LATE REFERRAL FOR BILIARY ATRESIA—
MISSED OPPORTUNITIES FOR EFFECTIVE
E. R. HOWARD
ALEX P. MOWAT
Departments of Child Health and Surgery and Liver Unit, King’s
College School of Medicine and Dentistry of King’s College London,
Denmark Hill, London
To assess whether clinicopathological
features other than the age at operation
influence prognosis after surgery for extrahepatic biliary
atresia (EHBA) and to determine whether the age at referral
has fallen since a previous survey, 50 consecutive cases with
EHBA referred between February, 1985, and December,
1987, were reviewed. Liver or spleen size, liver function
tests, or histological appearance of liver biopsy specimen
before surgery were not predictive of outcome. The jaundice
cleared up in 12 of 14 children operated on by age 8 weeks,
but in only 13 of 36 operated on later. In 41 referral was
delayed. All 25 children in whom surgery was successful are
alive and well, while 13 of 25 with unsuccessful surgery
have died, at a median age of 1 year. To improve the
prognosis of infants with EHBA parents and health staff
need a better awareness of the early clinical features of
EHBA and of the necessity for prompt referral. Liver
disease should be suspected in any infant jaundiced after 14
days of age.
EXTRAHEPATC biliary atresia (EHBA), the commonest
life-threatening hepatobiliary disorder in childhood, affects
at least 1 in 20 000 liveborn infants.’ Bile drainage can be
improvement in prognosis; Ohi et al reported a 10-year
survival in 90% of children in whom serum bilirubin
concentration returned to normal. Such surgery must be
done before all intrahepatic bileducts leading to the porta
hepatis are destroyed, usually by 8 weeks of age, otherwise
the prognosis is very poor, the majority of the children dying
by 2 years of age, and survival beyond age 4 years being
exceptional.’’ Failed surgical correction of EHBA is the
most frequent indication for liver transplantation during the
first decade of life.5.6
TABLE I-PRE-OPERATIVE CLINICAL, LABORATORY, AND
Results are expressed as median and range.
Serum bilirubin: normal < 20 lrnol/1
AST = aspartate aminotransfcrase: normal < 50 lUll.
GGT = gammaglutamyl transpeptidase : normal < 50 IU/1.
Albumin: normal >
Our survey published 4 years ago, on the incidence and
short-term prognosis of EHBA in the United Kingdom,!
showed that, like previous studies 3,7 restoration of bile
drainage was more likely with early (by 8 weeks of age) than
with late surgery, especially in experienced referral centres.
Unfortunately, 75% of the children identified had been
treated after 60 days of age, when short-term success rate
was 33%, compared with 72% in younger infants.
The aim of the present study was to investigate whether
age at referral has diminished since the publication of the
survey and to identify reasons for continued delay in
referral. We also compared the influence of age at surgery on
prognosis with that of clinical, laboratory, and pathological
PATIENTS AND METHODS
The subjects were 50 consecutive British infants with EHBA
referred to the Paediatric Liver Service at King’s College Hospital
between February, 1985, and December, 1987. 4 had multiple
anomalies (1 child had situs inversus, 1 polysplenia and intestinal
malrotation, 1 atrial septal defect, situs inversus, and preduodenal
portal vein, and 1 polysplenia and situs inversus with several
intra-abdominal vascular abnormalities). 27 were females. The 50
patients represent about one-third of all infants witn conjugated
hyperbilirubinaemia referred during the same period.
Liver and spleen sizes were recorded in centimetres below the
costal margin. Liver function was assessed by use of the’Technicon
SMAC II’ (Technicon, Basingstoke). Percutaneous liver biopsy
specimens obtained before surgery
quantitatively for fibrosis (0 = none, 1 = portal, 2 = portoseptal,
3 = severe portoseptal + bridging), lobular hepatitis (0 = none,
1 - mild, 2 = patchy collapse, 3 = bridging), bileduct damage
(0 = none, 1 = present), and cholestasis (1 = perivenular, 2 = severe
perivenular + periportal, 3 = perivenular + periportal + prominent
cholangiolar casts). Biopsy specimens were scored by a
histopathologist (B.P.) unaware of the surgical outcome.
