Biliary Atresia Registry, By Frederick
M. Karrer, John R. Lilly, Barbara A. Stewart, Denver,
Biliary atresia;
Kasai procedure;
portoen-
I
N 1976, THE SURGICAL Section of the American Academy of Pediatrics established a nationwide registry of children with extrahepatic biliary atresia. Reports of the early experience with Kasai’s new operation, portoenterostomy, showed widely disparate results.“2 The long-term success of the procedure
From the Department of Surgery, University of Colorado School of Medicine, Denver, CO. Supported in part by the Surgical Section of the American Academy of Pediatrics, Elk Grove Village, IL; by Grant No. RR-00069 from the General Clinical Research Centers Program, Division of Research Resources, National Institutes of Health, Bethesda, MD; by the Pediatric Liver Center, University of Cole rado Health Sciences Center, Denver, CO: and by the Johnson & Johnson Company. Presented at the 38th Annual Meeting of the Surgical Section of the American Academy of Pediatrics, Chicago, Illinois, October 21-23.1989. Address reprint requests to Frederick M. Karrer, MD. Pediatric Surgery (C314). University of Colorado School of Medicine, 4200 E Ninth Ave. Denver, CO 80262. o I990 by W.B. Saunders Company. 0022-3468/90/2510-0014$03.00/0
1076
and Roberta J. Hall
Colorado
8 The Registry provides information about 904 children with biliary atresia from more than 100 institutions. There was a 1.4 to 1 female predominance: racial distribution was 62% Caucasian, 20% black, 11% Hispanic, 4.2% asian, and 1.5% American Indian. Eight hundred sixteen (90%) underwent corrective surgery (median age at operation, 69 days). Intraoperatively, 70% had totally obliterated extrahepatic bile ducts, 22% had patency of the gallbladder and distal common duct, whereas only 8% had “correctable” biliary atresia (proximal duct patency). A variety of reconstructions were used, but the majority of patients had a Roux-en-Y portoenterostomy with or without exteriorization. Follow-up was available for 670 children (74%) with average length of follow-up of 5 years (range, 1 to 16 years). Five-year actuarial survival was 48% following Kasai’s operation, but was less than 10% (at 3 years) if no operative correction was done. Survival was unaffected by sex, type of reconstruction, or cholangitis. Predictors of a bad outcome were (1) Caucasian race: (2) operative age greater than 60 days; (3) presence of cirrhosis at initial biopsy; (4) totally nonpatent extrahepatic ducts; (5) absent ducts at the level of transection in the liver hilus; and (6) subsequent development of varices or ascites. Identification of factors predictive of the ultimate outcome provide a basis for either continued efforts with management of Kasai’s operation or for early referral for liver transplantation. Q 1980 by W.B. Saunders Company. INDEX WORDS: terostomy.
1976 to 1989
Journ
remained unproven and its widespread application was questioned.3 Because each surgeon treated only relatively few infants with biliary atresia, a nationwide survey was organized to accumulate larger numbers. By combining data from multiple centers, more valid statistical analysis of the early success and long-term prognosis are available.
MATERIALS
AND
METHODS
Members of the Surgical Section of the American Academy of Pediatrics were provided with initial patient entry forms to obtain general demographic information (sex, race, birthplace, date of birth), operative data (age at operation, cholangiographic findings, hepatic and portal histology, type of reconstruction), and early results (bile drainage, operative complications, death). The status of the patients was periodically updated to determine late complications (cholangitis, varices, rickets, ascites, growth retardation), the extent of liver disease and length of survival until death or last follow-up date. More recently, surveys have partially addressed the issue of transplantation. Nine hundred eighty-five patients were entered into the Registry from 136 surgeons representing a cross-section of practices; university hospitals, general hospitals, and private practice. The patients represented a broad geographic area from across the United States and Canada. We also received patient data from Mexico and Hong Kong. Of the 985 patients, 904 had true extrahepatic biliary atresia, 45 had biliary hypoplasia, and 36 had choledochal cyst. The latter two groups were deleted from the statistical analysis. Of those with true extrahepatic biliary atresia, there were 363 boys and 541 girls (M:F ratio, 0.7:1), with a racial distribution of 62% Caucasian, 20% black, 11% Hispanic, 4.2% asian, and 1.5% American Indian. Follow-up information was obtained on 670 children (75%). The data were entered into a computerized data base and periodically updated and revised. Statistical analysis was by x’ and analysis of variance using Statpac (Walonick Assoc, Minneapolis, MN). The survival data were analyzed as of October 1, 1989, by the method of Kaplan-Meier, and the statistical comparisons made by the Cox-Mantel test using Solo statistical system (BMDP Statistical Software, Los Angeles, CA).
