histopathology of fatal adenovirus infection respiratory

J. clin. Path. (1967), 20, 561

Histopathology of fatal adenovirus infection of therespiratory tract in young children

D. M. O. BECROFT

From the Princess Mary Hospital for Children, Auckland, New Zealand

SYNOPSIS Type 7 adenoviruses were isolated from lung tissue obtained at the necropsies on fivechildren aged 8 to 15 months. The clinical and radiological findings were those of bronchiolitisand pneumonia. Four children had an extensive necrotizing bronchitis and bronchiolitis. Thesurviving bronchial epithelium, and all bronchial epithelium in the fifth case, was proliferating.There was evidence that this proliferation was a virus-induced effect and not reparative. There was

virus tropism to alveolar lining cells with a pneumonia showing mononuclear cellular reaction,necrosis, and hyaline membranes. Distinctive necrotizing lesions were found in the bronchialglands of all cases. Two contrasting types of intranuclear inclusions were present in all lesions andtheir significance and inter-relationships are discussed. These pathological changes are specific forinfections by a limited number of adenovirus serotypes.

The usual clinical effects of adenovirus infectionare well defined. Various serotypes have been identi-fied as the cause of minor febrile, respiratory, con-junctival, and pharyngeal infections occurring atall ages (Stuart-Harris, 1962). Epidemics of infectionmay occur, particularly in closed communities, butotherwise the adenoviruses account for only a min-ority of respiratory illnesses in the general population(van der Veen, 1963). These infections usually aremild and there is less certainty about the role ofthe virus in causing severe disease of the lowerrespiratory tract. During adenoviral epidemicsevidence of this infection in cases presenting withpneumonia does not prove a causal relationship(Maisel, Pierce, Crawford, and Rosenbaum, 1962).On the other hand, there are a number of reportsof the isolation of the virus from the lung tissue ofchildren dying of bronchopneumonia and the simi-larity of the histopathological findings in thesecases has supported an adenoviral aetiology.

In Auckland, since 1959, type 7 adenoviruseshave been isolated from the lungs of five childrendying from acute lower respiratory tract infections.In this paper the clinical, histopathological, andepidemiological features of the five cases are des-cribed, compared with those in previous reports,and the specificity of the pathological findings isexamined further.Received for publication 24 January 1967.

CASE REPORTS

Findings in individual cases are presented in the Table.All patients were infants aged 8 to 15 months. Four weremale, two were European, and three Polynesian. Allwere well-grown, well-nourished children with no historyof undue susceptibility to respiratory infections and,with the exception of case 4, no previous illnesses ofsignificance.

CASES 1, 2, 3, AND 5 These children had closely similarillnesses. There had been respiratory symptoms of in-creasing severity for three to four days. One child de-veloped a morbilliform rash two days before the respira-tory symptoms and two had diarrhoea. On admissionthese four children had temperatures of 37-8°C. to40)6°C. and tachycardia. Distressed, rapid respirations,marked wheeze, rib recession, and auscultatory findingsled to a provisional diagnosis of bronchiolitis in allpatients. Two children had additional signs suggestingpneumonia. The chest radiographs showed a generalizedincrease in lung markings indicative of peribronchialor interstitial inflammation. The lung fields were hyper-ventilated in three patients and three had more extensiveshadowing suggesting consolidation or collapse, partic-ularly in the lower zones. Two children had mild hypo-chromic anaemias. The results of other investigationsare tabulated.

All patients were treated with antibiotics, three weregiven adrenal corticosteroids, and one child was digital-ized. One or more antispasmodics were used in eachcase. Oxygen was given, but respirations were not assisted.Treatment had little effect, fever persisted, and there

561

on Novem

ber 2, 2021 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.20.4.561 on 1 July 1967. D

ownloaded from

http://jcp.bmj.com/

D. M. 0. Becroft

TABLEFATAL CASES OF ADENOVIRUS INFECTION

Case I Case 2 Case 3

Date of admissionAge- sexRacePresenting symptoms

Clinical diagnosison admissionChest radiographs

White blood cells

Throat swabs

Other

Antibiotics

SteroidsInterval between(1) Onset and death(2) Admission and

death

August 195912 months- maleMaoriMorbilliformrash - 5 days'Respiratoryinfection' - 3 daysBronchiolitis,bronchopneumoniaIncreased lungmarkings, hyper-ventilation

15,000/c.mm.(91 % neutrophils)

Urine analysisnormalChloramphenicol,novobiocin

Prednisone, A.C.T.H.

