Rotavirus, adenovirus, andnon-viral enteropathogens in diarrhoea · Rotavirus, adenovirus, andnon-viral enteropathogens in diarrhoea TVESIKARI, MMAKI,HKSARKKINEN,PPARSTILA,ANDPEHALONEN

Archives of Disease in Childhood, 1981, 56, 264-270

Rotavirus, adenovirus, and non-viral enteropathogensin diarrhoeaT VESIKARI, M MAKI, H K SARKKINEN, P P ARSTILA, AND P E HALONEN

Department of Paediatrics, Tampere Central Hospital, and Department of Virology, University of Turku,Finland

SUMMARY The aetiology of rotavirus and adenovirus in acute gastroenteritis was studied in aprospective series that comprised 283 children admitted consecutively with diarrhoea during a

1-year period. Rotavirus was associated in 49% of the cases by solid-phase radioimmunoassay andelectron microscopical examination of stool specimens, or by serology. Adenovirus was detected byradioimmunoassay in the stool specimens of 29 (11 %) patients, including 8 cases of possible dualinfection with rotavirus. Rotavirus infections showed a typical age distribution and seasonalclustering, between January and June, whereas the adenovirus-associated cases did not form adistinctive subgroup. Enteropathogenic bacteria were found in 10% of cases, and were nearly ascommon in association with rotavirus infection as not. Respiratory symptoms accompanied diar-rhoea in 34% of the patients with rotavirus and in 25% of those with neither rotavirus nor adeno-virus. Therefore we could not confirm the existence of a 'rotavirus syndrome', nor could we confirman association of respiratory symptoms with rotavirus infection. Use of antibiotics before the onsetof diarrhoea was more common among those with non-viral diarrhoea (23 %) than in the rotavirusgroup (13 %). Rotavirus infections appeared to be common among cases of 'antibiotic-induced'diarrhoea.

Since its discovery in 197312 rotavirus has been welldocumented as a causative agent of infantile gastro-enteritis.4 Occasionally rotaviruses are encounteredwith other enteropathogens, but their significance insuch dual infections is unknown.5 Also uncertain isthe role of rotaviruses in the aetiology of respiratorysymptoms often associated with rotavirusdiarrhoea.68

Interest in adenoviruses as significant entericpathogens in children has recently been revived.9Electron microscopy (EM) has contributed a greatdeal to this increased interest, since adenovirusesare found in the stools more often by EM thanby isolation'0 1' and there now appear to be severaltypes of new 'noncultivable' adenoviruses.12 Again,the association of the virus in the faeces with thedisease is questionable. Adenoviruses may be foundin the stools in the absence of diarrhoea, with orwithout respiratory symptoms.13 14 Adenoviruseshave also been found in diarrhoeal stools in com-bination with other enteropathogens.5

In the present study we have used the newlydeveloped solid-phase radioimmunoassay (RIA)techniques which are sensitive and specific for the

264

detection of rotaviruses'5 and adenoviruses28 instools. We have studied the presence of rotavirusand adenovirus by these methods, and by viralserology, during a 1-year period in children admittedto hospital with diarrhoea. Additionally, bacterialenteropathogens were searched for from the samepatients. We have specifically monitored respiratorysymptoms during the diarrhoea and we recorded theuse of antibiotics so as to correlate these factors withthe aetiology of diarrhoea.

Materials and methods

Patients. The study group comprised 283 childrenadmitted consecutively for diarrhoea to this hospitalduring a period of one year (1 December 1977 to30 November 1978). The patients were chosen ifthey had diarrhoea on admission, irrespective ofother concomitant diseases. Children who had beentreated with antibiotics before the onset of diarrhoeawere included too. Those who developed diarrhoeawhile in hospital were termed nosocomial infectionsand were excluded from the study. If a child wasadmitted for a second time with diarrhoea he was

on Novem

ber 29, 2020 by guest. Protected by copyright.

