5 JClin Pathol 1993;46:154-158

Adenovirus and intranuclear inclusions inappendices in intussusceptionH J Porter, C J H Padfield, L C Peres, L Hirschowitz, P J Berry

AbstractAims: To examine the light and electronmicroscopic features of appendices re-moved at the time of surgical reductionof intussusception in children; and toconfirm that the viral inclusions seen insome ofthem are due to adenovirus.Methods: A series of 39 appendices fromcases of intussusception and 15 controlappendices were reviewed. Light micro-scopic examination of haematoxylin andeosin stained sections was performed onall of them and one appendix with largenumbers of inclusions was examined byelectron microscopy. Non-isotopic in situhybridisation using a biotinylated DNAprobe was carried out on sections ofappendix from 30 of the cases of intus-susception and from the 15 controls.Results: Light microscopic examinationshowed viral inclusions in 19 of theappendices from the cases of intussus-ception and in none of the controls.Electron microscopic examinationshowed virus particles with the typicalfeatures of adenovirus. Most of theappendices with viral inclusions in thehaematoxylin and eosin stained sectionsalso contained adenovirus DNA as shownby in situ hybridisation.Conclusions: Viral inclusions seen inappendices from cases of intussusceptionare caused by adenovirus. AdenovirusDNA was not demonstrable in appen-dices from cases of intussusception with-out viral inclusions and the aetiologicalfactors involved in intussusception inthese children remain unknown.

(7 Clin Pathol 1993;46:154-158)

Intussusception of the bowel may occur atany age but is particularly common in babiesand young children under the age of 2 years.Any part of the intestine may be affected butileocolic intussusceptions are the most com-mon. In adults usually a lesion such as apolyp or tumour acts as a lead point, but inchildren only rarely is such an abnormalityfound.' It has therefore been suggested thathyperplasia of the lymphoid tissue around theterminal ileum is the cause of the intussuscep-tion. The association between intussusceptionand adenovirus infection was first suggestedby Strang,2 and following this three differentstudies showed that adenovirus could berecovered from 50% or more of children with

intussusception, significantly more than incontrols.-5 More recently a study by Nicolaset al compared the incidence of adenovirusand rotavirus in 64 children with surgicallyconfirmed intussusception.6 Adenovirus wasfound in 25 of the 64 cases (39%). Serotypes1 and 2 were most common (10 cases each),followed by type 5 (three cases) and types 3and 4 (one case each).

Yunis et al performed a histological andultrastructural study of 682 appendices ofwhich 30 were from cases of ileocolic intus-susception.7 These 30 cases were examinedelectron miroscopically. They described twotypes of intranuclear changes seen on lightmicroscopy as presumptive evidence of aden-ovirus infection, namely eosinophilic nuclearinclusion bodies and "smudge cells", in whichthe nucleus is slightly enlarged, homoge-neous, and basophilic. Electron microscopicexamination showed clusters of replicatingviral particles with a dense central nucleoidsurrounded by a capsid, appearances typicalof adenovirus. In 11 of the 30 (36 6%) casesviral inclusions were demonstrated by lightand electron microscopy. Cultures wereobtained in five cases and these showed aden-ovirus serotypes 2 (three cases), 3, and 5 (onecase each).The aim of the present investigation was to

examine the light microscopic findings in aseries of appendices removed from childrenwith intussusception and to confinn the elec-tron microscopic appearance of the viralinclusions.