7 days41 hours

March 19608 months- maleEuropeanCoughWheeze 4 daysDyspnoeaFever J

Bronchiolitis

Increased lungmarkings, hyper-ventilation,consolidation R.upper lobe.7,500/c.mm.(50% neutrophils)Commensals only

C.S.F. normal

Chloramphenicol,tetracyclineerythromycin

7 days3 days

June 196013 months- maleEuropeanCoughWheezeDyspnoea 3 days

Fever J

Bronchiolitis

Increased lungmarkings, hyper-ventilationconsolidation L.lower lobe

Chloramphenicol,penicillin, streptomycin

Prednisone

4 days31 hours

September 196010 months- maleNuie Islander'III, 4 or 5 days

August 196315 months - FemaleMaoriCough ? 3 weeksWheeze 3Dyspnoes 3 days

Bronchopneumonia Bronchiolitis,pneumoniaConsolidation bothlower zones

3,800/c.mm.(27 Y. neutrophils)Staphylococcus aureus(also from noseand faeces)Blood culture sterile,urine analysis normal

Penicillin (oral) Erythromycin,tetracycline,ampicillinHydrocortisone

4-5 days15 minutes

? 6 days3 days

was inexorably increasing respiratory distress withrestlessness, cyanosis, and tachycardia. Two childrenhad convulsions terminally. Although case 3 was severelyill his death was sudden and unexpected. All patientsdied within three days of admission to hospital, afterillnesses lasting four to seven days from the first symp-toms.

CASE 4 This 10-month-old infant died within 15 minutesof his admission and treatment was limited to resuscita-tive measures. The only history obtained was that hehad been ill for four to five days and had been givenpenicillin by mouth. Two weeks before he had been dis-charged from hospital after an illness diagnosed asbronchopneumonia with bronchospasm. The onset ofthis illness was similar to that described above and, al-though his recovery had appeared complete, it is possiblethat the fatal illness was a relapse of this earlier infection.

NECROPSY FINDINGS

All five children were well-developed and well-nourished and there were no skin lesions or otherexternal features of note. In two cases the pleuralcavities contained small serous effusions. All lungswere voluminous and maintained this overexpandedstate after removal from the pleural cavities. Thelungs of cases 1, 2, 4, and 5 contained multiple,mostly small, brick-red or purple-red areas of con-solidation (Fig. 1). These lesions were irregularlydistributed through all lobes, but were more ex-tensive posteriorly and in the lower lobes. Case 4was most severely affected, showing consolidation

of the greater part of both lower lobes, the rightmiddle lobe, and the posterior part of the rightupper lobe. Intervening tissue was either emphy-sematous, mildly congested and oedematous or,less frequently, atelectatic. There was no consolida-tion in the lungs of case 3, but overinflation wasparticularly severe and there was an unsually exten-sive interstitial and mediastinal emphysema (Fig. 2).Interlobular septa throughout the lungs of thiscase were expanded by large air-filled blebs andthere were bullae up to 2 cm. in diameter at the lungmargins and at the hila. The soft tissues of the anter-

FIG. 1. Case 2. Posterior aspect of the lungs showingareas of consolidation.

Case 4 Case 5

562

on Novem


http://jcp.bmj.com

/J C


ownloaded from

http://jcp.bmj.com/

Histopathology offatal adenovirus infection of the respiratory tract in young children

FIG. 2. Case 3. Severe interstitial emphysema andpneumo-mediastinum.

ior mediastinum were transformed into a honey-comb of air-filled spaces.Many bronchi were filled with exudate, usually

mucoid or fibrinous, but occasionally purulent or

haemorrhagic. The mucosa of larger bronchi was

congested and the walls of smaller bronchi were

thickened and the lumina narrowed. The mucosalsurfaces of the pharynx, larynx, and trachea in eachcase were mildly congested but there was no evidenceof severe inflammation nor of ulceration. Thetracheobronchial lymph nodes were enlarged,congested and oedematous in all cases. Lymphnodes elsewhere were of normal size. The spleensof cases 1, 4, and 5 were enlarged, but to less thantwice the normal weight for age. Altered blood was

found in the stomach, oesophagus, and respiratorypassages of case 4. Bleeding appeared to have beenfrom small pyloric ulcers, which later proved tobe due to monilial infection. These were the onlyabnormalities found outside the respiratory tract.

HISTOLOGY

LOWER RESPIRATORY TRACT The lesions common

to these cases were characterized by cellular necrosisand by distinctive intranuclear inclusions in thesurviving cells. The destructive changes frequentlywere accompanied or replaced by proliferation.The epithelium of the bronchial tree, the bronchialglands and the alveolar septa were affected particul-arly, although the severity of the lesions at eachsite differed considerably from case to case.

