segmentalnecrotising pathological clinicalfeatures of22 ... · biutler, dalims, litidpaitittier,...

Gut, 1987, 28, 1433-1438

Segmental necrotising enterocolitis: pathological andclinical features of 22 cases in BangladeshT BUTLER, B DAHMS, K LINDPAINTNER, M ISLAM, M A K AZAD,AND P ANTON

Froml the Inttern(ational Centrefor Diarrhoeal Disease Research, Batngladesh; anid Departments of Medicinle atIdPatlology, Case Western Reserve University, Cleveland, Ohio, USA

SUMMARY To describe the pathology and clinical features of segmental necrotising enterocolitis(SNE) in children and adults, 22 diarrhoeal patients (median age two years, range two months to 50years) in Bangladesh with this lesion detected at autopsy were examined and compared with twogroups of diarrhoeal control patients. Gross pathology consisted of purplish or black mucosal or

transmural discoloration with erosions or ulcerations in segments of the jejunum or ileum of 18cases and of the colon alone in four cases. Two patients had intestinal perforations. Microscopicallyall specimens showed coagulation necrosis or haemorrhagic necrosis indicative of mucosalischaemia. In 20 cases there was submucosal oedema and nine showed pneumatosis of the bowel.From 1 1, one or more of the invasive diarrhoeal pathogens Shigella, Campylobacter andEntamoeba histolytica were detected. From the comparison with controls significant associationswere found for a long duration of diarrhoea, blood and mucus in stool, abdominal distension or

tenderness, shock not attributable to hypovolaemia, septicaemia, and low concentration of serumprotein (p<0-05). These findings indicated that segmental necrotising enterocolitis developssometimes as a fatal complication of prolonged diarrhoeal illnesses associated with shock andhypoproteinaemia and is caused by ischaemic injury to the intestinal mucosa.

Segmental necrotising enterocolitis (SNE) is an acuteand often fatal complication of infectious diarrhoea,usually occurring in children in developing countries.This intestinal disease is a recently recognised entity,which has been reported as identical clinical syn-dromes or closely related variants under the names of'segmental infarcts of the small intestine' and 'acutesegmental ischaemic enteritis' in Thailand," 'acutesegment.al necrotising enteritis' and 'acute ischemicenteritis' in India,' 'necrotising enterocolitis' inKuwait,' 'necrotising enteritis' in Sri Lanka," and'enteritis necroticans' (pig-bel) in New Guinea.Although various aetiologic agents have been pro-posed for SNE, including round worms,4 the n-toxinof Clostridium perfringens,7 and sweet potato trypsininhibitor,2the pathogenesis of SNE remains unclear.

In an autopsy series of diarrhoeal patients recentlydone in Bangladesh" 22 patients with SNE wereidentified. We describe in this report for the first time

Address for corrcspondence: Dr I Butler, Dcpt of Internal Medicine, l'exasTech University Health Sciences C'enter, Lubbock, Tcxas 79431). USA.Received for publication 9 April 1987.

the findings of SNE among diarrhoea patients inBangladesh and compare systematically the clinicalfeatures of patients with SNE to the features ofsimilar groups of diarrhoeal patients who did notdevelop SNE.

Methods

PATI ENTSThe International Centre for Diarrhoeal DiseaseResearch in Dhaka, Bangladesh maintains a hospitalfor diarrhoeal diseases. Twenty two cases of SNEwere identified among 140 autopsies on diarrhoealpatients carried out between May 1982 and May1985. Autopsies were carried out as early as writteninformed consent could be obtained from families.Bodies were stored in a cold room before theautopsy. The time after death that autopsies weredone ranged from one to 48 hours, with a mediantime of seven hours. In all cases the diagnosis of SNEwas made att autopsy. All patients showed one ormore discrete segments of the gastrointestinal tractwith purple or black mucosal discolourations and

1433

on February 4, 2021 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.28.11.1433 on 1 N

ovember 1987. D

ownloaded from

http://gut.bmj.com/

Biutler, Dalims, Litidpaitittier, Islam, Az-ad, atiditAtitoi

erosions or ulcers recognised on gross examination.Microscopic examination used routinely formalinfixed and paraffin embedded sections stained withhaematoxylin and eosin. All cases showed acutemucosal ischaemia with coagulation necrosis (ghostlike villi) or haemorrhagic mucosal necrosis. Someinflammation was usually also present in the affectedareas.

