IDENTIFICATION BY DNA HYBRIDISATION OF ENTEROTOXIGENIC ESCHERICHIA COLI IN HOMES OF CHILDREN WITH DIARRHOEA

Saturday 14 January 1984

IDENTIFICATION BY DNA HYBRIDISATION OF ENTEROTOXIGENIC ESCHERICHIA COLI IN HOMES OF CHILDREN WITH DIARRHOEA PETER ECHEVERRIA UDOM LEKSOMBOON WANPEN CHAICUMPA

JITVIMOL SERIWATANA CHALARD TIRAPAT BERNARD ROWE

Armed Forces Research Institute of Medical Sciences, Bangkok, Thailand; Children’s Hospital, Bangkok; Faculties of Public Health and Tropical Medicine, Mahidol University, Bangkok; and Division of Enteric Pathogens, Central Public Health Laboratory, London

The DNA hybridisation technique to detect coding for Escherichia coli enterotoxin was used to identify enterotoxigenic E coli (ETEC) in homes of children with diarrhoea in Thailand. ETEC was found in 30 (14%) of 221 children with diarrhoea and in 9% (8/88) of their household contacts, 8% (8/101) of their neighbours, and 2% (32/1379) of inhabitants of 382 homes not associated with ETEC infections. ETEC was found significantly more often in water and food and on mothers’ hands in homes of children with ETEC-associated diarrhoea and of their neighbours than in homes of children without ETEC infections (8/360 vs 3/2290; p<0·001). ETEC was identified in 80% (71/89) of specimens that hybridised with the enterotoxin gene probes by testing E coli isolated from the same specimen in the Y- 1 adrenal and suckling-mouse assays. The DNA hybridisation assay to detect genes coding for E coli enterotoxin is an effective method of identifying ETEC in a large number of human and environmental specimens and will be a valuable tool to define further the epidemiology of this enteric

Summary

genes

pathogen. Introduction ENTEROTOXIGENIC Escherichia coli (ETEC) is an cause of diarrhoea worldwide.1-3 In developing countries ETEC is one of the most common enteric pathogens in young children.4,5 However, because ETEC assays are cumbersome and time consuming, it has been difficult to identify sources of this enteric pathogen. To identify ETEC in a large number of specimens, we applied molecular genetic and DNA technology. Specimens collected from homes of children with diarrhoea and from their household and neighbourhood contacts were examined for DNA homologous with genes coding for E coli enterotoxin.

important

Subjects and Methods

Subjects From Feb 15 to July 31,1982, the first 2 children under 5 years of age with diarrhoea of less than 48 h duration who were seen at each session of an outpatient clinic of a children’s hospital in Bangkok, Thailand, were enrolled in the study (221 children). Diarrhoea was defined as the passage of three or more loose stools in less than 24 h. On the day of the clinic visit two nurses visited the homes of the child and of another family with children under 5 years of age living nearby. Rectal swabs were collected from all family members and neighbours. Individuals were asked if they or their children had had diarrhoea during the preceding week. Specimens of water were collected in sterile open-mouthed glass bottles from clay jars used to store drinking and bathing water, and from the source of drinking water in both homes. Two specimens of food which had recently been eaten by the child with diarrhoea and two similar foods from the neighbour’s home were also collected. Hand imprints were collected on MacConkey agar from the mothers in the index and

neighbouring houses.

Processing of Specimens Rectal swabs.-Two were obtained from each person, inoculated into Cary-Blair medium, and processed within 4 h of collection. One swab was cultured on MacConkey agar and the resultant growth inoculated onto six pieces of nitrocellulose paper (Schleicher & Schuell, Keene, New Hampshire) layered on MacConkey agar. After incubation at 37°C overnight, the DNA of the bacterial 6 growth was fixed and examined in the DNA hybridisation assay. Ten E coli colonies were also selected from the initial MacConkey agar and stored on nutrient agar slants. Colonies from specimens which hybridised with one or more of the three DNA probes for genes coding for enterotoxins (L T, 7 SToll, 8ST -P) were tested for heat-labile (LT) and heat-stable (ST) enterotoxin production in the Y-1 adrenallO and suckling-mouse assays,l1 (the standard assays). All toxigenic isolates were serotyped12 and assayed for colonisation antigens by mannose-resistant haemagglutination of human and bovine erythrocytes and agglutination by antisera to colonisation factor antigens I and II and E coli E8775. * The second swab was cultured on thiosulphate-citrate-bile-salts (TCBS) agar (for identification of vibrios) before and after enrichment in alkaline peptone water pH 8’0. Environmental samples.-Food samples were homogenised in sterile glass homogenisers with 2 ml brain-heart infusion broth and processed in the same manner as the rectal swabs. Six 100 ml samples of each water specimen were passed through O’ 45 m nitrocellulose filters and the bacteria on the filters were processed in the DNA hybridisation assay. Two additional 100 ml samples of water were passed through 0-45 m ’Millipore’ filters