BRONCHI AND BRONCHIOLES In cases 1, 2, 4, and 5the changes were largely destructive and so extensive

that no surviving bronchial or bronchiolar epitheliumremained in some blocks of lung tissue. The medium-sized intrapulmonary bronchi of 1 to 2 mm. dia-meter were affected most severely and only a fewfragments of respiratory epithelium remained atthis level of the bronchial tree. The majority ofthe larger intrapulmonary bronchi were at leastpartially ulcerated and all the residual epitheliumwas abnormal. There was marked exudation fromthe ulcerated surfaces. Some of the denuded bronchiand bronchioles had narrow lumina encircled bya thick layer of homogeneous eosinophilic P.A.S.-positive material which, apart from its bulk, re-sembled the typical hyaline membrane present inthe alveolar ducts (Fig. 3). More often the exudatefilled the lumen and was pleomorphic, includingin various proportions similar hyaline material,mucus, fibrin, neutrophils, macrophages, cellulardebris, and desquamated epithelial cells, some ofthe latter bearing inclusions (Fig. 4). The ulcerationincluded the basement membrane but rarely extendedmore deeply. The congested and swollen laminapropria of the larger bronchi was infiltrated bylymphocytes, plasma cells, macrophages, and a fewneutrophils, while a similar but more intense in-filtration involved the full thickness of the wallsof many smaller bronchi and bronchioles, and theadjacent alveoli (Fig. 4). Muscle bundles and elasticfibres were widely dispersed by this inflammatoryreaction. Full-thickness necrosis was rare, however,and was confined to a few terminal and respiratorybronchioles. Otherwise these smaller bronchioleswere affected least severely. Some were entirelynormal while others, although inflamed and oftenoccluded by exudate, had an intact epithelium.No ulceration was found in the bronchial tree of

case 3, but the respiratory epithelium of all bronchiand bronchiolesshowed proliferativechanges (Fig. 5).There was similar proliferation of all residualepithelium in the larger bronchi of the other cases(Fig. 6). Cilia and goblet cells were absent through-out and a stratified arrangement of polyhedral cellswith swollen vesicular nuclei replaced the regularpseudo-stratified columnar structure. There wereoccasional mitoses in the basal layers. Many cellsnear the surface were degenerate or were desquam-ating, and these often contained inclusions (Fig. 6).Alternatively, and particularly where the epitheliumwas six or more cells thick, the surface layers weresquamous, but not keratinizing, and inclusions anddesquamating cells were infrequent.

BRONCHIAL GLANDS This was the only site at whichall cases had typical necrotizing lesions. The changewas least severe in case 1 but in the others severaldegenerate acini were found about most bronchi

563

on Novem


http://jcp.bmj.com

/J C


ownloaded from

http://jcp.bmj.com/

D. M. 0. Becroft

', 4E ;. .t'", .44'

i.r~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~A

44

ko X N.,'+e Z f s.1N4

FIG.3

.

FIG. 5.

4.

* ;.I--

., O ^ m

fat,

AW.

.wR

. .4.*44.4

IA

`..

Kti'=* 4 >

84,

." S

4,'

44'';

* .vffiiP.,wwS;4A

* .P t.tt

C'; ~ 4 sTP 4C~~~~ft44

A

FIG. 4.

FIG. 6.

FIG. 3. Case 4. Necrotizing bronchitis. A thick eosino-philic membrane lines the wall of a small bronchus.Bronchial glands are necrotic. Haematoxylin and eosinx 165.

FIG. 4. Case 1. Necrotizing bronchiolitis. The lumen isfilled withpleomorphic exudate. Alveoli are emphysematous.Haematoxylin and eosin x 65.

FIG. 5. Case 3. Proliferating bronchial epithelium.Haematoxylin and eosin x 200.

FIG. 6. Case 1. Proliferating bronchial epithelium. Manynuclei contain one or more discrete eosinophilic inclusions.Inclusion material is distributed diffusely in the nucleiofdegenerate surface cells. Haematoxylin and eosin x 230.

564

on Novem


http://jcp.bmj.com

/J C


ownloaded from

http://jcp.bmj.com/

Histopathology offatal adenovirus infection of the respiratory tract in younig children

sectioned (Figs. 3 and 7). The peripheral lobules,situated outside the cartilage plates, were particul-arly affected. Changes were focal, sometimes in-volving only a single acinus, but often several acinior a whole lobule were replaced by a granulareosinophilic coagulum. Both the mucous and theserous acinar cells had either been completely des-troyed or were degenerate and desquamating intothe lumina. Many residual cells contained inclusions(Fig. 7). The acinar basement membranes usuallyremained intact, but the lobular and perilobularconnective tissue was heavily infitrated by lym-phocytes and plasma cells. Fibrin, neutrophils,macrophages, and mucus filled other affected acini.The severity of the acinar lesions had no correlationwith the degree of epithelial destruction in the assoc-iated bronchus. Changes in the larger ducts, however,were similar to those in the bronchi.