http://adc.bmj.com

/A

rch Dis C

hild: first published as 10.1136/adc.56.4.264 on 1 April 1981. D

ownloaded from

Rotavirus, adenovirus, and non-viral enteropathogens in diarrhoea 265

regarded as a new entry; there were 9 such cases.Three children found to have non-infectious causesof diarrhoea (cows' milk intolerance, bleedingMeckel's diverticulum, and Crohn's disease) wereexcluded. So finally 280 children were studied.The parents were asked about the duration of

symptoms before the admission, and about the use ofantibiotics. Any sign of respiratory infection wasregistered by nurses in charge of the patient. Thediagnosis of otitis media in any patient in hospitalwas confirmed by middle-ear aspiration. Afterdischarge, the patients were seen again by onephysician (M M) at about 4 weeks, and the parentswere asked about any illnesses during the interimperiod.

Specimen collection and storage. Between 1 and 3stool specimens were collected from each patientevery 2 days if possible. No stools for viral studieswere obtained from 7 children, but the patients wereincluded in the series for the sake of completeness.Convalescent stage stool specimens were collectedfrom 193 children 4 weeks after discharge from thehospital. All the specimens for viral studies were keptat -200C and tested retrospectively under code forrotavirus and adenovirus antigens by RIA. Any stoolthat was positive in RIA, or by serology using acomplement-fixation (CF) test (see below), wasfurther studied by EM for the presence of rotavirusand adenovirus particles. Stool specimens forbacteriology were taken to the laboratory in transportmedium and were generally cultured the same day.Serum samples were obtained at the acute stage on

the first weekday in hospital and again at theconvalescent stage 4 weeks later. Convalescent serawere not obtained from 22 children, but the childrenare included in the total number of patients.

Solid-phase radioimmunoassay. The solid-phase RIAfor rotavirus has been described in detail.15 In brief,rotavirus antigen in the test samples was captured onto polystyrene beads precoated with guinea-pigantirotavirus antibody. The presence of antigen wasassayed by adding rabbit antirotavirus antibody and125I-labelled sheep antirabbit immunoglobulin. Incases with borderline positive results confirmationwas obtained by adding guinea-pig antirotavirusserum before the rabbit antirotavirus antibody andobserving the blocking effect. The RIA procedure foradenovirus antigen consisted of analogous steps andreagents.28 The sensitivities of the RIA methods weredetermined to be between 1 and 10 ng/ml of purifiedvirus protein.

Electron microscopical examination. About 10%stool suspensions in phosphate buffered saline

centrifuged at 2500 rev/min for 30 minutes werestudied. If a RIA-positive specimen was foundnegative by the first EM, the supernatant wascentrifuged at 30 000 rev/min for 90 minutes, and thepellet resuspended in a small volume of distilledwater. The negative staining on carbon-coated gridswas done with 2% phosphotungstic acid. The speci-mens were examined using a JEM 100 electronmicroscope at a magnification of 30 000 x. At least 5grid squares of faecal material of each specimen werestudied.

Viral serology. A standard CF test with two full unitsof complement was used for all serology. Forrotavirus serology the antigen preparation was madeby growing Nebraska Calf Diarrhea Virus (NCDV)in continuous LLC-MK2 cells in the presence oftrypsin.16 At 48 hours the cells and the medium werefrozen and thawed three times, centrifuged at 2500rev/min for 30 minutes, and the supernatant usedas the CF antigen. The adenovirus CF antigen wasbought from Orion Diagnostica, Espoo, Finland. Inaddition, the following antigens were used in thestandard serology performed on all the paired sera:influenza A and B, parainfluenza 1, 2, and 3,respiratory syncytial virus, herpes simplex, varicellazoster virus, cytomegalovirus, mumps, measles,rubella, coxsackie B5, and Mycoplasma pneumoniae.

Bacteriological studies. The following enteropatho-genic bacteria were searched for: enteropathogenicEscherichia coli (EPEC), enterotoxigenic (LT) E. coli(ETEC), enteroinvasive E. coli (EIEC),'7 Salmonellasp., Shigella sp., Yersinia enterocolitica, andCampylobacter jejuni. In addition, cysts of Giardialamblia were studied from stool specimens using aformalin-ether concentration method.