The contents of referral letters and the parents’ statements on
admission as recorded in the infants’ case-records were
retrospectively analysed for reasons for delayed referral. "1..2 and
Wilcoxon rank tests were used for statistical analysis. Results are
expressed as median and range.
were assessed semi-
Outcorne of Surgery
48 infants underwent portoenterostomy and 2 hepatico-
jejunostomy, all operations being done by the same surgeon
(E.R.H.). 25 (50%) became jaundice-free (serum bilirubin
< 20 mol/1 = successful surgery) after surgery. No
significant difference was found in liver and spleen size or in
serum bilirubin, aspartate
gammaglutamyl transpeptidase (GGT) and albumin levels
between children who became jaundice-free after surgery
and those who did not (table I). Prothrombin time was
normal in all children immediately before surgery and at last
follow-up. Similarly, the semiquantitative pre-operative
liver biopsy scores were not predictive of outcome. Median
age at operation was 10 weeks (4-25 weeks). Children in
whom surgery was successful were younger (9 weeks, 4-25
weeks) at time of operation than those who did not become
jaundice-free (10 weeks,
Furthermore, surgery was successful in 12 of 14 (86%)
infants who had surgery before age 8 weeks, but in only 13 of
36 (36%) of those who were aged more than 8 weeks at time
of surgery (p < 0-0005). 11 of 27 (41 %) children operated on
between 8 and 12 weeks of age, but only 2 of 9 (22%)
operated on after age 12 weeks (p < 0-0005), became
jaundice-free (table II).
TABLE II-OUTCOME ACCORDING TO THE AGE AT SURGERY
All 25 children in whom surgery was successful are alive.
24 of them remain jaundice-free at a median age of 28
months (range 10-43 months), whereas 1 child became
jaundiced again 13 months after successful surgery and is
alive and well
at the age of 21 months, with
hepatosplenomegaly (liver 5 cm, spleen 6 cm), abnormal
serum bilirubin (199 amol/1), aspartate aminotransferase
(AST) (269 IU/1), and gammaglutamyl transpeptidase
(GGT) (890 IU/1), and normal albumin (36 g/1).
Of the 25 children in whom surgery was unsuccessful, 13
died at a median age of 12 months (range 5-36 months);
these 13 include 3 with multiple anomalies and 2 who
underwent liver transplantation 7 and 33 months after
surgery. 2 other children, who received transplants at 6 and
16 months of age, are alive and well 10 and 6 months later.
The remaining 10 children were alive, growing and
developing normally, and engaged in usual activities for
their age at the latest review (table ill). 9 of them are
regularly attending our unit (median follow up 30 months,
postoperatively; at the time, she was well, with serum
bilirubin 71 Imol/1, AST 96 IU/1, GGT 824 IU/1, liver 2
cm, spleen 2-5 cm, and albumin normal. The AST and
GGT levels in these 10 children are significantly higher than
those in the 25 in whom surgery was successful (table ill).
1 child went to America 5 months
Delay in Referral
The 50 infants were referred at a median age of 8 weeks
(range 2-5-23 weeks). Overall 72% of the children were
referred after 6 weeks of age, 81 % in 1985,54% in 1986, and
73 % in 1987. Referral after 8 weeks of age was 40% overall,
46% in 1985, 31 % in 1986, and 36% in 1987.
There were several reasons for late referral (table IV). In
11 cases, before referral to a paediatrician, worried parents
had been reassured, on up to three documented occasions,
by general practitioners, health visitors, midwives, or
community medical staff, that jaundice persisting for more
than 2 weeks after birth was physiological. In 5 others
persistent jaundice was judged to be of no importance, the
prime reason for referral to a paediatric unit being failure to
In 30 cases delayed referral to a specialist centre was
associated with inappropriate management by hospital
paediatric services. No investigations were done in 7 infants
TABLE III-CLINICAL AND LABORATORY DATA AT LAST FOLLOW UP
OF THE 35 SURVIVING CHILDREN*
Results are expressed as median and range.
*Excluding 2 patients who underwent liver transplant.
tl child had jaundice again at the age of 15 months.
TABLE IV-FACTORS CONTRIBUTING TO DELAYED REFERRAL
with jaundice lasting more than 2 weeks or in 4 with jaundice
and vitamin K responsive haemorrhagic diathesis. In
another 5 cases breast milk jaundice was diagnosed despite
clinical evidence of liver disease. In 7 cases a fall in serum
bilirubin concentration of up to 40 nmol/1 between 2 and 8
weeks of age was thought to exclude biliary atresia. Delays in
referral and appropriate investigations in 2 cases were due to
the presence of pigmented stools for a few weeks after birth.
In 2 infants with situs inversus this abnormality was not
appreciated and a left-sided liver was mistaken for an
enlarged spleen, so investigations were directed at causes of
splenomegaly. In 2 cases the diagnosis was suspected and
laparotomies were done, but corrective surgery was not
attempted in 1 case, and the bile drainage procedure done in
the other was inadequate.
investigated at age of 6 weeks in another hospital, but since
the liver biopsy was not characteristic of EHBA, the
diagnosis was no longer considered and appropriate follow-
up and investigations were delayed.