ALVEOLI The appearances were highly variablebut the consolidated areas in cases 1, 2, 4, and 5had features in common. Alveoli and alveolarducts were distended by an eosinophilic exudatepart proteinaceous oedema fluid, part fibrin strands,part granular coagulum. Some of this material mayhave tracked down from the bronchial tree, butmuch was a result of focal necrosis of the alveolarsepta and local inflammatory exudation. Manyalveolar ducts were lined by a P.A.S.-positive hyalinemembrane. There was an interstitial and intra-alveolar infiltration by macrophages, lymphocytes,and plasma cells. Neutrophils were abundant in afew areas, but no bacteria were identified. In themost severely affected lung there was extensivedisorganization as a result of necrosis of the cellularcomponents of the alveolar walls and of inflamma-tory cells, leaving much nuclear debris mingled witheosinophilic exudate and foci of haemorrhage.Even in these areas the connective tissues of thealveolar septa were intact and no occlusive vas-cular lesions were detected. The consolidated areasoften were bounded by interlobular septa and anecrotic terminal bronchiole was identified centrallyin others. Less severely affected alveolar tissue wascongested and oedematous, with or without fibrinousexudation and mononuclear reaction. Alveolarlining cells were swollen, many to cuboidal shape,particularly in the alveoli immediately adjacent toinflamed bronchiolar walls. Inclusion-bearing cellswere abundant in the necrotic areas. Some cellswere lying free but others were attached to remnantsof the alveolar septa and in the less degeneratetissue it was clear that inclusions were developingin the swollen alveolar lining cells (Fig. 8a). Otherfree-lying, inclusion-bearing cells may have beenmacrophages or cells desquamated from the respira-

tory epithelium. No section of the lung was entirelynormal, as in addition to widespread bronchialand bronchiolar lessions other alveolar tissue waseither emphysematous or atelectatic (Fig. 4). Therewas severe alveolar and interstitial emphysemathroughout the lungs of case 3.

INCLUSIONS Two types of nuclear change wereattributed to viral proliferation. The first was adiffuse accumulation in the nucleus of homogeneous,amphophilic or basophilic, phloxinophilic, stronglyFeulgen-poitive material (Figs. 7b and 8a). Thiswas not sharply demarcated from the surroundingnuclear chromatin. The second was a discrete, eosin-ophilic, phloxinophilic, but Feulgen-negative in-clusion, surrounded by a clear zone or 'halo' in thenuclear chromatin (Figs. 6 and 8b). These differedfrom the type A viral inclusions of herpes simplexor varicella in that the outlines were irregular andpoorly demarcated and the surrounding 'halo' in-complete (Fig. 8b). There was a marked differencein the size of the two types of inclusions. The majorityof discrete eosinophilic inclusions were less than 5,uin diameter and the nuclei were only slightly en-larged. By contrast the diffuse Feulgen-positiveinclusions filled nuclei ranging from normal sizeup to giant forms as much as 14[t in diameter. Thelatter were a striking feature even under low magnifi-cation (Fig. 7a). Amphophilia was more evident inthe larger inclusions, but Feulgen-positivity per-sisted in the majority. A thin rim of chromatin re-mained about the large inclusions, sometimesseparated by a narrow pale zone, but a distinct 'halo'was unusual. Many inclusions in desquamatingor degenerate cells were crenated, vacuolated, orfragmenting. With increasing eosinophilia of cyto-plasm and disappearance of all nuclear basophilia,other cells were transformed into homogeneouseosinophilic bodies accompanying the debris withinthe lumina of many alveoli, glands, and bronchioles.The interrelationships of the two types of intra-

nuclear inclusions were not clear. Both types werefound together in some tissues and could thereforehave been different stages of a single developmentalsequence (Fig. 8a). However, the large diffuse in-clusions predominated in alveoli and in glands anda sequence could be traced from normal nucleito the largest inclusions, without the discrete eosin-ophilic inclusions appearing to be a necessary partof this sequence. Alternatively, the eosinophilicinclusions often occurred alone in the proliferatingbronchial epithelium and here a second sequencecould be traced from the appearance of small, some-times multiple, eosinophilic particles within thevesicular chromatin, through Feulgen-negative in-clusions of increasing size until the nucleus was

565

on Novem


http://jcp.bmj.com

/J C


ownloaded from

http://jcp.bmj.com/

566

*S,s, Al,..