Results

The diagnostic criteria for rotavirus-associated andadenovirus-associated diarrhoea were established asfollows. RIA of the stools was the primary test forboth. For rotavirus, the demonstration of sero-conversion or a significant rise in CF antibodies toNCDV was also considered diagnostic of diarrhoea,since it was thought unlikely that such a serologicalfinding was associated with any disease other thangastroenteritis. In contrast, for adenovirus only thosepositive in RIA were regarded as possible adenovirus-associated cases of diarrhoea, since a diagnostic risein adenovirus antibodies alone may be found in anumber of other clinical conditions as well. In fact, inthe present series there were 18 children withdiarrhoea and no demonstrable adenovirus in thestools who showed a significant rise in adenovirusCF antibodies over a period of 4 weeks.

on Novem

ber 29, 2020 by guest. Protected by copyright.

http://adc.bmj.com

/A

rch Dis C

hild: first published as 10.1136/adc.56.4.264 on 1 April 1981. D

ownloaded from

266 Vesikari, Mdki, Sarkkinen, Arstila, and Halonen

From these criteria there were 128 cases ofrotavirus-associated diarrhoea and 29 cases ofadenovirus-associated diarrhoea. Of the latter, in 8children evidence of a concomitant rotavirusinfection was also found or, conversely, there were 8cases of rotavirus-associated cases with simul-taneous presence of adenovirus in the stools. Finally,there were 131 cases in which neither rotavirus noradenovirus could be shown. In 21 instances either thestool specimen or the convalescent serum wasmissing, so the possibility of rotavirus infectioncannot be excluded in these patients. This left 110cases of non-rotavirus-associated, non-adenovirus-associated diarrhoea with adequate specimens. Thetotal number of children in whom viral studies werecomplete was 259, and a diagnosis of rotavirus couldbe made in 49% and adenovirus in 11% of themincluding dual infections; neither virus was found in42% of cases.

Rotavirus-associated diarrhoea. Epidemiologicallythe rotavirus-associated cases formed a fairlydistinct subgroup (Figs 1 and 2). The epidemicseason lasted from January to June with a few casesbetween July and October. There was only one casein an infant aged less than 3 months, and theinfection was rare up to 6 months. The peak incidencewas between 7 and 12 months, and rotavirusdiarrhoea remained common up to age 3 years.Most children with rotavirus diarrhoea showed

seroconversion by CF test with NCDV (Table 1). Theseroconversion rate in the RIA-positive cases was85%. In only 8 of the 104 RIA-positive cases wasthere no detectable CF antibody, and in 3 others theCF antibody titre did not change. Therefore most ofthe rotavirus-associated cases appeared to beprimary infections. The CF antibody titres in theconvalescent sera were generally low (Fig. 3), butthere was not much difficulty in serological diagnosisbecause in most cases the acute stage specimens werenegative for NCDV antibody. As expected, EM wasfound to be less sensitive than RIA for detecting

A]iT

IL

I. RotavirusAdenovirusNon-rota,non -adenovirus

12 1 2 3 4 5 6 7 8Month of year

rotavirus in stool specimens (Table 2). Of the 108RIA-positive specimens rotavirus was found by EMin 88 (81 %) cases.There were no second admissions for rotavirus-

associated diarrhoea during the 1-year period. At theconvalescent stage none of the initially rotavirus-positive patients had demonstrable rotavirus in thestools (95 tested). Rotavirus was detected at theconvalescent stage in one child who initially had hada bacterial diarrhoea (EPEC 0 11 1): in this case there

40 -

30 -LA

a

0

, 20

Ez

10

F-* X: E-1.1 fi rl -

E Rotavirus_ Adenovirus

r INon-rota.non - adenovirus

0-3 4-6 7-12 13-18 19-24 25-36 37-72 6-10 11-15Months Years

Fig. 2 Age distribution of 259 children who hadrotavirus, adenovirus, or non-rotavirus, non-adenovirus-associated diarrhoea.