MethodBlocks from 39 appendices removed at thetime of surgical reduction of intussusceptionbetween 1985 and 1991 were retrieved fromthe files of Bristol Children's Hospital. Onehaematoxylin and eosin stained section con-taining tip, middle, and proximal end fromeach of these cases and from each of 15 agematched controls was examined blind by lightmicroscopy. The control cases were appen-dices removed incidentally during othersurgery for example, for Hirschsprung's dis-ease and intestinal malrotation. The featuresnoted were viral inclusions, epithelial disarraywith loss of polarity of the surface epithelium,shedding of the epithelium into the lumen,and lymphoid hyperplasia as indicated bylarge lymphoid follicles with active germinalcentres containing many tingible bodymacrophages. Electron microscopy was car-ried out on one formalin fixed appendix with

Department ofPaediatric PathologyBristol Royal Hospitalfor Sick ChildrenBristolH J PorterL C PeresP J BerryDepartnent ofHistopathology BristolRoyal InfirmaryBristolC J H PadfieldDepartment ofHistopathology StMichael's HospitalBristolL HirschowitzCorrespondence to:Dr H J PorterDepartment of PaediatricPathology, Bristol RoyalHospital for Sick Children,St Michael's Hill, BristolBS2 8BJAccepted for publication7 August 1992

154

on April 24, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.46.2.154 on 1 February 1993. D

ownloaded from

Adenovirus intranuclear inclusions in appendices in intussusception

V

Figure 1IAppendix with epithelial disarray and

intranuclear viral inclusions (haeinatoxylin and eosin).

particularly abundant viral inclusions, which

was post-fixed in osmium.

In s-itu hybridisation for adenovirus was

performed on 30 appendices from cases of

intussusception (15 with viral inclusions and

15 without inclusions by light microscopy)

and on the 15 control appendices. In situ

hybridisation was perform ed using Enzo

Pathogene Tissue Preparation Kit and Enzo

Pathogene DNA Probe Assay for identifica-

tion of adenovirus (British distributers

Cambridge Bioscience). The biotinylatedprobe is described as detecting adenovirus

serotypes 4,5,7,11,20, 40 and 41. The kituses proteinase K for digestion and 3%

hydrogen peroxide to block endogenous per-

oxidase. After denaturation at 950G, hybridis-

ation occurs at room temperature. The

detection system is avidin biotinylated horse-

radish peroxidase with aminoethyl carbazole

as the chromogen. The sections were coun-

terstained with 50% Harris's haematoxylin

for 30 seconds.

For the first batch of sections the positive

control for the in situ hybridisation was a

slide with HeLa cells infected with adenovirus

5, provided by Enzo. An appendix in the first

batch which gave a positive result was used as

positive control in subsequent in situ hybridi-sation procedures. The negative control was a

section from the positive appendix treated in

the same way as the test sections but with no

probe applied to it.

Mte clinical notes of the 39 children with

intussusception were reviewed to determinethe results of any virological investigation.

ResultsThe children were aged 1 month to 4 years

(mean 8f5 months) and all but three were 1

year or under. Table summarises the mor-

phological findings on light microscopicexamination of the appendices. Viral inclu-sions were present in 19 of the 39 appendicesfrom cases of intussusception (49%) and innone of the controls. The number of inclu-sions varied from a few to hundreds in anyone section. Most of the inclusions were ofthe eosinophilic rather than the "smudgecell" type. Epithelial disarray (fig 1) includingbudding of the epithelium into the lumen was

a consistent feature in appendices with viralinclusions but it was also found in appendicesfrom cases of intussusception without viralinclusions and in the controls, although inthese it was less prominent. Shedding of theepithelium into the lumen was present in justover half the appendices with inclusions butin only two without inclusions and in none ofthe controls. Lymphoid hyperplasia was pre-sent in most of the appendices, although itwas slightly less common in the controls.Electron microscopic examination of one case

with inclusions showed viral particles with thetypical appearance of adenovirus (fig 2).

Table 2 shows the results of the in situhybridisation. The method gave strong posi-tive staining of the inclusions with very littlebackground staining. The nuclei of surfaceepithelial cells with inclusions stained posi-tively as did cells that had been shed into thelumen (fig 3). There were also scattered posi-tive epithelial cells deeper within the glands(fig 4) and occasional positive cells in lym-phoid follicles. Both these last two groups ofpositive cells were difficult to identify inhaemotoxylin and eosin stained sections dueto their scarcity. Positive lymphoid cells werepresent in nine of the 12 appendices thatwere positive with in situ hybridisation.