trapped

(Millipore, Bedford, Massachusetts)

and 8368

one

was

layered

on

64

MacConkey

agar and the other

on

TCBS agar. After incubation

TABLE II-DNA HYBRIDISATION ASSAY EVIDENCE OF ETEC IN

overnight at 37°C, ten E coli colonies selected from the MacConkey agar were stored on nutrient agar slants for testing in the standard

ENVIRONMENTAL SPECIMENS FROM HOMES OF INDEX CHILDREN AND

assays. Vibrios were identified from filters layered on TCBS medium.16 Bacteria isolated on MacConkey agar from imprints of the mothers’ hands were examined with the DNA hybridisation assay. Ten E coli colonies, if present, were picked from food, water, and hand cultures to be stored on nutrient agar slants. Colonies from samples which hybridised with the enterotoxin gene probes were tested for ETEC in the standard assays.

ETEC INFECTIONS

DNA

Hybridisation

Bacterial growth from all specimens was fixed on nitrocellulose paper and examined with two different preparations of each of the three a-32P-labelled enterotoxin gene probes.E coli K12 C600 (pRIT 10036) and E coli K12 C600 (pSLM 004) containing multicopy recombinant plasmids from which the 7Bf* and ST-H probes were derived, ETEC B2C (producing LT and ST enterotoxin) which hybridised with the LT and ST-H probes, and non-enterotoxigenic E coli Kl2 Xac were included as probe controls on

OF THEIR NEIGHBOURS AND FROM HOMES NOT ASSOCIATED WITH

*Both neighbouring homes’ bathing serotype as in the index child.

water

contained ETEC of the

same

TABLE III-COMPARISON OF THE DNA HYBRIDISATION AND STANDARD METHODS TO IDENTIFY ETEC

each sheet of nitrocellulose paper.

Results ETEC was isolated from 30 (14%), Shigella spp from 9%, Aeromonas hydrophila from 3%, non-O Vibrio cholerae from 2%, and Salmonella spp from 1% of 221 children with diarrhoea. The DNA hybridisation assay identified ETEC in 9% (8/88) of household contacts of the index children (32% of the inhabitants if the index cases are included) and in 8% (8/101) of their neighbours, but in only 2% (32/1379) of individuals from 382 homes not associated with children with ETEC infections. 5 of 12 other children under 5 years living in the homes of the 30 index children were infected with ETEC, compared with only 3 of 43 children of the same age in the 30 neighbouring homes (p<0’ 025; table I). 2 of these 8 infected children had diarrhoea. None of the 8 older household and neighbourhood contacts with ETEC infections had diarrhoea. In 1 home both parents and a child were infected with the same LT-producing ETEC (07:H-). LTST-producing ETEC 078:H 12 was isolated from a child with diarrhoea and a 19-year-old girl in an adjacent home who frequently looked after the child. In 5 other homes 6 family contacts of the index children were infected with ETEC of different serotypes from those of the children with diarrhoea. Evidence of ETEC was found in environmental samples in the homes of 6 of the 30 index children (table II). Evidence of ETEC was more frequently found in homes of index children than in homes not associated with ETEC infections (p<0 001). The bathing water in homes adjacent to those of two index children contained the same ETEC as was found in TABLE I-DNA HYBRIDISATION ASSAY EVIDENCE OF ETEC INFECTIONS IN INHABITANTS OF HOMES OF INDEX CHILDREN AND OF THEIR

NEIGHBOURS AND HOMES NOT ASSOCIATED WITH ETEC INFECTIONS

*25 children under 1 and 5 children 2-4 years are also inhabitants of these homes. They are excluded from the "contacts" because they were index cases seen at the outpatient clinic.