..;...gri.A X ..s%; Wi F * Z Xt~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.

'ft '.'r " s .t$fN4!, '

* 4r8t#'*}

FIG. ~ 4 ~ ~ 47a.

* .~~~~~~~~~~~~~~~~~~~~~~~~

FIG. 7a.

FIG. 7. Case 2. Bronchial glands. Haematoxylin and eosin. (a) Some acini are completely necrotic. Other acinar cells

contain large basophilic inclusions x 200.(b) Detail of (a) showing large basophilic inclusions and necrotic material in lumen x 675.

FIG. 8. Case 5. Intranuclear inclusionsin alveolar tissue Haematoxylin andeosin. (a) Four adjacent alveolar liningcells containing inclusions. The dis-crete eosinophilic inclusion with 'halo'contrasts with the large diffusebasophilic or amphophilic inclusionsx 1,000. (b) An eosinophilic inclusionat higher magnification to show irregularoutline and incomplete 'halo' x 2,000.

FIG. 8b.

D. M. 0. Becroft

*. *

FIG. 7b.

FIG. 8a.

.:

"'i.11?Am.a

"'

on Novem


http://jcp.bmj.com

/J C


ownloaded from

http://jcp.bmj.com/


eventually filled or the cell degenerated (Fig. 6).The discrete eosinophilic inclusions occasionallymay have developed from the diffuse type, but thereverse probably did not occur and in general thetwo types appeared to represent alternative andlargely independent reactions of the nucleus.

UPPER RESPIRATORY TRACT The epithelium of thetrachea in case 1 showed proliferative changes andintranuclear inclusions similar to those in the bronchi.In other cases the tracheal epithelium had lost ciliaand had a tattered appearance, but there was noulceration and no inclusions were detected. Thetracheal glands in cases 4 and 5 showed typicalnecrotizing lesions with inclusions. Sections fromthe tongue, posterior pharynx, tonsils, and larynxin three cases revealed minor subepithelial inflamma-tory changes, but there were no specific lesions.

OTHER ORGANS The swollen tracheo-bronchiallymph nodes present in all cases had dilated sinusescontaining many large macrophages. Althoughoccasional cells had large hyperchromatic nucleino inclusions were detected. Lymphoid tissue else-where was normal. The liver, spleen, kidney, pancreas,adrenals, thyroid, and myocardium were sectionedin all cases and the small and large intestines infour. The only abnormality found was patchycentrilobular necrosis in the liver of case 5 but thisappeared due to anoxia and not to virus infection.No evidence of meningo-encephalitis was found inthe three patients in whom the central nervoussystem was examined.

VIROLOGY

Viruses were isolated from lung tissue obtained atthe necropsies of all five cases and from a throatswab taken from case 5 on the day of her admissionto hospital. Primary inoculations were into Hela orHEp 2 tissue cultures in which cytopathic changesappeared promptly and on passage. The virusesisolated from cases 1, 2, 3, and 4 were investigatedfurther at the Department of Virus Diseases, NationalInstitutes of Health, Bethesda, Md., where all wereidentified as type 7 adenoviruses by haemagglutina-tion inhibition techniques (Rosen, 1960). Theadenoviruses isolated more recently from the throatand lung of case 5 were identified as type 7 byneutralization techniques (Grayston, Johnston,Smith, and Loosli, 1956). The antisera used wereprepared by a technique similar to that describedby Rowe, Huebner, Hartley, Ward, and Parrott(1955).

DISCUSSION

The extent of the necrotizing lesions in bronchialand bronchiolar epithelium determined the clinicalcourse of four patients. The initial clinical andradiological findings were those of bronchiolitisand the patients' death followed progressivelysevere anoxia largely due to airway obstruction.It is doubtful if even the best modern methods ofrespiratory management could have maintainedpatency of the bronchial tree against such an out-pouring of exudate. The presence of viral inclusionsin alveolar lining cells proved a true viral pneumoniain each case, but clinically the alveolar lesionsappeared to be of secondary importance.The fifth patient had an intact bronchial epithelium