Clinical features of SNE cases were obtained byreviewing clinical records. To determine whichfeatures were significantly associated with SNE,retrospective case control analyses were performed.Two groups of diarrhoeal controls were selected forcomparison. One group was 44 autopsy controls thatwere selected from the same series of autopsies bychoosing two patients without SNE with serialautopsy numbers as close to each SNE case aspossible. The other control group consisted of 44surviving patients. Using a random number table toselect serial hospital numbers, four controls wereselected for each of the first 11 SNE cases, admittedduring May 1982 to March 1983, from among surviv-ing diarrhoeal patients admitted to the intensive careunit during the same two week period. Criteria foradmission were signs of serious illness includingrespiratory distress, shock, altered mental status,seizures, and malnutrition. During the followingperiod May 1983 to May 1985 an additional 11 casesof SNE were detected at autopsy. Chart reviewwas then performed for cases and controls. Datacollected included information concerning age,sex, history, physical examination, and laboratorydata. Shock was defined as absent or thready radialpulse that persisted after intravenous infusion ofisotonic fluid in an amount :5°/o of the body weight.Septicaemia was defined by positive antemortemblood cultures for pathogenic bacteria or positivepostmortem blood cultures if taken less than sixhours after death. Mean values were determined forboth SNE cases and control groups and compared byStudent's t test. Frequencies were evaluated by the X'test or Fisher exact test.

Results

PATHOl.OGIC FINDINGS IN SNE

On opening the abdominal cavity at autopsy one ormore discrete segments of intestine were noted to bepurple or black in colour. This was transmural andapparent on the serosal surface in 15 patients (Fig.1); in the remaining seven the mucosa alone wasinvolved. The small intestine, particularly distaljejunum and ileum, was the most common site of thesegmental necrosis. In eight individuals, lesions werelimited to small intestinal segments. Ten patients hadinvolvement of both small intestine and colon. In

only four patients was the colon alone involved.When the bowel was opened, the involved segmentsrevealed deeply congested or haemorrhagic mucosawith superficial erosions or ulcerations. The length ofinvolved segments was highly variable, some patientshad patchy involvement with ischaemic areas onlyseveral centimetres long. In half the patients, onethird or more of the length of the small intestine wascongested and haemorrhagic. In one infant the entiresmall intestine was grossly purple. Two adult patientsdemonstrated one or more perforations through theinvolved segments (one small intestinal, one colonic).

Microscopically the lesions most closely resembledischaemic bowel disease. Foci of mucosa in all casesshowed coagulative necrosis, usually with ghost likevilli devoid of distinct cellular outlines and nucleardetail (Fig. 2). In some cases only villus tips wereinvolved; more advanced lesions showed completemucosal and even submucosal extension. Haemor-rhagic necrosis was also seen focally in all cases (Fig.3) and is responsible for the dark red to black colourwhich allows for gross recognition of involved intest-inal segments. Usually this coagulative and haemor-rhagic necrosis was combined with inflammation(Fig. 2). In some cases where only the superficialmucosa showed coagulative necrosis, the inflam-mation was subjacent, with acute and chronicinflammatory cells filling the viable lamina propria.In cases with coagulative necrosis of the entiremucosa, the inflammation was apparent immediatelyadjacent. No bacteria were identified in these areas,even though they may have been present elsewherein the gastrointestinal tract. Submucosal oedema waspresent in 20 of the cases, expanding this layer of thebowel to approximately two or three times normalthickness (Fig. 3). Nine patients demonstratedextensive pneumatosis intestinalis of the submucosa(Fig. 4). These submucosal gas bubbles were presentin both small intestine and colon (small intestinealone in four patients, colon alone in one patient,both in four patients). The muscularis propria wasintact except in two patients with transmuralgangrene and intestinal perforation. Thrombi werenot seen in any large blood vessels in the intestinesexcept in one patient who had hyphae of mucor-mycosis invading blood vessels in necrotic lesions ofthe terminal ileum and ascending colon."' Rarecapillary platelet fibrin thrombi were observed in thesubmucosa in areas of coagulative necrosis. Thesewere secondary to the necrosis and not aetiologic.

PATHOILOGY OF OTHER TISSUES

Mesenteric lymph nodes were enlarged in 14 casesand showed increased numbers of polymorpho-nuclear leukocytes. In two cases the lymph nodesshowed erythrophagocytosis. The livers of 18

1434



j.com/


ovember 1987. D

ownloaded from

http://gut.bmj.com/

Segmental necroti.sing entero(olitis: pathological (a1(1 clinlicalfeatlures of22 cases in Banigladesh

[.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~%a i 6

1 - ^l a- a Pz. ff .

1435

0 10 15cmFig. I Gross photograph (fsegmental necrotising enterocolitis shows di.stension and black di.scolouration oflong and shortsegments of small intestine. Intervening lighter areas are spared.