RESULTS
Eight hundred sixteen of the 904 infants with biliary atresia underwent corrective surgery, the remainder either had no operation or exploratory laparotomy only. At the time of the operation, 17% were less than 30 days old, 10% were over 120 days old (median age at operation, 69 days). Cholangiograms were attempted on 880 infants. Seventy percent of patients had totally obliterated nonpatent extrahepatic bile ducts, and 22% had patency of the gallbladder and distal common duct (Fig 1). Only 8% were reported to have proximal patency, so-called “correctable” biliary of Pediatric
Sorgern Vol 25, No 10 (October), 1990: pp 1076-1081
BILIARY ATRESIA REGISTRY
Fig1. Cholangiographic findings at exploration. Patency of the gallbladder and distal common bile duct (lower right). totally obliterated extrahepatic ducts (left). and proximal duct patency, the so-called correctable type (upper right).
atresia. At the time of surgery, 25% had minimal fibrosis on liver biopsy, 43% had moderate fibrosis, and 26% had severe fibrosis/cirrhosis. Only 5% had no evidence of fibrosis at the time of initial operation. Sections of the bile ducts at the liver hilus showed microscopic ducts in over 60% and moderate inflammation in 44%. Eighty-two infants (9%) had other associated anomalies, most commonly cardiac (eg, septal defects), absent inferior vena cava, preduodenal portal vein, polysplenia, and malrotation. The type of reconstruction varied widely (Fig 2). Roux-en-Y portoenterostomy with double-barrelled exteriorization was chosen most often (38.3%) and a simple Roux-en-Y without exteriorization (Kasai I) was used in 23%. Other reconstructions were less popular--double-Y (Kasai II, 4.5%), simple intestinal conduit (Sawaguchi, 5.3%)-or had selected indications-choledochoenterostomy, 5.5%, used for correctable types and portocholecystostomy (gallbladder Kasai, 7.7%) where the distal bile ducts were patent. The early complication rate was low, with eight intraabdominal abscesses, 14 instances of small bowel obstruction, seven episodes of postoperative hemorrhage, and a total of four early deaths (0.4% operative mortality). The majority of patients achieved bile drainage after the Kasai operation, but 140 children never drained or had only temporary bile drainage. Bile drainage was verified by radionucleide scans (rose bengal or iminodiacetic acid), a decrease in serum bilirubin by more than half, bile bilirubin concentration greater than serum bilirubin, bile pigment in the stool, or other means in those who drained bile. Cholangitis was reported in 5 1% of infants during the
1077
first year postoperatively and in 44% the second year. Usually this consisted of one to three episodes per year. One quarter of patients had one or more revisions of the portoenterostomy. Of 4 15 reconstructions in which exteriorization was done, the stoma remains in 94 patients. When internalized, 25% were closed before 6 months, 39% between 6 and 12 months, 30% between 1 and 3 years, and 6% after 3 years. The late complication rate was high. Of long-term survivors (>3 years), varices were present in 49%, ascites in 63%, rickets in 32%. and developmental retardation in 65%. Subjects in the Registry were born between 1967 and 1989 (mean, 1977). Follow-up was available in 670 with average length of follow-up of 5 years (range, 1 to 15). The actuarial survival curves are shown in Fig 3. Long-term survival rate was for the entire group approximately 25%. For the subset of patients having no corrective operation (or exploratory laparotomy only), survival was limited to 3 years. Of those who died, the primary cause of death was liver failure in 64%, infections in 16.4%, gastrointestinal bleeding in 9%, and other causes in 10.9%. Prognostic Factors Liver biopsy.