and might have survived had he not suddenly de-veloped interstitial pulmonary emphysema andpneumomediastinum. In this case there was wide-spread proliferation of the bronchial epitheliumwhich, in the absence of ulceration, clearly was notreparative. In all cases the presence of viral inclusionsin the proliferating cells indicated that this changewas virus induced. Hyperplasia of bronchial epi-thelium infected by adenoviruses has also beenobserved in vitro (Maddi, 1963) and is of interestin relation to the oncogenic properties of type 7adenovirus (Larson, Girardi, Hilleman, and Zwickey,1965). The necrotizing lesions in bronchial glandswere a striking feature of all cases and of particulardiagnostic importance in case 3 in which they pro-vided the first evidence of an adenoviral aetiology.The uniformity of these pathological findings is

strong evidence that the type 7 adenovirus isolatedfrom each case was the primary aetiological agentand not an incidentally detected prevalent virus.Additional points confirming this relationship are:(1) that similar changes have been described inother cases from which certain adenovirus serotypeshave been isolated; (2) that the intranuclear in-clusions resemble those produced by these adeno-viruses in other cell systems; (3) that no otherinfectious agent is known to cause this combinationof lesions; and (4) that there was no evidence of aconcurrent adenovirus epidemic.

FATAL ADENOVIRUS INFECTION The first reports ofthe isolation of adenoviruses from children dyingwith pneumonia were from France, two case withsimilar pathological findings reviewed by Chany,Lepine, Lelong, Le-Tan-Vinh, Satge, and Virat(1958). By 1964 Wright, Beckwith, and Gwinnwere able to collect reports of 28 fatal cases ofadenovirus infection in children, but there wereonly three further descriptions of the pathology incases from which viruses had been isolated, one

567

on Novem


http://jcp.bmj.com

/J C


ownloaded from

http://jcp.bmj.com/

D. M. 0. Becroft

from Japan (Kawai, 1959) and two from the UnitedStates (Benyesh-Melnick and Rosenberg, 1964;Wright et al., 1964). However, Wright et al. did notinclude the reports of a large number of deaths dueto pneumonia during an epidemic of adenovirusinfection in China in 1958 and 1959 (Teng, 1960),when 3,398 children presented with pneumonia atone hospital in Peking, 528 died, and Hsiung (1963)later described the pathology of 40 of these casesin which adenoviruses were isolated from lung tissue.With the inclusion of other reports from Russia(Dreizen, Zhukova, and Kniazeva, 1960) andGermany (Guthert, Meerbach, and Wockel, 1964)there are pathological descriptions of at least 60cases from which adenoviruses have been isolated.The majority of these adenoviruses have been type7 and all but one of the remainder type 3. Mostpatients were previously healthy infants under 1year of age. Necrotizing bronchitis and bronchiolitis,and a pneumonia with mononuclear cellular reaction,hyaline membranes, and necrosis have been des-cribed in varying detail in all cases from which type7 or type 3 viruses were isolated. These changesappear identical to those found in the present series,but not all authors have described the acidophilicgranular necrosis of mucous glands and hyper-plastic epithelium which were a feature of all theAuckland cases. The presence of inclusions or theisolation of a virus from the appropriate tissuestrongly supports the adenoviral aetiology ofmeningo-encephalitis (Chany et al., 1958), hepatitis(Benyesh-Melnick and Rosenberg, 1964) and lympha-denitis (Kawai, 1959) in some cases, but theseappear to be unusual complications.

INCLUSIONS All reports of the pathology of type 7infection describe intranuclear inclusions in thelesions in the respiratory tract and most make adistinction between the smaller eosinophilic formswith surrounding halo and the much larger baso-philic inclusions which fill greatly expanded nuclei.There has, however, been a difference of opinionon the interrelationship of the two types of in-clusion, with each referred to as the late or matureform and the other as the precursor. The conclusionstated above, that these were alternative and largelyindependent reactions of the nucleus to infection,receives support from consideration of the cyto-logical changes induced in tissue cultures by type 7adenoviruses (Boyer, Denny, Miller, and Ginsberg,1960). In tissue cultures the mature inclusions arelarge, basophilic, and strongly Feulgen-positive.They contain up to 10 times as much D.N.A. asuninfected cells and large amounts of viral antigenand infective virus. In contrast to the sequence inherpes simplex infection there is no later stage in

which the inclusions become eosinophilic andFeulgen-negative. In vivo as in vitro it is likely thatthe large basophilic Feulgen-positive inclusion isthe mature form, typical of a cell heavily, and pre-sumably lethally, infected by adenovirus.There is no stage described in tissue culture

corresponding exactly to the solitary eosinophilicinclusion, but multiple small eosinophilic inclusionsappear as an early cytopathic effect when there islittle infective virus and viral antigen present. Thissuggests that the eosinophilic inclusions found in vivooccur in cells which are less heavily infected andcan be correlated with the observation in the presentseries that these inclusions were found particularlyin the proliferating bronchial epithelial cells. Manyof the nuclear changes observed in tissue cultures,including crystalline structures and 'flower' or'rosette' forms of mature inclusions, were rare inthese cases, but this difference may be due to thepoor cytological detail in necropsy material com-pared with that in monolayer preparations.