'vbi ,:f,

Fig. 2 Coagulative necrosis of .small intestin7al mucosa, seenl at the centre is characterised by ghost like villi devoid ofcellulardetail. Also note superficial epithelium lifting offmucosal surface to the right. (Haematoxylin and eosin.)



j.com/


ovember 1987. D

ownloaded from

http://gut.bmj.com/

Blutler, Dalmtnis, Lindl)piittier, lslamtti, Azad, nd(lAnton

Ar ,. . _.

i_ --v . _-

Fig. 3 Haeinorrhagic Inw((os.at/ nec rosis is seen at the ulpperrigght. Cell/lar detail of surf ice and glanidular epithelium isefa'ced, a(ltd lamnina propria isfilled with erythrocytes,ialipirtitlg a dark colour to the affected area. The submucosashiows. pronounced oedema. Themuscularis laYersarenotinvolved. (Haematoxyli/n anid eositl.)

p'atients showed fatty infiltration. The degree of fattyinfiltration was marked in 1() cases, moderate in six

cases, and mild in two cases.

C INIC AL ANI) ILABORATORY FINDINGS

The mediain age of the SNE cases was two years

(range two months to 50 years) and there were more

female than male patients (Table). Eleven patientswere hospitalised during the summer months July toSeptember. Segmental necrotising enterocolitispatients reported a meain duration of diarrhoea ofII -9 days aind were hospitalised for a mean of threedays before death. Most of the patients with SNEhad presented with blood and mucus in stool withvomiting and signs of dehydration. Most of themdeveloped abdominal distension or tenderness andshock not attributed to saline depletion. Fever(T-37-8XC on the day of admission) was detected in

nine of the SNE patients. One or more of the invasivediarrheal pathogens Shigella sp, Emttanioebahistolytica, and Campylobactersp were detected in 11cases. Cholera was diagnosed in none of the SNEcases. Septicaemia was diagnosed in eight cases(36%/). The mean white blood cell count was17,300/mm. and the mean concentration of serumprotein was 4-5 g/100 ml. Metabolic acidosis withtotal serum CO, concentrations less than 20) mmol/lwas detected in 16 patients with SNE.Comparison of the SNE cases with autopsy diar-

rhoeal controls revealed comparable features exceptfor a greater frequency of blood in stool of the SNEcases (Table) (p<0-05). On the other hand, compari-son of SNE cases with surviving diarrhoeal controls,showed that SNE cases had a longer mean duration ofdiarrhoea before admission, greater frequencies ofblood and mucus in stool, abdominal distension ortenderness, shock, and septicaemia, and a lowermean concentration of serum protein (p<t).05).

Discussion

Our patients with SNE were mostly children over awide range of ages after the neonatal period. Theyhad been ill with diarrhoea, vomiting and fever for amean period of about 12 days, which is longer than thethree to four days of illness described in Thailand,'as well as in India"' and Kuwait.' Most of our patientspresented with dysenteric diarrhoea, in contrast withthe patients in Thailand who had watery diarrhoea.'The portion of the bowel most commonly affected bysegmental necrosis was the jejunum, followed by theileum and colon. The microscopic pathology found inour cases was ischaemic or haemorrhagic necrosis,with oedema and without significant thrombosis ofvessels, affecting the musoca and, sometimes, the fullthickness of the intestinal wall and is typical of thedisease described in other countries. '' Other vari-able features, including acute inflammation of thebowel, pneumatosis, and perforation, were some-times present in our cases. None of our patientsshowed vasculitis or fibrinoid necrosis of intestinalblood vessels, as was described in Sri Lanka.'The aetiology of SNE is not known. Heaidington et

a!' who described the clinical and pathologic findingsof this condition in 1967, iegarded it as segmentalinfarction. Most of the patients reported fromThailand' ' were not in shock at presentation and hadno vessel occlusions or other discernible vascularlesions to explain the pathologic resemblance toischemic bowel disease. There is a striking similarityof the pathology of SNE to neonatal necrotisingenterocolitis (NEC), a disease largely of prematureinfants in intensive care units."' The anatomic distri-bution of disease in NEC is most commonly the

1436



j.com/


ovember 1987. D

ownloaded from

http://gut.bmj.com/

Segmtiental necrotising enterocolitis: pathological and clinicalfeatures of22 cases in Bangladesh

Fig. 4 Severe pneumatosis intestinalis is present in the submucosa ofthe small intestine. (Haematoxylin and eosin.)

terminal ileum and proximal colon, however incontrast the jejunum in SNE. The combination ofischaemic and inflammatory findings in NEC has ledto the same aetiologic impasse in that disease: is itcaused by ischaemia, infection, or a combination offactors?