The amount of fibrosis or cirrhosis on liver biopsy was found to be most strongly correlated with the patient age at the time of the operation. As the operative age increased, there was a greater amount of fibrosis or cirrhosis on liver biopsy (P < .OOl). There was no relationship between liver biopsy findings and
KASAI
23%
I
f..-B & T
DOUBLE ENTEROSTOMY 38.3%
II 4.5%
KASAI
F
. 6 SAW;GFHZ
NO OPERATION OR EXPLORATION ONLY '.,i, 1, f- F
11.9%
CHOLEOOCHOENTERO:TOMY FOR “COR;E;TABLE TYPE .%
GALLBLADDER KASAI 7.7%
Fig 2. Types of reconstruction and percentages of each. Not shown are 35 (3.8%) coded as other types of reconstruction.
1078
KARRER ET AL
80 4 f
60
6 K
40
3 -dQ
20 0
0
1
2
3
4
5
6
7
8
g 10 11 12 13 14 15 16
YEARS Fig 3. Actuarial survival curves for those undergoing operative correction of biliary atresia (solid line) and for those not having operative intervention or exploratory laparotomy only (shaded line). The groups are statistically different (P = 6001). Note that the extended survival in the “No Kasai” group represents one patient reported to be alive 11 years after exploratory laparotomy in infancy.
patient sex, race, cholangiographic findings, or bile duct histology at the liver hilus. Cholangiographicfindings. The older the infant at the time of surgery, the more often nonpatent ducts were found (P = .018). Distal patency and proximal correctable-type lesions were found more often in younger infants. Boys were significantly more likely to have distal patency of the gallbladder and common bile duct (boys 28%, girls 18%; P = .OOl). Cholangitis. The frequency of cholangitis during the first year and subsequently was significantly influenced by the type of reconstruction. Those having a “gallbladder Kasai,” portocholecystostomy, had the lowest incidence of cholangitis; 35% versus roughly 55% for the other reconstructions (P = .02). The presence of distal bile duct patency, permitting portocholecystostomy, also correlated with a lower incidence of cholangitis. The degree of hepatic fibrosis also influenced the rate of cholangitis; increasing fibrosis correlated with more cholangitis. Race had an effect as well; the incidence of cholangitis was 37% for blacks, 54% for Caucasians, 54% for asians, and 6 1% for Hispanics (P = .05). There was no correlation with patient sex, age at operation, or survival. Hepatic decompensation. The development of varices and ascites was found to correlate strongly with the amount of fibrosis on operative liver biopsy and with the cholangiographic findings. Those with more hepatic fibrosis/cirrhosis, and those with nonpatent extrahepatic ducts had significantly higher incidence of both varices and ascites. Cholangitis was also associated with the development of varices and ascites. Mortality. The lo-year actuarial survival for children having a Kasai procedure was over 30%, com-
pared with only 10% at age 3 years for those not undergoing any drainage procedure (Fig 3). Many factors influenced patient survival. The most unexpected finding was a significantly decreased survival among Caucasians versus other races (Table 1). The finding could not be explained by difference between the racial groups in terms of cholangiographic findings, liver biopsy, bile duct histology, type of reconstruction used, or age at operation. There was a lower incidence of cholangitis among blacks but cholangitis did not correlate with mortality (vide supra). Follow-up was greater for Caucasians, asians, and Hispanics (81%, 87%, 92%, respectively) than for blacks (65%). The age of the child at the time of operation had an important impact on bile drainage and on long-term survival. Infants under 30 days of age had the best chance of achieving bile drainage and for survival (Table 2). The beneficial impact of younger age seemed to plateau over 100 days of age. The operative age was also related to the bile duct size; with increasing age, total fibrosis of the bile duct remnant was more common (P = .05). The findings on exploration were also predictive of ultimate outcome. Cholangiographic findings strongly influenced survival; proximal duct patency (correctable) had the best prognosis, followed by patent distal ducts, and totally obliterated extrahepatic ducts had the worst prognosis (5-year survival rates, 72%, 62%, and 38%, respectively). Bile duct size similarly predicted outcome with mortality rates for those with a totally fibrotic porta of 77%, versus only 45% if sections at the hilum showed a grossly patent, visible duct. Inflammation of the extrahepatic ducts also had no impact on long-term outcome. On liver biopsy, minimal fibrosis at the initial operation was found to be a favorable finding, with increasing fibrosis yielding lower survival and, paradoxically, normal or near normal hepatic architecture on biopsy falling in between. The presence of associated malformations did not influence longevity (30.1% with, 35.6% without). The type of reconstruction had no effect on outcome when separated from the cholangiographic findings. That is, although the gallbladder Kasai and choledo-
Table 1. Racial Influence on Survival
Race Caucasian
Total Patients
Survival
355
32.1%
Black
89
42.7%
Hispanic
71
62.0%
Asian
26
50.0%
NOTE: Only those patients undergoing some type of reconstruction and having follow-up information were included. P -z ,001.