DIFFERENTIAL DIAGNOSIS There is agreement on themajor pathological features of type 7 infection andthat similar changes are induced by the type 3 virus.These lesions have considerable specificity and innecropsy material a firm diagnosis will usually bepossible on histological grounds alone. The twoserotypes are members of a subgroup of adeno-viruses sharing many biological properties but fatalinfections by other serotypes of the group have yetto be described. Cases apparently having fatal in-fections by the less closely related serotypes 1 and21 have not had extensive necrotizing lesions norsimilar inclusions (Deinhardt, May, Calhoun, andSullivan, 1958; Clarke, Corner, Gambier, Macrae,and Peacock, 1964).

Other virus infections may induce intranuclearinclusions in the respiratory tract but these will notusually cause diagnostic difficulty. Although someof the adenovirus inclusions resemble the type Ainclusions of varicella and herpes simplex the lesionsof the disseminated forms of the latter infectionsare typically focal and involve other organs, whilethe larger basophilic inclusions characteristic ofadenovirus infection will be absent. Cytomegalicinclusion disease will be readily excluded becauseof the difference in the clinical setting, in the local-ization of lesions in the respiratory tract, and amarked difference in size and form of all but thedegenerate cytomegalic inclusions. In morbillithere may be some destructive changes in the res-piratory epithelium and bronchial glands, plussmall amphophilic intranuclear inclusions, butthese inclusions will be in multinucleated epithelialgiant cells which also contain prominent cyto-

568

on Novem


http://jcp.bmj.com

/J C


ownloaded from

http://jcp.bmj.com/


plasmic inclusions. Confusion between the twodiseases is more likely to be on clinical than patho-logical grounds because of morbilliform rashescaused by adenovirus infection Chany et al., 1958).Many pathological features of types 3 and 7 in-

fection closely resemble those due to the influenzaviruses. Necrotizing lesions in the respiratoryepithelium of similar extent, depth, and maximumlocalization in smaller bronchi have been describedin influenza uncomplicated by bacterial infection(Walsh, Dietlein, Low, Burch, and Mogabgab, 1961).Necrosis of cells in bronchial glands (Spencer, 1962)and tropism to alveolar lining cells (Hers andMulder, 1961) have also been reported. Fatal in-fluenza infection is rare in infants, but histologicaldifferentiation from adenoviral infection wouldrest largely on the demonstration of typical inclus-ions in the latter.The more recently identified organisms causing

the majority of childhood respiratory infections-the rhino-viruses, the respiratory syncytial virus,the para-influenza viruses, and the Eaton agent-would not have been detected by the methods usedto isolate adenoviruses from these cases. Little isknown of the pathology of these rarely fatal in-fections, particularly that due to the rhino-virusesand Eaton agent, but it appears unlikely that theywould have caused such widespread necrotizinglesions in the bronchi, or have induced intranuclearinclusions. The respiratory syncytial virus is themajor cause of bronchiolitis in infancy, but inthe few cases reported the bronchial epitheliumhas shown syncytium formation and the inclusionswere cytoplasmic (Adams, Imagawa, and Zike,1961; Holzel, Parker, Patterson, Cartmel, White,Purdy, Thompson, and Tobin, 1965). The bronchialepithelium does not appear to have been severelyaffected in fatal cases of infection by para-influenzaviruses (Ishida, 1957; Arrobio, 1964). The differingpathological effects of adenoviruses and other res-piratory viruses parallel the changes, necrotizingor otherwise, induced in organ cultures ofrespiratory epithelium (Tyrrell and Hoom, 1965).

EPIDEMIOLOGY Several studies have demonstrateda low incidence of antibodies to type 7 adenovirusesin childhood, indicating that sporadic infection isuncommon at this age (van der Veen, 1963). Thisis not one of the serotypes commonly latent innasopharyngeal lymphoid tissue. Nevertheless, epi-demic infection of children is well documented andalthough the incidence of serious lower respiratorytract infection has been low in some epidemics(Sterner, 1962), others have provided the majorityof the fatal cases previously reported. This varia-bility in severity is unexplained, as is the maximum

occurrence of fatal cases in an age group generallylittle affected by this virus. The epidemiologicalbasis of the Auckland cases is obscure. Four pre-sented within 13 months, but at three-to-six-monthintervals and without geographical relationship.There is no indication from the records of hospitaladmissions that there was an unusually high inci-dence of any category of childhood respiratory in-fection during this 13 months. Although adeno-viruses were isolated from other patients, nonewere type 7. It is possible that the virus was wide-spread in latent form (Andrews, 1964), but thespecial susceptibility of these previously healthychildren remains unexplained.