Table 1 Comparison of radioimmunoassay for rotavirusantigen in the stools with serological studies ofpairedsera by complement fixation using NCDV antigen in122 children

Serology RIA-positive RIA-negative

Seroconversion 88 15Significant rise 5 3No significant change 3 0No detectable CF antibody 8 0

Fig. 1 Seasonal distribution of259children admitted with gastroenteritis inwhom a viral diagnosis could be made.

9 10 11

30-

,,, 20

0

0" 100

z

,.,:E

D;...,...,.,'.'

on Novem

ber 29, 2020 by guest. Protected by copyright.

http://adc.bmj.com

/A

rch Dis C

hild: first published as 10.1136/adc.56.4.264 on 1 April 1981. D

ownloaded from

Rotavirus, adenovirus, and non-viral enteropathogens in diarrhoea 267

had been a second attack of diarrhoea at home a fewdays before specimen collection.

Adenovirus-associated diarrhoea. There were 29adenovirus-positive cases by RIA including caseswith evidence of a concomitant rotavirus infection byRIA, and 4 cases with a simultaneous rise of rota-virus antibodies. One other case was associated withM. pneumoniae infection. Each of the remaining 20patients is listed in Table 3.Of the 20 adenovirus-only cases the presence of

adenovirus in the stools was confirmed by EM in 12instances. The results of the serological tests varied.In 7 patients there was no demonstrable adenovirusCF antibody response. Seroconversion by CF wasseen in 8 children, a significant rise of CF titre in 2more, and in 3 cases there was a high CF titre withoutrise.There was some clustering of the adenovirus-

associated cases of diarrhoea in the spring (Fig. 1).Eight out of the 20 cases (Table 3) occurred in April

30

(Aw 20-

0

0zE:: 10

......

......

...I...

.......

...I...

<4 4CF titre

128 >2568 16 32 64

Fig. 3 Distribution of rotavirus complement fixation(NCD V) antibody titres in convalescent sera of 103children in whom rotavirus antigen was demonstrated inthe stools at the acute stage of diarrhoea.

Table 2 Comparison of electron microscopicalexamination with radioimmunoassay in 125 cases ofrotavirus-associated diarrhoea

RIA-positive RIA-negativediagnosis by serology

EM positive 88 0EM negative 20 17

Table 3 Adenovirus-associated cases of diarrhoea withno evidence of concomitant rotavirus or bacterialpathogenAge Adenovirus in stools Serology by CF

(months) EM RIA Acute Convalescent

4 + + <4 <44 - + 8 45 + + <4 <45 + + <4 46 - + <4 327 - + <4 87 + + <4 168 - + <4 416 - + <4 1616 + + <4 <418 + + <4 <419 - + <4 12824 + + 8 6424 + + <4 <426 - + 512 25628 + + 32 25630 + + <4 <435 + + 64 6451 - + 1024 102467 + + <4 128

and May. Similarly, of the 18 serological diagnoseswith negative RIA 10 were made in these months.The age distribution of the adenovirus-associated

cases of diarrhoea was wide (Table 3). Eight patientswere infants aged between 4 and 8 months, only fourwere between 1 and 2 years of age, and eight wereolder. There was some correlation between the age ofthe patient and the serological response, as most ofthose who failed to produce a good CF antibodyresponse by the convalescent stage were younginfants.No adenovirus was detected by RIA in stool

specimens at the convalescent stage of 19 patientswho had been initially positive. However, adenoviruswas detected on two occasions in the convalescentspecimens from patients who had been negative foradenovirus at the acute stage of diarrhoea. These2 children did not have diarrhoea at the time ofspecimen collection.

Association of viruses and other enteropathogens.Bacterial pathogens were encountered only rarely(Table 4). Enteropathogenic bacteria were found in29 cases representing 10% of the total material. Inaddition G. lamblia was demonstrated in 4 cases, butit was not associated with rotavirus or adenovirusinfection. Similarly, neither rotavirus nor adenoviruswas demonstrated in the 4 children with Salmonellasp. Diarrhoeagenic E. coli (EPEC and ETEC) wereisolated both in association with rotavirus andwithout. Isolations of C. jejuni and Y. enterocoliticawere almost equally divided between viral andapparently non-viral cases of diarrhoea.

