retrieval rates behoves us toaddress the issue now. Are we creating anenvironment where the need to produceconsistently high nodal harvests may impacton our operative timing or compromise theuse of chemo/radiotherapy? Sphincter pre-servation has untold positive impact onpatient quality of life and it is essential thatany questioning of this intervention isperformed in an evidence-based manner.Regardless of the uncertain validity of anisolated quality measure in the age of multi-disciplinary patient management, the appli-cation of guidelines based on an incompar-able patient cohort must raise concern.

CONCLUSIONThere is an over-riding drive to find para-meters of quality in surgery for rectal cancer.The need for standardised care motivates thesearch and wide variance in survival datavalidates it. When faced with such animperative, solutions are required; however,it may now be the time to hasten slowly. Thehigh standards we seek for our patients mustalso be applied in the ongoing search forperformance benchmarks and quality control.

R Kennelly, D C Winter

Institute for Clinical Outcomes Research and Education(ICORE), St Vincent’s University Hospital, Dublin, Ireland

Correspondence to: Dr R Kennelly, Institute for ClinicalOutcomes Research and Education (ICORE), St Vincent’sUniversity Hospital, Dublin 4, Ireland; kennellyrory@yahoo.ie

Competing interests: None.

Study background and organisationRK is a registrar in general surgery with a special interest incolorectal disease. DCW is a consultant laparoscopiccolorectal surgeon in a university teaching hospital. Thisarticle originated from a discussion following amultidisciplinary meeting in our hospital. DCW noted thatinternational guidelines were quoted by all interest groupsand we wanted to satisfy ourselves that the literaturesupported this uniform acceptance. Analysis of the literaturewas performed by RK and both authors contributed todrafting and editing. DCW is guarantor.

Provenance and peer review: Not commissioned;externally peer reviewed.

Gut 2010;59:139–140. doi:10.1136/gut.2009.185611

REFERENCES1. Morris E, Quirke P, Thomas JD, et al. Unacceptable

variation in abdominoperineal excision rates for rectalcancer: time to intervene? Gut 2008;57:1690–7.

2. Rich T, Gunderson LL, Lew R, et al. Patterns ofrecurrence of rectal cancer after potentially curativesurgery. Cancer 1983;52:1317–29.

3. Heald RJ, Ryall RDH. Recurrence and survival aftertotal mesorectal excision for rectal cancer. Lancet1986;i:1497–82.

4. Macfarlane JK, Ryall RDH, Heald RJ. Mesorectalexcision for rectal cancer. Lancet 1993;341:457–60

5. Reynolds JV, Joyce WP, Dolan J, et al. Pathologicalevidence in support of total mesorectal excision in themanagement of rectal cancer. Br J Surg1996;83:1112–5.

6. Swanson RS, Compton CC, Stewart AK, et al. Theprognosis of T3N0 colon cancer is dependent on thenumber of lymph nodes examined. Ann Surg Oncol2003;10:65–71.

7. Chang GJ, Rodriguez-Bigas MA, Skibber JM, et al.Lymph node evaluation and survival after curativeresection of colon cancer: systematic review. J NatlCancer Inst 2007;99:433–41.

8. Wong SL, Ji H, Hollenbeck BK, et al. Hospital lymphnode examination rates and survival after resection forcolon cancer. JAMA 2007;298:2149–54.

9. Joseph NE, Sigurdson ER, Hanlon AL, et al. Accuracyof determining nodal negativity in colorectal cancer onthe basis of the number of nodes retrieved onresection. Ann Surg Oncol 2003;10:213–8.

10. Baxter NN, Virnig DJ, Rothenberger DA, et al. Lymphnode evaluation in colorectal cancer patients: apopulation-based study. J Natl Cancer Inst2005;97:219–25.

11. Murphy J, Pocard M, Jass JR, et al. Number andsize of lymph nodes recovered from dukes B rectalcancers: correlation with prognosis and histologicantitumor immune response. Dis Colon Rectum2007;50:1526–34.

12. Galon J, Costes A, Sanchez-Cabo F, et al. Type,density, and location of immune cells within humancolorectal tumors predict clinical outcome. Science2006;313:1960–4.

13. Grade M, Hormann P, Becker S, et al. Geneexpression profiling reveals a massive, aneuploidy-dependent transcriptional deregulation and distinctdifferences between lymph node-negative and lymphnode-positive colon carcinomas. Cancer Res2007;67:41–56.

14. Crane CH, Skibber JM, Feig BW, et al. Response topreoperative chemoradiation increases the use ofsphincter-preserving surgery in patients with locallyadvanced low rectal carcinoma. Cancer2003;97:517–24.

15. Habr-Gama A, Perez RO. Non-operativemanagement of rectal cancer after neoadjuvantchemoradiation. Br J Surg 2009;96:125–7.

16. Rullier A, Laurent C, Capdepont M, et al. Lymphnodes after preoperative chemoradiotherapy for rectalcarcinoma: number, status, and impact on survival.Am J Surg Pathol 2008;32:45–50.

CORRECTION

doi:10.1136/gut.2008.169250corr1

G E Boeckxstaens, W J de Jonge.Neuroimmune mechanisms in postoperativeileus. Gut 2009;58:1300–11. On page 1303,‘‘mMCP-1 (murine monocyte chemoattrac-tant protein)’’ should read ‘‘mMCP-1 (mur-ine mast cell protease 1)’’.

PostScript

140 Gut January 2010 Vol 59 No 1

group.bmj.com on October 5, 2010 - Published by gut.bmj.comDownloaded from