I wish to thank Dr. Alice Bush, Dr. Elizabeth Hughes,Dr G. T. Fox, and Dr. W. R. Lang; Dr. Selwyn Hillsand Dr. J. F. Burton for their direction of studies inthe Virus Department of the Central Laboratory, Auck-land Hospital Board, and to Captain Paul D. Parkmanof the National Institute of Health, Bethesda, Md. forfurther virus studies; Mr. R. J. Patterson for technicalassistance and Mr. A. D. Fraser for assistance withphotography.

REFERENCES

Adams, J. M., Imagawa, D. T., and Zike, K. (1961). J. Lancet, 81, 502.Andrewes, C. H. (1964). Science, 146, 1274.Arrobio, J. (1964). Clin Proc. Child. Hosp. (Wash.) 20, 298.Benyesh-Melnick, M., and Rosenberg, H. S. (1964). J. Pediat., 64, 83.Boyer, G. S., Denny, F. W. Jr., Miller, I., and Ginsberg, H. S. (1960).

J. exp. Med., 112, 865.Chany, C., Lepine, P., Lelong, M., Le-Tan-Vinh, Satg6, P., and

Virat, J. (1958). Amer. J. Hyg., 76, 367.Clarke, S. K. R., Corner, B. D., Gambier, D. M., Macrae, J., and

Peacock, D. B. (1964). Brit. med. J., 1, 1536.Deinhardt, F., May, R. D., Calhoun, H. H., and Sullivan, H. E.

(1958). Arch. intern. Med., 102, 816.Dreizen, R. S., Zhukova, E. K., and Kniazeva, L. D. (1960). Vop.

Okhrany. Materin. Dets., 5, 19.Grayston, J. T., Johnston, P. B., Smith, M. E., and Loosli, C. G.

(1956). J. infect. Dis., 99, 188.Guthert, H., Meerbach, W., and Wockel, W. (1964). Z. ges. innt. Med.,

19, 865.Hers, J. F. P., and Mulder, J. (1961). Amer. Rev. resp. Dis. 83, (no 2,

pt. 2), 84.Holzel, A., Parker, L., Patterson, W. H., Cartmel, D., White, L. L. R.,

Purdy, R., Thompson, K. M., and Tobin, J. O'H. (1965).Brit. med. J., 1, 614.

Hsiung, C.-C. (1963). Chin. med. J., 82, 390.Ishida, N. (1957). Ann. N. Y. Acad. Sci., 67, 299.Kawai, K. (1959). Jap. J. exp. Med., 29, 359.Larson, V. M., Girardi, A. J., Hilleman, M. R., and Zwickey, R. E.

(1965). Proc. Soc. exp. Biol. (N. Y.) 118, 15.Maddi, F. V. (1963). Ann. Inst. Pasteur, 104, 43.Maisel, J. C., Pierce, W. E., Crawford, Y. E., and Rosenbaum M. J.

(1962). Amer. J. Hyg., 75, 56.Rosen, L. (1960). Ibid., 71, 120.Rowe, W. P., Huebner, R. J., Hartley, J. W., Ward, T. G., and

Parrott, R. H. (1955). Ibid, 61, 197.Spencer, H. (1962). Pathology of the Lung. Pergamon Press, Oxford.Sterner, G. (1962). Acta paediat. (Uppsala), suppl. 142.Stuart-Harris, C. H. (1962) Brit. med. J. 2, 869.Teng, C. H. (1960). Chin. med. J., 80, 331.Tyrrell, D. A. J., and Hoorn, B. (1965). Brit. J. exp. Path., 46, 514.van der Veen, J. (1963). Amer. Rev. resp. Dis., 88, suppl, 167.Walsh, J. J., Dietlein, L. F., Low, F. N., Burch, G. E., and Mogabgab,

W. J. (1961). Arch. intern, Med., 108, 376.Wright, H.T., Jr., Beckwith, J. B., and Gwinn, J. L. (1964). J. Pediat.,

64, 528.

569

on Novem


http://jcp.bmj.com

/J C


ownloaded from

http://jcp.bmj.com/

histopathology of fatal adenovirus infection respiratory

Documents