Virological studies were performed on onlynine of the children with intussusception. Ofthose without viral inclusions, four patientshad virological investigations: adenovirus was

cultured from the mesenteric lymph node ofone patient, one patient had negative faeces,one a negative throat swab and one a negativerectal swab, mesenteric lymph node, andappendix. Of the five patients with inclusionswho had virological investigations, the appen-dix of one yielded adenovirus type 1, on cul-ture, although the faeces were negative. Inthree patients adenovirus was found by elec-tron microscopic examination of the faecesand in one patient the faeces were negative.There was no seasonal variation in the inci-dence of intussusception or in the cases ofintussusception in which adenovirus was

detected. There was a large increase in thenumber of intussusceptions requiring surgicalreduction in 1991 (17 cases compared with22 cases for 1985-1990 inclusive), but no

Table I Morphologicalfindings in stained sections of appendices

Lymphoid Epithelial EpithelialNumber hyperplasia disarray shedding

IntussusceptionsWith viral inclusions 19 18 19 11Without viral inclusions 20 18 1 1 2

Controls 15 1 1 8 0

155

on April 24, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.46.2.154 on 1 February 1993. D

ownloaded from

Porter, Padfield, Peres, Hirschowstz, Berry

a

Figure 2 Electron micrograph ofepithelial cell nucleus showingparticles.

Figure 3 Adenovirus in surfacdemonstrated by in situ hybridis

increase in the proportisions (six out of 17 casesof 22 cases for 1985-199

DiscussionIntussusception is the cintestinal obstruction instill has a substantial moiare ubiquitous and cautract infections, conjundiarrhoea and mesenteiclinical and virological s

relation between adencintussusception. Gardne

Table 2 In situ hybridisation

Nt

With viral inclusions 15Without viral inclusions 15Controls 1 5

Figure 4 Glandular epithelial cells containing adenovirusdemonstrated by in situ hybridisation.

children admitted with intussusception andcompared them with 62 children with diar-rhoea, 38 with pneumonia and bronchiolitisand 10 normal controls.9 Faeces and throatswabs were cultured from all children andmesenteric lymph nodes from 21 of the 38children with intussusception. Thirty five hadpaired sera examined. Adenovirus was isolat-ed from the faeces of 46% of the childrenwith intussusception. When adenovirus wasisolated from more than one site in a child itwas of the same type and the authors consid-ered that this, together with a positive anti-body response in 18 children, indicated activeand probably recent systemic infection.Adenoviruses were isolated from the faeces in

47 3-2% of children with diarrhoea, 2-6% ofthose with pneumonia and bronchiolitis, andfrom 3-7% of normal children. Two childrenwith intussusception were also found to haveherpes simplex virus and one to have poliovirus in their faeces. Ross et at4 and Potter5performed similar studies and found aden-ovirus in 61% and 63%, respectively, ofpatients with intussusception compared with

e epithelial cells between 2% and 6% of controls. There hasation. been only one large study from the United

States which documented the presence ofion with viral inclu- viral inclusions in appendices from cases offor 1991 and 13 out intussusception, and suggested that they were

attributable to adenovirus.7The finding in the present investigation of

viral inclusions in 49% of appendices fromcases of intussusception is higher than that

-ommonest cause of found in the American study but within theyoung children and range of the British virological studies. There

rtality8 Adenoviruses have been no large studies correlating histo-lse upper respiratory logical appearances with virological investiga-lCtiVitiS, pneumonia, tions. In the series of 30 appendicesric adenitis. Several examined by Yunis et al7 five patients with;tudies have shown a viral inclusions had virological culture per-~virus infection and formed on rectal swabs and adenovirus was)r et al studied 38 grown from all of them. In only one of these

cases was adenovirus detected in the faeceselectron microscopically. The histologicaldiagnosis of adenovirus is obviously subject tosampling error but the incidence in the pre-

umber ISH positive sent investigation correlates well with the12 incidence found in the larger virological stud-o ies. Although there is one case in this studyo where adenovirus was cultured from mesen-