In 1 case late referral was due to parental reluctance to
accept the diagnosis of serious liver disease in a child who
looked so well.
In only 5 occasions was referral from primary care to
district paediatric services and subsequent transfer arranged
1 child was thoroughly
This study confirms that surgery before age 8 weeks is
important to obtain effective bile drainage in a high
percentage (86%) of patients with EHBA. Clinical features,
standard biochemical tests of liver function, and severity of
histological changes on percutaneous liver biopsy are not of
prognostic value. It also shows that in the United Kingdom
corrective surgery is still done too late. Several factors may
account for this late referral. Possibly it is still believed that
EHBA is a congenital malformation and therefore affected
children are expected to be jaundiced at birth, to have
persistently acholic stools, and to have continuously
increasing levels of serum bilirubin. Since EHBA is not a
malformation but an obliterative process often present at
birth, but sometimes starting later, and involving previously
patent extrahepatic bileducts,9 the presence of pigmented
stools postnatally or fluctuating levels of serum bilirubin do
not exclude this condition. Furthermore, practitioners
familiar with the cachexia and ascites of the advanced
cirrhosis in late EHBA may be unware that in the early stage
of the disease, when surgery is most effective, the majority of
patients are well nourished. Indeed, in only 5 of the 50
children reported here was failure-to-thrive a presenting
symptom. The severity of clinical and biochemical features
423 Download full-text
of liver disease in the early stages of EHBA is variable. The
liver and spleen are generally only slightly enlarged (the
median liver size being only 1 cm above normal for infancy)
and the jaundice mild (the mean bilirubin being 165 umol/1
The ominous signs in a jaundiced infant-urine that is not
colourless and stools with no yellow or green pigment-are
easily spotted but are rarely volunteered by parents,
especially those of first-born babies. Health education
publications do not mention these important signs.’"" Since
most parents will have been advised that jaundice in the first
week of life is "normal", it is understandable that they may
remain unconcerned or be easily reassured that jaundice
persisting after 1 week of age is not pathological. In fact
jaundice after 11 days in term infants and 14 days in preterm
infants is very unusual, occurring in only 0-5 % of infants in a
paediatric unit.12 Some of these children will have benign
jaundice associated with breast feeding, but a pathological
cause should always be excluded.
It is essential that conjugated hyperbilirubinaemia is
recognised promptly, since it is always pathological and
reflects an underlying hepatobiliary disease. Nearly all
parents informed us, on direct questioning, that their
infant’s urine had never been colourless. In over 50% of the
cases referring medical and paramedical staff had not asked
about this sign or, if they had observed it, they had not
taken appropriate action to investigate the underlying
hepatobiliary problem. Almost universally parents were
unaware that the yellow colour of their infant’s urine after
the first week of life was abnormal. When infants underwent
phototherapy in the first week of life, parents had not been
advised that the jaundice should have cleared by 3 weeks of
life and that the urine should have become colourless at the
end of treatment. In 6 cases the reassuring diagnosis of
breast milk jaundice was given without ascertaining whether
or not the bilirubin was conjugated. Failure to recognise
hepatobiliary disease by hospital staff is surprising. No
adequate investigations, for example, were made in babies
with prolonged neonatal jaundice and dark urine or with a
haemorrhagic diathesis beyond the first few days of life that
responded rapidly to parenteral vitamin K and that should
suggest serious hepatobiliary dysfunction.13 Even when
liver disease was suspected, the child was often inadequately
investigated for several weeks before referral to a specialist
centre. Early identification of the cause of the underlying
liver dysfunction is always important for adequate
management and often for genetic counselling. In the case of
EHBA it is life saving.