*E coli from 5 specimens that hybridised with enterotoxin gene probes not available for testing in the standard assays. fNumber of specimens from which enterotoxigenic E coli isolated. :l:No E coli isolated from 2 specimens. SNT=E coli recovered from specimen, but isolates died on storage.

were

the index children (LTST-producing ETEC 0128:H12 and LT-producing ETEC 07:H-). Drinking water and powdered milk were identified as possible sources of ETEC infections. ST-producing ETEC of serotype 078:H12 with colonisation factor antigen I was isolated from a 7-month-old girl with diarrhoea and from water in the clay jar used to store drinking water in her home. ST-producing ETEC of serotype 017:H45 was isolated from an 11-month-old boy with diarrhoea and from powdered milk in his home. LTproducing ETEC 0167 :H5 was isolated from a 2-year-old girl with diarrhoea and from her mother’s hand. 89 specimens hybridised with the enterotoxin gene probes (table in). ETEC was isolated from 80% (71/89) of specimens that hybridised with the enterotoxin gene probes. ETEC was not detected in 1 specimen that hybridised with both the LT and ST-H probes, in 14 that hybridised with the LT probe, and in 3 that hybridised with the ST-H probe, by the more time-consuming procedure of isolating, storing, and testing individual colonies in the standard assays. E coli died during storage of 1 specimen that hybridised with the ST- H probe and no E coli was recovered from 4 specimens, including 2 hand cultures, that hybridised with the LT gene probe. Thus, if only specimens from which E coli was available for testing in the standard assays are considered, ETEC was isolated from 84% (71/84) of specimens that hybridised with the enterotoxin gene probes.

65

Discussion In this study of children with ETEC-associated diarrhoea’ in Thailand, the DNA hybridisation technique identified ETEC in 9% of family contacts, and the rate of infection was greatest in siblings under 5 years of age. The infection rate in neighbours was similar. In a study in Bangladesh 11% of household contacts were infected, and both the rate of infection and proportion of infected persons with diarrhoea dropped with increasing age.17 Food and water have previously been identified as sources of ETEC in epidemics of diarrhoea.18-22 In our study drinking water and powdered milk were identified as possible sources of infections in 1 home each. In the other 28 homes studied these sources were not implicated either because the

relevant food or water was no longer available for testing, or because the infection was acquired elsewhere. The DNA hybridisation assay was reliable in detecting bacteria containing genes coding for E coli enterotoxin. ETEC were isolated by means of the Y-1 adrenall° and suckling-mousell assays from 80% of 89 specimens that hybridised with the enterotoxin gene probes or, if only the 84 specimens from which E coli was available for testing are considered, from 84%. The DNA hybridisation assay has previously been shown to be very specific in identifying ETEC in Asia.23 There are several possible reasons why the technique detected more ETEC than the standard assays in our study. First, enterotoxin production is plasmid mediated/4 ETEC may have lost plasmids coding for enterotoxin during storage. In one study many ETEC from a variety of serotypes not usually associated with ETEC gradually lost their ability to produce toxin.25 In the enterotoxin gene probe assay bacterial- DNA is fixed onto nitrocellulose within 48 h. Once fixed it is extremely stable; enterotoxin genes on nitrocellulose papers were still detectable after a year.Secondly, other bacteria may interfere with the isolation of E coli, especially from symptom-free carriers and hand imprints. Numbers of ETEC may be low in comparison with other bacteria, so they are obscured and not selected for testing in the standard assays. In contrast, DNA from non-enterotoxigenic bacteria does not interfere with the DNA hybridisation assay.26 Thirdly, the hybridisation method is at least 10 000 times more sensitive in detecting ETEC in water26 and 1000 times more sensitive in stool6 than testing ten individual E coli colonies from the same specimen in the standard assays. Fourthly, although bacteria other than E coli have been reported to produce enterotoxin,27,28 other species of bacteria were not tested in the standard assays. It is possible that bacteria of other genera carried genes coding for E coli enterotoxin. Vibrios can share some degree of homology with the LT probe.29 However, it is unlikely that similar, but not identical, DNA sequences were detected by the enterotoxin gene probes under the hybridisation conditions we used. Furthermore, no vibrios were found in the 14 LT-probepositive specimens from which no LT-producing ETEC was isolated. B There are two difficulties in defining the epidemiology of ETEC-separating E coli from the biomass present in most environmental samples and testing large numbers of these colonies for enterotoxin production rapidly. The major advantage of the DNA hybridisation assay over previous methods of identifying ETEC1O,11 is that it increases logarithmically the number of samples that can be examined. Although certain environments are clearly contaminated with ETEC, contacts of index cases in 5 homes were infected