156

on April 24, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.46.2.154 on 1 February 1993. D

ownloaded from

Adenovirus intranuclear inclusions in appendices in intussusception

teric lymph node even though histology of theappendix was negative, it seems likely thatmost cases can be diagnosed histologically ifthe appendix is removed. This study has con-

firmed the previous reports of a significantassociation between adenovirus and intussus-ception in that none of the control appen-

dices had viral inclusions.The morphological changes of epithelial

disarray and shedding are clearly not specificto cases of intussusception associated withadenovirus infection as they were found,albeit to a lesser extent, in cases of intussus-ception without viral inclusions and in con-

trols. These changes presumably simplyrepresent disturbance of the epithelium whichmay be due to a number of different causes

including the appendix being involved in theintussusception.The use of in situ hybridisation has

demonstrated that the viral inclusions are dueto adenovirus in most cases. There are severalpossible explanations for the three appendiceswith viral inclusions on haematoxylin andeosin stained sections that were negative within situ hybridisation. First, of the serotypesusually associated with intussusception,'-6 theprobe used is described as only detecting fourand five so it is possible that the inclusionsnegative for in situ hybridisation are ofserotypes 1, 2, 3 or 6. Serotypes 1 and 2 seem

to be the most common and it is thereforeperhaps surprising that as many as 12 of thecases were positive with in situ hybridisationbecause the probe is not supposed to detectthese serotypes. Secondly, there may not havebeen any viral inclusions in the particular sec-

tions on which in situ hybridisation was per-formed. Thirdly, the inclusions may havebeen caused by another virus. Against thislast possibility is the fact that the inclusions inthe in situ hybridisation negative cases hadthe same morphology on haematoxylin andeosin staining as those in the in situ hybridisa-tion positive cases. The finding of viral inclu-sions in lymphoid cells is not surprisingbecause adenovirus seems to cause lymphoidhyperplasia, can be cultured from mesentericlymph nodes, and was first isolated fromadenoids. However, this does not seem to bewell documented probably because thenumber of lymphoid cells with inclusions issmall.

There were no cases positive with in situhybridisation that were negative for inclusionson haematoxylin and eosin staining, indicat-ing that routine examination of the appendixis at least as sensitive as in situ hybridisationfor detecting viral inclusions. Thus in cases ofintussusception without viral inclusions aden-ovirus DNA is not demonstrable by in situhybridisation. This raises the question of theaetiology of the intussusception in these chil-dren which make up about 50% of the cases.

Konno et al' suggested that rotavirus may

also have a role in the aetiology of intussus-ception. There were 30 patients in their studyand the diagnosis was made on clinical andradiological findings. Faecal specimens were

obtained from all and electron microscopy

performed. Paired sera were available in 18cases and complement fixation tests were car-ried out for rotavirus and adenovirus type 2.Eleven of the 30 patients shed rotavirus(37%) compared with eight who shed aden-ovirus (27%). Serological findings showed afour-fold or greater increase in antibodies torotavirus in six patients including one whodid not shed rotavirus in the faeces.Serological evidence of adenovirus infectionwas present in 11 patients including threewho did not shed adenovirus in their faeces.The authors concluded that rotavirus andadenovirus are both associated with intussus-ception, although there were no controls inthe study and no virus isolation was per-formed.The more recent study by Nicolas et a16

examined 64 children with intussusceptionrequiring surgical reduction. Virologicalinvestigations consisted of virus isolationfrom rectal and throat swabs and serology foradenovirus and rotavirus. Twenty five (39%)patients shed adenovirus. Five patients shedother viruses (one polio, three echo, and oneherpes simplex). Six patients showed serolog-ical evidence of rotavirus infection but in fourit was associated with isolation of adenovirus.Unlike Konno et al, these authors considerthat rotavirus is not an aetiological agent inintussusception.