We are concerned that the interval between referral and
surgery is 2 weeks. Although there are now techniques that
give results quickly14 about 1 week is needed to exclude the
many infective, genetic, metabolic, endocrine, and biliary
disorders which may mimic EHBA. These must be
carefully excluded, to avoid unnecessary laparotomy in
children with intrahepatic disease. Cystic fibrosis and
3-anti-trypsin deficiency, which can mimic EHBA, must
be excluded. The deficiency of cx1-anti-trypsin must be
excluded by phenotyping, since serum levels are often
misleading in the presence of liver disease in infancy. After
all results are available, a few more days still elapse before
surgery. Biliary atresia presents in our unit in an episodic
fashion. Surgery must be fitted in around a full operating
schedule and to minimise delay it is often done as a
The only alternative to portoenterostomy for children
with EHBA is liver transplantation. At present the 90%
10-year survival following successful portoenterostomy3
compares favourably with the 62% 5-year survival after one
or more liver transplantation procedures.15 Furthermore,
liver transplantation is more costly, both acutely and in the
long term, and may have more morbidity.16 Even if donors
were available it should be reserved for children in whom
portoenterostomy has been unsuccessful. Because of the
shortage of donors for small recipients, liver transplantation
was available to only 4 of 12 infants without major
contraindications to transplantation in this series. Those
who have survived so far despite unsuccessful surgery have
obtained sufficient palliation to be well grown at a mean age
of 30 months and thus to have much better prospects of
obtaining a size-matched orthotopic liver transplant, or
orthotopic liver transplant following hepatic reduction.17
While advocating portoenterostomy as the primary surgical
management for EHBA, we avoid repeated laparotomy or
cutaneous biliary drainage which can complicate any later
Early identification and referral of babies with conjugated
hyperbilirubinaemia is likely to be essential if porto-
enterostomy is to be offered at the optimum time. This has
been successfully achieved by two measures in Japan.l8
Firstly, parents receive written advice on the serious
implications of yellow stained urine and pale stools
associated with jaundice in early infancy. This information
is reinforced with posters in infant welfare clinics and in
family practitioners’ surgeries. Secondly, the "well-baby"
review is done at 4 weeks of age rather than at 6 weeks as in
the United Kingdom. We advocate that similar measures be
instituted in the United Kingdom.
G. M.-V. is supported by the Michael McCough Foundation against Liver
Disease in Children.
Correspondence should be addressed to A. P. M., Department of Child
Health, King’s College Hospital, Denmark Hill, London SE5 8RX.
1. McClement JW, Howard ER, Mowat AP. Results of surgical treatment for
extrahepatic biliary atresia in United Kingdom 1980-2. Survey conducted on
behalf of the British Paediatric Association Gastroenterology Group and the British
Association of Paediatric Surgeons Br Med J 1985; 290: 345-47.
2. Kasai M, Suzuki S. A new operation for "non correctable" biliary atresia; hepatic
portoenterostomy. Shujitsu 1959, 13: 733-39
3. Ohi R, Hanamatsu M, Mochizuki I, Chiba T, Kasai M. Progress in the treatment of
biliary atresia. World J Surg 1985; 9: 285-93.
4. Mowat AP, Psacharopoulos HI, Williams R Extrahepatic biliary atresia versus
neonatal hepatitis A series of 137 prospectively investigated infants. Arch Dis Child
1976, 51: 763-70.
5 Pett S, Pelham A, Tizard J, et al Pediatric liver transplantation. Cambridge King’s
series, December 1983 to August 1986. Transpl Proc 1987; 19: 3256-60.
6 Zitelli BJ, Gartner JC, Malatack JJ, et al. Pediatric liver transplantation patient
evaluation and selection, infectious complications, and life style after
transplantation. Transpl Proc 1987; 19: 3309-16.
7 Kasai M Advances in treatment of biliary atresia. Jpn J Surg 1983; 13: 265-76.
8. Hayes DM, Altman P, Hitch DC, Lilly JR, Smith EI, Urceda JE. Biliary atresia in the
United States: the survey of the surgical section of the American Academy of
pediatrics. In: Kasai M, ed. Biliary atresia and its related disorders. Amsterdam;
Excerpta Medica, 1983 161-16.
9 Mowat AP. Liver disorders in childhood London: Butterworth, 1987 72-86.
Butterworth, 1987 72-86
10 Kohner N Pregnancy book. London Health Education Council 1984.
11 Scowen P, ed New baby A health visitors’ and midwives’ handbook for parents.
London Newbome Holdings, 1987.
12. Matthew PM, Wharton BA. Investigation and management of neonatal Jaundice, a
problem-oritentated case record. Arch Dis Child 1981; 56: 949-53
13 Lane PA, Hathaway MD Vitamin K in infancy J Pediatr 1985; 106: 351-59.
14. El Tumi MA, Clarke MB, Barrett JJ, Mowat AP. Ten minute radiopharmaceutical
test in biliary atresia Arch Dis Child 1987, 62: 180-81
15 Iwatsuki S, Starzl TE, Todo S, et al Experience in 1000 liver transplants under
cyclosporin-steroid therapy: a survival report. Transpl Proc 1988; 20 (suppl 1).
16 Stewart BA, Hall RJ, Lilly JR Liver transplantation and the Kasai operation in biliary
atresia. J Pediatr Surg 1988, 23: 623-66.
17 de Hemptinne B, Salizzoni M,Tan KC, Otte JB The technique of liver size reduction
in orthotopic liver transplantation Transpl Proc 1988; 20 (suppl 1) 508-11
18. Hayes DM, Kimura K. Biliary atresia The Japanese experience. Cambridge,
Massachussets: Harvard University Press, 1980 24