Previous efforts to identify an aetiologic agent forSNE have not succeeded. Stool cultures for entero-pathogens have shown negative results, except forone case of Salmonella."' Although Cperfringens hasbeen isolated from some patients,'7 the relevanttoxin of this organism is the 13-toxin of C perfringenstype C, which was identified in only nine of 54 isolateswith toxin from patients in New Guinea.7 Further-more, the toxin is degraded by pancreatic proteasesin the intestinal lumen. Half of our patients wereinfected with the invasive diarrhoeal pathogensShigella sp, Entamoeba histolytica, and Campylo-bacter sp. The lesions of SNE were in the jejunumand those of shigellosis and amoebiasis in the colon,indicating that these primary intestinal infections didnot directly cause the lesions of SNE.

Hypoproteinaemia probably contributed to hypo-perfusion of the intestine in some of our patients by areduction in plasma oncotic pressure. At serumconcentrations of protein below 5 g/100 ml, oedemaof tissues occurs because fluid leaves the vesselsunder normal hydrostatic pressure in the arterialcirculation and the diminished osmotic pressure inthe venous side of the capillaries fails to draw tissuefluid back into the vessels.'" Oedema of the intestinalwall, present in our cases and reported in other

studies of ischaemic enteritis,' is consistentwith this mechanism. Severe body burns and non-tropical sprue, which cause large losses of plasmaproteins, are also associated with necrosis of theintestinal mucosa."4' In the carcinoid syndromeand the acquired immunodeficiency syndrome, thepresence of diarrhoea is associated with hypo-albuminaemia. I" Thus, hypoalbuminaemia couldserve as a common stimulus for both diarrhoea andSNE through a deficit in plasma oncotic pressure.1The causes of hypoproteinaemia in our cases werenot identified but probably include decreased intakeof protein, protein losing enteropathy, malabsorp-tion, and decreased albumin synthesis by the liver.'6

Local factors in the intestine may also act topredispose the bowel to ischaemic injury. When thesmall intestine is stimulated by an enterotoxin tosecrete fluid during diarrhoea, the intestine willrequire additional perfusion to obtain fluid for secre-tion as well as to maintain its oxygen supply andnutrients. Thus, the enterotoxin stimulated jejunumor ileum may become more vulnerable to ischaemicinjury when there is not an ample reserve of perfu-sion pressure and volume. Most of our patientspresented with arterial hypotension often caused bysepticaemia, which produces a sympathetic nervousreflex leading to secretion of catecholamines.Catecholamines will act to redistribute blood flowaway from the intestine, thus aggravating or provok-ing intestinal ischaemia."' Another factor in thepathogenesis of SNE is autodigestion of the mucosaby pancreatic proteases, including trypsin, chymo-

1437



j.com/


ovember 1987. D

ownloaded from

http://gut.bmj.com/

1438 Biitler, Dahmzs, Lindlpainttner, Islatni, Azacd, atndAnitoi

Taible Cliniical and laboratory features ofcases withsegmenital necrotising enterocolitis (SNE) and diarrhoealcontrol groups without SNE*

A utop.s) SurvivingSNE c ases controls controls(n =22) (it=44) (n =44)

Median age (range),years 2 (0-2-50) 2-5 (0-1-60) 1-3 (0-1-50)

Males/females(n) 8/14 23/21 30/14Duration of diarrhoea

before admission,dayst 11-9±9-5 14-1±18-1 5-2±6-5

Percent with:Blood in stooli 55 25 11Mucus in stool§ 77 55 32Vomiting 55 45 86Dehydration 59 73 52Abdominal

distcnsion ortenderness§ 64 41 11

Shock not attributedto salinedcplction§ 55 45 11

.Shigella sp 18 32 1()Enttatnoeba histolytica 18 16 7Ca(npvylobacter.sp 18 17Vibrio cliolerae 0 9 24Scpticaemia§ 36 36 14

White blood ccll count/mm3x l(O(X) 17-3±9-1 17-2± 15-4 17-6±15-8Total scrunm protein

g/tt)t)mlt 4 5±1+1 5-0±1( 6 6-5±1 6

*Valucs are expressed as means±SD or pcr cents. Sce Methods fordcscription of control groups; t-Mean value of SNE cascssignificantly different from mean value of surviving controls byStudent's t test (p<O.(X)5); tFrequency of SNE cascs significantlygrcater than autopsy controls and surviving controls by x2 test(p<)-0)5); §Frequency of SNE cases significantly greatcr thansurviving controls by x2 and Fishcr's exact test (p<(0.0)5).