with ETEC of different serotypes, suggesting that future studies of the epidemiology of ETEC infections should consider the transmission of genes coding for enterotoxin as well as the serotype of ETEC infecting the index case. By employing this application of molecular genetic technology to larger populations we hope to identify further sources of genes coding for enterotoxin in the community and to develop strategies for the prevention of this infectious diarrhoea. We

thank

C. T.

Pitarangsi, P. Ratarasarn, D. Plianbangchang, O. Sakuldaipeara, P. Nabumrung, S. Boonak, S. Piyaphong, V. Khungvalert, and P. Siripraivan for technical assistance; S. Tumrongrachaniti, S. Pinnoi, N. Kramomthong, and C. Pravichpram for obtaining specimens in the homes; andS. MoseleyandS. Falkow for supplying E coli C600 containing cloned genes for E coli enterotoxin and help in establishing the DNA hybridisation assay in Thailand. Supported in part by contract C6-181-81 from the World Health Organisation. Chivaratanond,

Correspondence should be addressed to P. E., Armed Forces Research Institute of Medical Sciences, Rajvithi Road, Bangkok 10400, Thailand, or APO, San Francisco, California 96346, USA: .

REFERENCES 1. Black RE, Merson MH, Rahman ASMM, et al. A two-year study of

bacterial, viral, and parasitic agents associated with diarrhea in rural Bangladesh. J Infect Dis 1980; 142: 660-64.

2. 3.

Ryder RW, Sack DA, Kapikian AZ, et al. Enterotoxigenic Escherichia coli and reoviruslike agent in rural Bangladesh. Lancet 1976; i: 659-62. Sack RB. Human diarrheal disease caused by enterotoxigenic Escherichia coli. Annu Rev

Microbiol 1975; 29: 333-53. RW, Brown KH, Becker S, et al. Longitudinal studies of infectious diseases and physical growth of children in rural Bangladesh. II. Incidence of diarrhea and association with known pathogens. Am J Epidemiol 1982; 115: 315-24. 5. Leksomboon U, Echeverria P, Suvongse C, et al. Viruses and bacteria in pediatric diarrhea in Thailand: a study of multiple antibiotic-resistant enteric pathogens. Am J Trop Med Hyg 1981; 30: 1281-90. 6. Moseley SL, Echeverria P, Seriwatana J, et al. Identification of enterotoxigenic

4. Black

Escherichia coli by colony hybridization using three enterotoxin gene probes. J Infect Dis 1982; 145: 863-69. WS, Gill D, Falkow S. Cistrons Bacteriol 1979; 139: 850-58.