BILIARY ATRESIA REGISTRY
1079
Table 2. influence of Operetive Age
at Operation WYS)
Age on Bile Drainage end Survival
BileDrainage (%)
SurvivalRate (%I
530
66.7
31-60
69
6 l-90
62.3
39.5
91-120
53.3
28.6
>120
62.5 *
35 P = .004
43.6
28.8 P = .023
choenterostomy groups had improved survival, they were only applicable to certain types of biliary anatomy, distal patency and proximal patency, respectively. The other reconstructive procedures, when applied to infants with totally obliterated extrahepatic ducts, had equal survival rates and the frequency of achieving bile drainage with each of the reconstructions was not significantly different. However, the vented procedures (Roux-en-Y with exteriorization, Kasai II, and Sawaguchi) were associated with a higher frequency of cholangitis. Patients with no, or only temporary bile drainage had a poor prognosis with 3-year survival rate of only 15%. Postoperative development of varices, ascites, rickets, or developmental retardation also reduced chances of long-term survival. DISCUSSION
The original intent of the Biliary Atresia Registry was to determine the feasibility of the Kasai procedure. Following Kasai’s report in English in 1968,4 the portoenterostomy operation began to be used in centers in the United States. Published articles reported widely disparate, even contradictory results, with Campbell et al’ reporting no bile drainage in a series of 12 patients, and Lilly and Altman’ finding more than 50% successful bile drainage. This, and the absence of long-term studies, led to the establishment of a nationwide registry. By accumulation of a large number of cases from multiple centers, it was hoped that more valid conclusions about the long-term survival would result. In that regard, the Registry data quite clearly shows a 30% IO-year survival rate with Kasai’s operation, and a less than 10% 3-year survival rate without treatment. The reported poorer operative results in the United States (versus Japan),’ have been blamed on racial influence. The Registry data show that Caucasians have significantly reduced survival compared with nonwhite races (Table 1). The racial mix of patients in the Registry was dissimilar from the racial distribution in the general population of the United States, ie, fewer Caucasians (68 v 80%) and more nonwhite races. Whether this is merely a sampling error or represents a true difference in racial incidence of biliary atresia is
undetermined.6 The survival differences could not be explained by differences of age at operation, operative findings (cholangiographic, bile duct size, hepatic and porta histology), or type of reconstruction, but they were less striking when patients reported from Mexico and the Orient were eliminated. We believe the racial differences to be genuine but unexplained. There was no difference in survival between the various types of reconstruction with the exception of choledochoenterostomy for the correctable type, and a slight but not statistically significant improvement with the gallbladder Kasai for patients with distal duct patency. The implication from these data is that the most important technical factor of any reconstruction is the anastomosis at the porta hepatis, not the various loops, anastomoses, and conduits. If bile drainage is achieved, regardless of the reconstruction, survival is markedly improved; 8.5% without and 56.3% with drainage. The other point of interest is that the vented procedures had higher reported incidences of cholangitis, the complication they were designed to reduce. Possible explanations include (1) increased reporting of cholangitis in the vented group because of improved accuracy of diagosis (bile drainage visible); and (2) exteriorization, which may allow more colonization. However, cholangitis does not correlate with reduced survival (even after those with bile drainage failure are selected out [vi& infru] ). Patient age at operation was strongly correlated with bile drainage and survival (Table 2) but also with the degree of hepatic fibrosis, the type of biliary anatomy on cholangiography, and the histological findings at the porta hepatis. Increasing age was accompanied by increasing amounts of fibrosis on liver biopsy and the older patients tended to have a higher incidence of a totally obliterated tree on attempted cholangiography. There also was a trend toward smaller or absent biliary ducts in the porta hepatis in older children, supporting the concept of biliary atresia being a dynamic process with progression of disease.’ Although hepatic pathological changes varied in progression and severity, only a few patients older than 3 months were cured by corrective surgery. Therefore, prompt diagnosis with early reconstruction optimized success. Separate analysis of bile duct histology showed a strong association with survival, larger true ducts being associated with the best results and the smallest ductules and glands having reduced bile drainage and longevity. However, care must be exercised in analysis of this type of data because the histological criteria between different pathologists and surgeons may vary. Independent observations by a single pathologist or