Serological evidence (seroconversion or a signifi-cant rise in the paired sera studied by CF) of viral

-... .... ..... -,.- . - .--l

on Novem

ber 29, 2020 by guest. Protected by copyright.

http://adc.bmj.com

/A

rch Dis C

hild: first published as 10.1136/adc.56.4.264 on 1 April 1981. D

ownloaded from

268 Vesikari, Mcki, Sarkkinen, Arstila, and Halonen

Table 4 Association of rotavirus and adenovirus with non-viral enteropathogens in 280 cases of diarrhoea in childrenEscherichia coli Salmonella sp. Yersinia Campylobacter Giardia

enterocolitica jejuni lambliaEPEC ETEC EIEC

Rotavirus (n = 128*) 4 2 0 0 1 3 0Adenovirus (n = 29*) 0 0 0 0 0 (l)t 0Non-rota, non-adenovirus (n = 131) 7 2 1 4 1 4 4Total 11 4 1 4 2 7 4

*Includes 8 cases of dual infection with rotavirus and adenovirus. tOne case of Campylobacterjejuni isolation was associated with such adual infection.

Table 5 Serological evidence of associated virus infections in rotavirus diarrhoea and in cases ofdiarrhoea withoutdemonstrable rotavirus or adenovirus in the stools. Paired sera were collected at the acute stage of the disease and4 weeks later

Adenovirus Parainfluenza Respiratory Mycoplasma Herpes Coxsackie B5 All- syncytial virus simplex viruses

2 3

Rotavirus (n = 128) 9 2 4 0 0 0 0 15Non-rota, non-adenovirus (n = 131) 9 1 3 1 2 1 1 18Total 18 3 7 1 2 1 1 33

Table 6 Antibiotic treatment before the onset of diarrhoea, and occurrence of respiratory symptoms in relation toviral aetiology ofdiarrhoea in 259 childrenAetiology of diarrhoea Antibiotics Respiratory symptoms Otitis media

No (7%) No (%) No (%)

Rotavirus (n = 128*) 17 (13) 43 (34) 19 (15)Adenovirus (n = 29*) 7 (24) 9 (31) 9 (31)Non-rota, non-adenovirus (n = 110) 25 (23) 28 (25) 24 (22)*Includes 8 cases of dual infection of rotavirus and adenovirus.

infection was obtained in 33 cases (Table 5). Thisfigure does not include rotavirus serology or adeno-virus serology in cases with demonstrable adenovirusin the stools. As the specimens were collected at4-weekly intervals it was not possible to ascertainwhether the serological diagnoses resulted from aconcomitant infection or from a second illness in themeantime. Adenovirus infections occurred mostfrequently (18 cases); in addition, parainfluenzavirus, respiratory syncytial virus, herpes simplexvirus, and M. pneumoniae were diagnosed. Theseserological diagnoses of respiratory viruses wereevenly distributed between rotavirus-associated andnon-rotavirus, non-adenovirus-associated cases ofdiarrhoea. There was only one case of enterovirusinfection demonstrable by a significant rise incoxsackie B5 virus antibodies.

Association of diarrhoea with other clinical conditions.Clinical signs of respiratory tract infection werepresent at admission or during the duration inhospital in almost one-third of all patients withdiarrhoea (Table 6). Otitis media was observed quiteoften, in about 20% of the cases. There was no cleardifference between the viral and the assumed non-viral cases of diarrhoea in this respect. A smallproportion of the respiratory symptoms could be

explained by another respiratory virus infectiondetectable using CF antibody response. There were 5serological diagnoses among the rotavirus group ofdiarrhoea with respiratory infection, 1 in the cor-responding adenovirus group, and 11 in the non-rotavirus, non-adenovirus group.