Other. viruses reported to be associatedwith intussusception include herpes virus 6(three cases),"I and E coli 0 1 57:H7.'12The large increase in the number of opera-

tive reductions of intussusception in Bristol in1991 seems to be the result of an increasedrate of failed enema reductions rather than anincreased number of cases of intussusceptionas a whole. The numbers are small but thereis no evidence of an increase in the numbersof intussusceptions associated with viralinclusions during 1991. This, together withthe fact that the virological studies whichincluded operated and non-operated cases,found a similar incidence of adenovirus infec-tion does not support the possibility thatcases requiring surgical reduction are morelikely to be associated with adenovirus infec-tion.

In conclusion, by use of in situ hybridisa-tion we have proved conclusively that most ofthe viral inclusions seen in the appendicesfrom children with intussusception are due toadenovirus. This investigation has shownserotypes 4 and 5 to be the most common inthe cases studied, if as is claimed by Enzo, theprobe used does not detect serotypes 1, 2, 3and 6. This is in contrast to the virologicalstudies which generally indicate thatserotypes 1 and 2 are the most common.None of the appendices that did not showviral inclusions in sections stained withhaematoxylin and eosin was positive with insitu hybridisation, indicating that straightfor-ward morphology is adequate for the diagno-sis of adenovirus infection. The role of otherpossible aetiological factors in adenovirusnegative cases of intussusception remainsunresolved.

157

on April 24, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.46.2.154 on 1 February 1993. D

ownloaded from

Porter, Padfield, Peres, Hirschowitz, Berry

The Department of Paedatric Pathology is supported by agrant from the Foundation for the Study of Infant Death.LCP is supported by a grant from the Brazilian NationalResearch Council.

1 Potter CW, Zachary RB. The etiology of intussusception.With particular attention to the adenovirus infections.Surg Clin North Am 1964;44:1509-20.

2 Strang R. Intussusception in infancy and childhood. Areview of 400 cases. Br Jf Surg 1959;46:484-95.

3 Bell TM, Steyn JH. Viruses in lymph nodes of childrenwith mesenteric adenitis and intussusception. Br Med1962;2:700-2.

4 Ross JG, Potter CW, Zachary RB. Adenovirus infection inassociation with intussusception in infancy. Lancet1962,ii:221-3.

5 Potter CW. Adenovirus infection as an aetiological factorin intussusception of infants and young children. JfPathol Bacteriol 1964;88:263-73.

6 Nicolas JC, Ingrand D, Fortier B, Bricout F. A one-year

virological survey of acute intussusception in childhood.Med Vir 1982;9:267-71.

7 Yunis EJ, Atchison RW, Michaels RH, DeCicco FA.Adenovirus and ileocecal intussusception. Lab Invest1975;33:347-51.

8 Stringer MD, Pledger G, Drake DP. Childhood deathsfrom intussusception in England and Wales, 1984-9. BrMedJ 1992;304:737-9.

9 Gardner PS, Knox EG, Court SDM, Green CA. Virusinfection and intussusception in childhood. Br Med Jf1962;2:697-700.

10 Konno T, Suzuki H, Kutsuzawa T, et al. Human rotavirusinfection in infants and young children with intussus-ception. J7Med Virol 1978;2:265-9.

11 Asano Y, Yoshikawa T, Suga S, Hata T, Yamazaki T,Yazaki T. Simultaneous occurrence of human her-pesvirus 6 infection and intussusception in three infants.Pediatr Infect DisJ7 1991;1O:335-7.

12 Lopez EL, Devoto S, Woloj M, Pickering LK, Cleary TG.Intussusception associated with Escherichia coli0157:H7. Pediatr Infect Dis _J 1989;8:471 -3.

158

on April 24, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.46.2.154 on 1 February 1993. D

ownloaded from