1080
KARRER ET AL
group would be preferrable. Nevertheless, the correlation has been corroborated by a number of investigators who reported a correlation between bile duct size and survival.*-‘* Significant negative factors noted in this study included the lack of correlation between survival and (1) cholangitis; (2) associated malformations; and (3) amount of portal hilar inflammation. Cholangitis was reported to be a negative prognostic sign,13 which we were unable to confirm. The aberration could be due to variation in criteria used for diagnosis between centers, but the correlation was so weak (P = 571) that it is doubtful small adjustments would result in significance. The more likely explanation is that cholangitis only occurs in patients with bile drainage and, therefore, improved survival, whereas those with less cholangitis have been mongrelized with patients with no cholangitis due to absent postoperative bile drainage. Associated malformations, especially those connected with the polysplenia syndrome (absent inferior vena cava, situs inversus, malrotation, preduodenal portal vein, bilobed midline liver, polysplenia) are well-recognized,*49*S and occurred in 10% to 15% of patients. There is an unsubstantiated belief that this group had an unfavorable prognosis, but data failed to support that idea. In many instances, examination of the resected tissue at the porta hepatis showed a marked inflammatory infiltrate. Hitch et al* found no correlation between the degree of portal hilar inflammation and survival. The Biliary Atresia Registry analysis confirms that finding. The degree of inflammation also bore no relationship to the biliary anatomy at opera-
tion, the size of the biliary ducts, or the degree of hepatic fibrosis in this study; therefore, it appears to be of uncertain importance. Biliary atresia continues to be a subject of considerable debate. However, the premise has now changed from Kasai versus no Kasai, to Kasai versus transplantation. The Registry data do not as yet address the transplantation issue. However, they do provide valuable information about the Kasai operation, the complications associated with it, and the long-term prognosis of children with biliary atresia. Although suffering from all of the limitations of any study performed on the basis of a survey, many significant prognostic indicators have been identified. The influence of age at operation makes continued early diagnosis and referral paramount. The unexpected racial influence on outcome deserves further investigation. The presence of totally obliterated extrahepatic ducts, grossly or microscopically, and hepatic cirrhosis forbode poor outcome, as does the development of varices or ascites. Unfortunately, none of these variables, solely or in combination, can accurately predict 100% failure. Further development of prognostic indicators may ultimately result in direct referral for transplantation, but at present the long-term survival with Kasai’s operation for biliary atresia supports our ongoing efforts. ACKNOWLEDGMENT The authors acknowledge the contributions of Drs R.P. Altman, D.C. Hitch, E.I. Smith, D.M. Hays, and J. Uceda, who are responsible for the development of the Registry and the initial patient survey.
REFERENCES 1. Campbell DP, Poley JR, Bhatia M, et al: Hepatic portoenterostomy-Is it indicated in the treatment of biliary atresia? J Pediatr Surg 9:329-333, 1974 2. Lilly JR, Altman RP: Hepatic portoenterostomy (the Kasai operation) for biliary atresia. Surgery 78:76-86, 1975 3. Arcari F: Discussion of Altman RP, Chandra RS, Lilly JR: Ongoing cirrhosis after successful porticoenterostomy in infants with biliary atresia. J Pediatr Surg 10~690,1975 4. Kasai M, Kimura S, Asakura Y, et al: Surgical treatment of biliary atresia. J Pediatr Surg 3:665-675,1968 5. Koop CE: Progressive extrahepatic biliary obstruction of the newborn. J Pediatr Surg 10:169-170. 1975 6. Shim WKT, Kasai M, Spense MA: Racial influence on the incidence of biliary atresia. Prog Pediatr Surg 6:53-62, 1974 7. Landing BH: Considerations of the pathogenesis of neonatal hepatitis, biliary atresia, and choledochal cyst-The concept of infantile obstructive cholangiopathy. Prog Pediatr Surg 6:113-139, 1974 8. Hitch DC, Shikes RH, Lilly JR Determinants of survival after Kasai’s operation for biliary atresia using actuarial analysis. J Pediatr Surg 14:310-314, 1979 9. Kasai M: Treatment of biliary atresia with special reference to
hepatic porto-enterostomy and its modifications. Prog Pediatr Surg 6:5-52,1974 10. Ohi R, Shikes RH, Stellin GP, et al: In biliary atresia duct histology correlates with bile flow. J Pediatr Surg 19:467-470, 1984 11. Miyano T, Suruga K, Tsuchiya H: A histopathological study of the remnant of extrahepatic bile duct in so-called uncorrectable biliaryatresia. J Pediatr Surg 12:19-25, 1977 12. Gautier M, Jehan P, Odievre M: Histologic study of biliary fibrosis remnants in 48 cases of extrahepatic biliary atresia: Correlation with postoperative bile flow restoration. J Pediatr 89:704-709, 1976 13. Schweizer P: Treatment of extrahepatic bile duct atresia: Results and long-term prognosis after hepatic portoenterostomy. Pediatr Surg Int 1:30-36, 1986 14. Miyamoto M, Kajimoto T: Associated anomalies in biliary atresia patients, in Kasai M (ed): Biliary Atresia and Its Related Disorders. Amsterdam, The Netherlands, Excerpta Medica, 1983, pp 13-19 15. Lilly JR, Chandra RS: Surgical hazards of co-existing anomalies in biliary atresia. Surg Gynecol Obstet 139:49-54, 1974