Altogether 54 children were receiving or hadreceived antibiotics within 3 days before the onset ofdiarrhoea. This 'antibiotic-induced' diarrhoeaoccurred somewhat more often among the non-rotavirus, non-adenovirus-associated cases (23 %)than in the viral diarrhoea group (13% for rotavirus,24% for adenovirus). Conversely it can also be saidthat many allegedly antibiotic-induced cases ofdiarrhoea were in fact associated with, and probablycaused by, rotavirus or adenovirus.

Discussion

Since the discovery of rotavirus there have beenseveral studies on infantile gastroenteritis looking forboth viral and bacterial enteropathogens.47 10 18-23It was hoped that this investigation would be acomprehensive study of diarrhoea of children inhospital. We believe that the diagnostic recovery wasthorough as we used sensitive methods for rotavirusand adenovirus in the stools, the viral studies were

on Novem

ber 29, 2020 by guest. Protected by copyright.

http://adc.bmj.com

/A

rch Dis C

hild: first published as 10.1136/adc.56.4.264 on 1 April 1981. D

ownloaded from

Rotavirus, adenovirus, and non-viral enteropathogens in diarrhoea 269

complemented by serology of paired sera, and mostrecognised non-viral enteropathogens were searchedfor by adequate methods.

Rotavirus diarrhoea appears epidemiologically toform a distinct entity, and this was observed in thepresent study. Rotavirus was present in 49% of thecases, which is one of the highest figures to bereported in a 1-year study of children. It is unlikelythat many cases were missed as we used bothserology and faecal examination by RIA. RIA hadpreviously been found to be more sensitive that EMfor the detection of rotavirus in the stools.28 Therewas a very good correlation between RIA forrotavirus in the stools and serology by CF of pairedsera, confirming previous observations on thecorrelation between EM of stools and serology byCF using NCDV as antigen.24 Serological studiescomplemented the examination of stools by RIA orEM and showed additional cases of rotavirusinfection. The relatively long period (about 4 weeks)between the acute and convalescent stage specimensapparently allowed these children to developdetectable antibody responses.18The finding of peak incidence of rotavirus diar-

rhoea between 7 and 12 months of age is in accord-ance with previous studies.6 7 19 24 The epidemicseason lasted until June, which is later than in reportsfrom countries with more temperate climates-suchas USA.14The replication of rotavirus is probably limited to

the small-intestine,25 26 and there are no reports ofrotavirus outside the gastrointestinal tract.8 In viewof this it is surprising that respiratory symptoms areso often observed in rotavirus diarrhoea,6 7 as wasthe case in the present series. Lewis et al.7 went so faras to call cases of rotavirus diarrhoea 'rotavirussyndrome' if they were preceded by respiratorysymptoms. We could not find such a clear associationbetween rotavirus and respiratory infection. Respir-atory infection and otitis were indeed common beforeand during diarrhoea, but they occurred equallyoften in adenovirus-associated cases and in patientsin whom neither rotavirus nor adenovirus wasdemonstrated. Use of antibiotics before the onset ofdiarrhoea was more common in patients with nodemonstrable virus, but it also occurred not in-frequently in rotavirus-associated diarrhoea. All thisprobably reflects the fact that in winter, and in younginfants of less than 2 years, both respiratory infectionsand diarrhoea, by rotavirus and other causes, occurfrequently and the chance of coincidence is great.

In contrast to rotavirus, the adenovirus-associatedcases of diarrhoea do not appear as a distinct sub-group in the series. There was a wide distribution ofage and only some seasonal clustering, many casesappeared to be associated with a concomitant

rotavirus infection, there were cases with adenovirusin the stools but no serological response and,conversely, many cases of serological diagnosiswithout adenovirus in the stools. Despite all thesearguments it appears that some cases of diarrhoea inthe present series were likely to have been caused byadenovirus. Therefore, continuous search for adeno-viruses seems worthwhile in children with diarrhoea,whenever a suitable test is available. The present RIAmethod is convenient and sensitive, and the use ofthis or comparable enzymatic assays should in thefuture lead to increased knowledge of the role ofadenoviruses in diarrhoea.The association of stool viruses and entero-