7. Dallas

encoding Escherichia colt heat-labile toxin. J

8. Moseley SL, Hardy JW, Huq MI, et al. Isolation and nucleotide sequence determination of a gene encoding a heat-stable enterotoxin of Escherichia coli. Infect Immun 1983; 39: 1167-74. R, Hirth P, De Wilde M, et al. Cell-free synthesis of enterotoxin of E. coli from a cloned gene. Nature 1980; 284: 473-74. Sack DA, Sack RB. Test for enterotoxigenic Escherichia coli using Y-1 adrenal cells in miniculture. Infect Immun 1975; 11: 334-36. Dean AG, Ching Y-C, Williams RB, et al. Test for Escherichia coli enterotoxin using infant mice: application in a study of diarrhea in children in Honolulu. J Infect Dis 1972; 125: 407-11. Merson MH, Rowe B, Black RE, Huq I, Gross RJ, Eusof A. Use of antisera for identification of enterotoxigenic Escherichia coli. Lancet 1980; ii: 222-24. Evans DG, Evans DJ, Tjoa WS, et al. Hemagglutination of human group A erythrocytes by enterotoxigenic Escherichia coli isolated from adults with diarrhea. Infect Immun 1977; 19: 727-36. Evans DG, Evans DJ. New surface-associated heat-labile colonization factor antigen (CFA/II) produced by enterotoxigenic Escherichia coli of serogroups 06 and 08. Infect Immun 1978; 21: 638-47. Thomas LV, Cravioto A, Scotland SM, et al. New fimbrial antigenic type (E8775) that may represent a colonization factor in enterotoxigenic Escherichia coli in humans. Infect Immun 1982; 35: 1119-24. Blair JE, Lennette EH, Truant JP. Manual of clinical microbiology, 3rd ed. Washington DC: American Society for Microbiology, 1980. Black RE, Merson MH, Rowe B, et al. Enterotoxigenic Escherichia coli diarrhea: acquired immunity and transmission in an endemic area. Bull WHO 1981; 59: 263-68. Kudoh Y, Zen Yoji H, Matsushita S, et al. Outbreaks of acute enteritis due to heatstable enterotoxin-producing strains of Escherichia coli. Microbiol Immunol 1977; 2: 175-78. Ryder RW, Wachsmuth IK, Buxton AE, et al. Infantile diarrhea produced by heatstable enterotoxigenic Escherichia coli. N Engl J Med 1976; 295: 849-53. Taylor WR, Schell WL, Wells JG, et al. A foodborne outbreak of enterotoxigenic Escherichia coli diarrhea. N Engl J Med 1982; 306: 1093-95. Sack RB, Hirschhorn N, Brownlee I, et al. Enterotoxigenic Escherichia coli associated diarrheal disease in Apache children. N Engl J Med 1975; 292: 1041-46. Rosenberg ML, Koplan JP, Wachsmuth IK, et al. Epidemic diarrhea at Crater Lake from enterotoxigenic Escherichia coli: a large waterborne outbreak. Ann Intern Med 1977; 86: 714-18. Seriwatana J, Echeverria P, Escamilla J, et al. Identification of enterotoxigenic Escherichia coli in patients with diarrhea in Asia with three enterotoxin gene probes. Infect Immun 1983; 42: 152-55.

9. Lathe

10. 11.

12. 13.

14.

15.

16. 17.

18.

19.

20. 21. 22.

23.

66

SURGICAL APPROACHES TO CHOLANGIOCARCINOMA AT CONFLUENCE OF HEPATIC DUCTS L. H. BLUMGART

I. S. BENJAMIN

N. S. HADJIS R. BEAZLEY

Hepatobiliary Surgical Unit, Department of Surgery, Royal Postgraduate Medical School, London of the surgical management of 94 consecutive patients with carcinoma at the confluence of the hepatic ducts showed that 20% of them (or 60% of those thought, after comprehensive pre-operative investigation, to have resectable lesions) underwent tumour resection—6 local resections and 12 liver resections. The 30-day hospital mortality was 11% and mean duration of survival was 17 months (7 patients are still alive, with a mean survival of 22· 2 months). The quality of life was also much improved after resection. In patients treated with palliative surgery the hospital mortality was 33% and the mean survival was 8· 5 months. The results support the treatment of bileduct carcinoma at the hilum of the liver by resection and suggest that the use of drainage techniques, whether employed surgically or non-surgically, be reserved until the possibility of treatment by resection has been fully considered.

Summary

Analysis

Introduction CHOLANGIOCARCINOMA involving the confluence of the hepatic ducts at the hilum of the liver, whose distinctive pathological features were first described by Altemeierl and Klatskin,2 used to be regarded as a rare lesion, but it is now much more frequently diagnosed as a cause of obstructive jaundice. Although the tumour is usually small and has a low propensity for metastasis, the prognosis is generally poor and surgical management has largely been palliative. Although endoscopic retrograde and percutaneous cholangiography has permitted earlier and more accurate diagnosis of malignant obstruction of the biliary tree, improvement in long-term survival has yet to be demonstrated. Management of hilar cholangiocarcinoma has recently been surrounded by controversy. Palliative surgical drainage (by means of a tube passed across the tumour), with or without radiotherapy, has been stated by some to be the treatment of choicé-7 and attempts at surgical cure have been considered inappropriate.8 With the introduction of methods for passing tubal endoprostheses percutaneously or endoscopically non-surgical drainage of malignant biliary obstruction has been made possible, and acceptable success, complication, and hospital mortality rates have been claimed for such treatment.9-14 Intrahepatic biliary-enteric 15-17 has only anastomosis, well-established in France, recently been practised in the United Kingdom and the HW, Linggood M. The transmissible nature of enterotoxin production of human enteropathogemc strains of Escherichia coli. J Med Microbiol 1971; 4: 301-05. 25 Evans DS, Evans DG, DuPont HL, et al. Patterns of loss of enterotoxigenicity by Escherichia coli isolated from adults with diarrhea: suggestive evidence of an interrelationship with serotype. Infect Immun 1977; 17: 105-11. 26. Echeverria P, Seriwatana J, Chityothin O, et al. Detection of enterotoxigenic Escherichia coli in water by filter hybridization using three enterotoxin gene probes. J Clin Microbiol 1982; 16: 1086-90. 27. Guerrant RL, Moore RA, Kirschenfeld PM, et al. Role of toxigenic and invasive bacteria in acute diarrhea of childhood. N Engl J Med 1975; 293: 567-73. 28. Jiwa SFH, Krovacek K, Wadstrom T. Enterotoxigenic bacteria in food and water from an Ethiopian community. Appl Environ Microbiol 1981; 41: 1010-19. 29. Kaper JB, Moseley SL, Falkow S. Molecular characterization of environmental and non-toxigenic strains of Vibrio cholerae. Infect Immun 1981; 32: 661-67. 24. Smith