BILIARY ATRESIA REGISTRY
1081
Discussion H. Filston (Durham, NC): I think it is important to look at John Lilly’s series that was just presented at the American Surgical Society and published in a recent issue of Annals of Surgery in conjunction with the results of this survey. This survey involved a tremendous amount of work, but it represents a mix of surgeons. I’ve heard many surgeons state that they do not believe in this operation, and if you don’t believe in an operation, you may not do very well with it. Even though it appears that about 35% or 40% of children benefit from this surgery, it is very important to look at how well some of them benefit from it. Many of the children who get a really superb result from this operation end up growing in the 50th to 75th percentile. They are on no medications, they eat normally, they grow normally, they participate in every activity from sports to cheerleading to music to whatever. We have to be very careful that we do not put this
operation down because we do not save everybody with it; we must continue to look at it as an operation that will benefit a certain number of children extremely well. Those who get that benefit are going to be much better off than a child with a liver transplant. A liver transplantation is not precluded for the 60% or so who may not, in the long run, get a lifetime benefit from it. So I hope we don’t hear any more comments that this is an operation “of historical interest only.” F. Karrer (response): I agree with all of your comments, Dr Filston, especially your point about this registry representing a very heterogenous mix of surgeons and surgical techniques. In order to get more consistent data and answer some more specific questions, I think that if this study is to be continued, it would be optima1 to choose a few selected centers that have an interest in this problem, rather than continue a general nationwide survey.
M. Karrer, John R. Lilly, Barbara A. Stewart, Denver,
Biliary atresia;
Kasai procedure;
portoen-
I
N 1976, THE SURGICAL Section of the American Academy of Pediatrics established a nationwide registry of children with extrahepatic biliary atresia. Reports of the early experience with Kasai’s new operation, portoenterostomy, showed widely disparate results.“2 The long-term success of the procedure
From the Department of Surgery, University of Colorado School of Medicine, Denver, CO. Supported in part by the Surgical Section of the American Academy of Pediatrics, Elk Grove Village, IL; by Grant No. RR-00069 from the General Clinical Research Centers Program, Division of Research Resources, National Institutes of Health, Bethesda, MD; by the Pediatric Liver Center, University of Cole rado Health Sciences Center, Denver, CO: and by the Johnson & Johnson Company. Presented at the 38th Annual Meeting of the Surgical Section of the American Academy of Pediatrics, Chicago, Illinois, October 21-23.1989. Address reprint requests to Frederick M. Karrer, MD. Pediatric Surgery (C314). University of Colorado School of Medicine, 4200 E Ninth Ave. Denver, CO 80262. o I990 by W.B. Saunders Company. 0022-3468/90/2510-0014$03.00/0
1076
and Roberta J. Hall
Colorado
8 The Registry provides information about 904 children with biliary atresia from more than 100 institutions. There was a 1.4 to 1 female predominance: racial distribution was 62% Caucasian, 20% black, 11% Hispanic, 4.2% asian, and 1.5% American Indian. Eight hundred sixteen (90%) underwent corrective surgery (median age at operation, 69 days). Intraoperatively, 70% had totally obliterated extrahepatic bile ducts, 22% had patency of the gallbladder and distal common duct, whereas only 8% had “correctable” biliary atresia (proximal duct patency). A variety of reconstructions were used, but the majority of patients had a Roux-en-Y portoenterostomy with or without exteriorization. Follow-up was available for 670 children (74%) with average length of follow-up of 5 years (range, 1 to 16 years). Five-year actuarial survival was 48% following Kasai’s operation, but was less than 10% (at 3 years) if no operative correction was done. Survival was unaffected by sex, type of reconstruction, or cholangitis. Predictors of a bad outcome were (1) Caucasian race: (2) operative age greater than 60 days; (3) presence of cirrhosis at initial biopsy; (4) totally nonpatent extrahepatic ducts; (5) absent ducts at the level of transection in the liver hilus; and (6) subsequent development of varices or ascites. Identification of factors predictive of the ultimate outcome provide a basis for either continued efforts with management of Kasai’s operation or for early referral for liver transplantation. Q 1980 by W.B. Saunders Company. INDEX WORDS: terostomy.