pathogenic bacteria appears a complicated issue.Looking at the present results one might firstconclude that there is no apparent division betweenbacterial and viral cases of diarrhoea, but entericbacteria are found with or without rotavirus in arandom fashion. Rotavirus gastroenteritis mayprecede infection by enteropathogenic bacteria byfavouring their growth. On the other hand thisconventional thinking that viral infection leads tobacterial superinfection may not necessarily alwaysbe valid. In the present series in at least two instancesinfection by enteropathogenic bacteria seemed to pre-cede rotavirus infection, and in several other casesrotavirus and bacteria were present simultaneously.Careful longitudinal follow-up studies are needed.

Finally, one should comment on the 92 (35%)cases of diarrhoea with adequate specimens for viraland bacterial studies but with no demonstrableaetiological agent. These probably form a hetero-geneous group. Stool viruses other than rotavirus andadenovirus were not fully covered in this study, butthey were not encountered by EM in specimens thathad been found positive for rotavirus or adenovirusby RIA. Among these unresolved cases there mayalso be bacterial diarrhoeas which so far are elusiveand this group includes 'true' cases of antibiotic-associated diarrhoea, whatever the pathogenicmechanism is. Another explanation for the aetiologyin some cases may be 'parenteral' diarrhoea associ-ated with other infections-such as otitis media,pneumonia, or urinary tract infection by a hithertounknown mechanism.27

We thank Ms Taimi Tammipuu, Ms Kerttu Kangas,Ms Tuire Laatta, Ms Leena Soini, and Ms MaritaMaaronen for technical assistance, and Ms TerttuRosenholm for secretarial help.

This work was supported by grants from theAcademy of Finland, the Medical Research Council,and the Sigrid Juselius Foundation.

on Novem

ber 29, 2020 by guest. Protected by copyright.

http://adc.bmj.com

/A

rch Dis C

hild: first published as 10.1136/adc.56.4.264 on 1 April 1981. D

ownloaded from

270 Vesikari, Maki, Sarkkinen, Arstila, and Halonen

References

Bishop R F, Davidson G P, Holmes I H, Ruck B J. Virusparticles in epithelial cells of duodenal mucosa fromchildren with acute non-bacterial gastroenteritis. Lancet1973; ii: 1281-3.

2 Flewett T H, Bryden A S, Davies H. Letter; Virusparticles in gastroenteritis. Lancet 1973; ii: 1497.

3Anonymous. Rotavirus gastroenteritis. Br Med J 1977;ii: 784-5.

4Walker-Smith J. Rotavirus gastroenteritis. Arch Dis Child1978; 53: 355-62.

5Madeley C R, Cosgrove B P, Bell E J, Fallon R J. Stoolviruses in babies in Glasgow. I. Hospital admissions withdiarrhoea. J Hyg (Camb) 1977; 78: 261-73.

6 Rodriguez W J, Kim H W, Arrobio J 0, et al. Clinicalfeatures of acute gastroenteritis associated with humanreovirus-like agent in infants and young children. JPediatr1977; 91: 188-93.

7Lewis H M, Parry J V, Davies H A, et al. A year'sexperience of the rotavirus syndrome and its associationwith respiratory illness. Arch Dis Child 1979; 54: 339-46.

8 Goldwater P N, Chrystie I L, Banatvala J E. Rotavirusesand the respiratory tract. Br MedJ 1979; ii: 1551.

9 Retter M, Middleton P J, Tam J S, Petric M. Entericadenoviruses: detection, replication, and significance.J Clin Microbiol 1979; 10: 574-8.

10 Bryden A S, Davies H A, Hadley R E, Flewett T H,Morrison C A, Oliver P. Rotavirus enteritis in the westMidlands during 1974. Lancet 1975; ii: 241-3.

1 Birch C J, Lewis F A, Kennett M L, Homola M, PritchardH, Gust I D. A study of the prevalence of rotavirusinfection in children with gastroenteritis admitted to aninfectious diseases hospital. J Med Virol 1977; 1: 69-77.