United States of America.18,19"It offers a means of palliative decompression not complicated by the presence of tubes, which may become dislodged or blocked or require replacement. Whether radical surgical resection improves survival rate and quality of life remains to be determined, though some encouraging results have been reported.20,21 The results of liver transplantation for hilar cholangiocarcinoma have been disappointing, with 7 of 8 transplant patients having tumour recurrence, often shortly after surgery, and only 2 patients have survived for just over a

year.22,23 In recent years all patients seen at the Royal Postgraduate Medical School, Hammersmith Hospital, with obstructive jaundice due to hilar cholangiocarcinoma have been considered for surgical resection or surgical bypass. Diagnostic and management policies have been directed at: defining the extent of disease and ascertaining non-

resectability by comprehensive pre-operative investigation, including cholangiography and selective angiography;24 obtaining histological or cytological confirmation of all lesions; and assessing prospectively the scope of radical surgery in the treatment of hilar cholangiocarcinoma, with an evaluation of its hospital mortality rate and its effects on longterm survival and quality of life. Patients and Methods

Subjects This report is based on 94 patients (55 men, 39 women, aged 29-75 years, average 56 - 5) with hilar cholangiocarcinoma defined anatomically as lesions arising from the right or left or both hepatic ducts or the common hepatic duct and affecting the confluence of the hepatic ducts. The site of the lesion was shown on percutaneous

transhepatic cholangiography or endoscopic retrograde cholangiography, and it was confirmed at laparotomy, on examination of the resected specimen, or at necropsy in all but 4 cases. The patients had been admitted to two institutions for treatment of obstructive jaundice. The first 15 patients were treated by L. H. B. at Glasgow Royal Infirmary, while the remaining 79 were referred for treatment from May, 1979, to June, 1983, to the hepatobiliary surgical unit in Hammersmith Hospital, London. 24 other patients referred for suspected cholangiocarcinoma were excluded: 4 who harboured peripheral cholangiocarcinoma not involving the bifurcation; 16 with gallbladder carcinoma; and 4 with biliary obstruction and cholangiographic evidence highly suggestive of malignant stricture at the confluence of the bile ducts, but whose strictures were found postoperatively to be benign . Methods

During initial assessment of the patients’ general condition, special attention was paid to nutritional status and evidence of impaired hepatic function. Renal function was monitored and a thorough search was made for sepsis. All patients underwent comprehensive pre-operative investigations to establish the extent of disease. All patients

were examined by grey-scale ultrasonography. If cholangiography had been done previously, the films were reviewed and if the study was incomplete percutaneous transhepatic cholangiography (PTC) was done with a fine needle. Frequently, separate punctures were necessary to fill segmentally obstructed ducts and display the entire biliary tree. The extent of ductal disease as shown cholangiographically was assessed and particular attention was paid to filling of the left ductal system to allow planning of biliary-enteric anastomosis.19 If cholangiographic

studies did not show that the lesions were too extensive for resection, and if the general condition of the patient did not preclude surgery, angiography was done. Selective coeliac and superior mesenteric arteriograms were obtained and visualisation of the portal venous system at the hilum by indirect portography was