trypsin, and elastase. After the initial ischaemicinjury, the brush border of the mucosa loses itsglycoprotein protection against luminal enzymes.This autodigestion is a later event in SNE and mayexplain the mucosal erosions and ulcers in ourpatients.When clinical features of our patients with SNE

were compared with autopsy controls, only blood instools occurred more frequently in the SNE cases. Onthe other hand, when clinical features were com-pared with surviving diarrhoeal controls, the SNEcases showed a longer duration of diarrhoea, a lowerserum protein concentration, and greater frequen-cies of blood and mucus in stool, abdominal disten-sion or tenderness, septicaemia, and shock. Thissuggests that the clinical features ofSNE are the sameas those associated with a fatal outcome in the settingof diarrhoeal diseases. Thus, the emergence of SNEbefore death may be determined by crucial timing ofan altered intestinal circulation in severely illpatients.

We thank M R Islam, Peter Speelman, and JohnBanwell for their assistance with the manuscript. Thiswork was supported, in part, by a grant from UnitedNations Development Programme and World HealthOrganisation.

References

1 Headington JT, Sathornsumathi S. Simark S.Sujatanond W. Segmental infarcts of the small intestineand mesenteric adenitis in Thai children. Lancet 1967; i:802-6.

2 Welch TP, Sumitswan S. Acute segmental ischaemicenteritis in Thailand. Br J Surg 1975; 62: 716-9.

3 Kalani BP, Shekhawat NS, Sogani KC. Acute segmentalnecrotizing enteritis in children. Am J Dis Child 1985;139: 586-8.

4 Gupta SD, D'Silva V, Salelkar AV, Reys M. Acuteischaemic enteritis in Goa. BrJ Surg 1977; 64: 420-3.

5 Takayanagi K, Kapila L. Necrotising enterocolitis inolder infants. Arch Dis Child 1981; 56: 468-71.

6 Arseculeratne SN, Panabokke RG, Navaratnam C.Pathogenesis of necrotising enteritis with special refer-ence to intestinal hypersensitivity reactions. Gut 1980;21: 265-78.

7 Murrell TGC, Roth L, Egerton J, Samels J, Walker PD.Pig-bel: enteritis necroticans. A study in diagnosis andmanagement. Lancet 1966; i: 217-22.

8 Lawrence G, Walker PD. Pathogenesis of enteritisnecroticans in Papua New Guinea. Lancet 1976; i:125-6.

9 Butler T, Islam M, Azad MAK, Islam MR, Speelman P.Causes of death in diarrheal diseases: an autopsy studyof 140 patients in Bangladesh who died in hospital afterrehydration. Bull Wld Hlth Org 1987; 65: 317-23.

10 Calle S, Klatsky S. Intestinal phycomycosis (mucomyco-sis). Am J Clin Pathol 1966; 45: 264-72.

11 Zeissler J, Rassfeld-Sternberg L. Enteritis necroticiansdue to Clostridium welchii type F. Br Med J 1949; i:267-9.

12 Kliegman RM, Fianaroff AA. Necrotizing cnterocolitis.NEnglJ Med 1984; 310: 1093-103.

13 West ES, Todd WR. Textbook of Biochemistry. NewYork: The Macmillan Co, 1962: 522.

14 Patterson M, Rosenbaum HD. Enteritis necroticians.Gastroenterology 1952; 21: 110-8.

15 Bounous G. Acute necrosis of the intestincal mucosa.Gastroenterology, 1982; 82: 1457-67.

16 Mariani G, Strober W, Keiser H, Waldman TA.Pathophysiology of hypoalbuminemia associated withcarcinoid tumor. Cancer 1976; 38: 854-60.

17 Kotler DP, Gaetz HP, Lange M, Klein EB, Holt PR.Enteropathy associated with the acquired immuno-deficiency syndrome. Ann Intern Med 1984; 101: 421-8.

18 Moss G. Plasma albumin and postoperative ileus. SlurgForum 1967; 18: 333-6.

19 Bynum TE, Jacobson ED. Nonocclusive intestinalischemia. Arch Intern Med 1979; 139: 281-2.

20 Bounous G, Menard D, de Medicis E. Role of pan-creatic proteases in the pathogenesis of ischemic entero-pathy. Gastroenterology 1977; 73: 102-8.



j.com/


ovember 1987. D

ownloaded from

http://gut.bmj.com/

segmentalnecrotising pathological clinicalfeatures of22 ... · biutler, dalims, litidpaitittier,...

Documents