12 Gary G W, Jr, Hierholzer J C, Black R E. Characteristicsof noncultivable adenoviruses associated with diarrheain infants: a new subgroup of human adenoviruses.J Clin Microbiol 1979; 10: 96-103.

13 Appleton H, Buckley M, Robertson M H, Thom B T.A search for faecal viruses in new-born and other infants.J Hyg (Camb) 1978; 81: 279-83.

14 Brandt C D, Kim H W, Yolken R H, et al. Comparativeepidemiology of two rotavirus serotypes and other viralagents associated with pediatric gastroenteritis. Am JEpidemiol 1979; 110: 243-54.

15 Sarkkinen H K, Halonen P E, Arstila P P. Comparison offour-layer radioimmunoassay and electron microscopyfor detection of human rotavirus. J Med Virol 1979; 4:255-60.

16 Sarkkinen H K, Meurman 0 H, Halonen P E. Solid phaseradioimmunoassay of IgA, IgG, and IgM antibodies tohuman rotavirus. J Med Virol 1979; 3: 281-9.

17 Maki M, Vesikari T, Gronroos P. Enterotoxigenic andinvasive Escherichia coli as causes of childhood diarrhoeain Finland. Acta Paediatr Scand 1980; 69: 219-24.

18 Kapikian A Z, Kim H W, Wyatt R G, et at. Humanreovirus-like agent as the major pathogen associated withwinter gastroenteritis in hospitalized infants and youngchildren. NEnglJ Med 1976; 294: 965-72.

19 Carr M R, Donald G, McKendrick W, Spyridakis T.The clinical features of infantile gastroenteritis due torotavirus. ScandJ Infect Dis 1976; 8: 241-3.

20 Gurwith M J, Williams T W. Gastroenteritis in children:a two-year review in Manitoba. 1. Etiology. J Infect Dis1977; 136: 239-47.

21 Echeverria P, Ho M T, Blacklow N R, et al. Relativeimportance of viruses and bacteria in the etiology ofpediatric diarrhea in Taiwan. J Infect Dis 1977; 136:383-90.

22 Pickering L K, Evans D J, Jr, Munoz 0, et al. Prospectivestudy of enteropathogens in children with diarrhea inHouston and Mexico. JPediatr 1978; 93: 383-8.

23 Hieber J P, Shelton S, Nelson J D, Leon J, Mohs E.Comparison of human rotavirus disease in tropical andtemperate settings. Am J Dis Child 1978; 132: 853-8.

24 Tufveson B, Johnsson T. Occurrence of reo-like virusesin young children with acute gastroenteritis. Diagnosesestablished by electron microscopy and complementfixation, using the reo-like calf virus as antigen. ActaPathol Microbiol Scand [B] 1976; 84: 22-8.

25 Theil K W, Bohl E H, Cross R F, Kohler E M, Agnes A G.Pathogenesis of porcine rotaviral infection in experi-mentally inoculated gnotobiotic pigs. Am J Vet Res 1978;39: 213-20.

26 Middleton P J. Pathogenesis of rotaviral infection. J AmVet Med Assoc 1978; 173: 544-6.

27 Nelson J D. Diarrhea. In: Vaughan V C, McKay R J, Jr,Behrman R E, eds. Nelson textbook ofpediatrics, eleventhedition. Philadelphia: Saunders, 1979: 710-2.

28 Halonen P, Sarkkinen H, Arstila P, Hjertsson E, TorfasonE. A four-layer radioimmunoassay for detection ofadenovirus in stool. J Clin Microbiol 1980; 11: 614-7.

Correspondence to Dr Timo Vesikari, Department ofPaediatrics, Tampere Central Hospital, 33520Tampere 52, Finland.

Received 22 January 1980

on Novem

ber 29, 2020 by guest. Protected by copyright.

http://adc.bmj.com

/A

rch Dis C

hild: first published as 10.1136/adc.56.4.264 on 1 April 1981. D

ownloaded from