1976 to 1989
Journ
remained unproven and its widespread application was questioned.3 Because each surgeon treated only relatively few infants with biliary atresia, a nationwide survey was organized to accumulate larger numbers. By combining data from multiple centers, more valid statistical analysis of the early success and long-term prognosis are available.
MATERIALS
AND
METHODS
Members of the Surgical Section of the American Academy of Pediatrics were provided with initial patient entry forms to obtain general demographic information (sex, race, birthplace, date of birth), operative data (age at operation, cholangiographic findings, hepatic and portal histology, type of reconstruction), and early results (bile drainage, operative complications, death). The status of the patients was periodically updated to determine late complications (cholangitis, varices, rickets, ascites, growth retardation), the extent of liver disease and length of survival until death or last follow-up date. More recently, surveys have partially addressed the issue of transplantation. Nine hundred eighty-five patients were entered into the Registry from 136 surgeons representing a cross-section of practices; university hospitals, general hospitals, and private practice. The patients represented a broad geographic area from across the United States and Canada. We also received patient data from Mexico and Hong Kong. Of the 985 patients, 904 had true extrahepatic biliary atresia, 45 had biliary hypoplasia, and 36 had choledochal cyst. The latter two groups were deleted from the statistical analysis. Of those with true extrahepatic biliary atresia, there were 363 boys and 541 girls (M:F ratio, 0.7:1), with a racial distribution of 62% Caucasian, 20% black, 11% Hispanic, 4.2% asian, and 1.5% American Indian. Follow-up information was obtained on 670 children (75%). The data were entered into a computerized data base and periodically updated and revised. Statistical analysis was by x’ and analysis of variance using Statpac (Walonick Assoc, Minneapolis, MN). The survival data were analyzed as of October 1, 1989, by the method of Kaplan-Meier, and the statistical comparisons made by the Cox-Mantel test using Solo statistical system (BMDP Statistical Software, Los Angeles, CA).
RESULTS
Eight hundred sixteen of the 904 infants with biliary atresia underwent corrective surgery, the remainder either had no operation or exploratory laparotomy only. At the time of the operation, 17% were less than 30 days old, 10% were over 120 days old (median age at operation, 69 days). Cholangiograms were attempted on 880 infants. Seventy percent of patients had totally obliterated nonpatent extrahepatic bile ducts, and 22% had patency of the gallbladder and distal common duct (Fig 1). Only 8% were reported to have proximal patency, so-called “correctable” biliary of Pediatric
Sorgern Vol 25, No 10 (October), 1990: pp 1076-1081
BILIARY ATRESIA REGISTRY
Fig1. Cholangiographic findings at exploration. Patency of the gallbladder and distal common bile duct (lower right). totally obliterated extrahepatic ducts (left). and proximal duct patency, the so-called correctable type (upper right).
atresia. At the time of surgery, 25% had minimal fibrosis on liver biopsy, 43% had moderate fibrosis, and 26% had severe fibrosis/cirrhosis. Only 5% had no evidence of fibrosis at the time of initial operation. Sections of the bile ducts at the liver hilus showed microscopic ducts in over 60% and moderate inflammation in 44%. Eighty-two infants (9%) had other associated anomalies, most commonly cardiac (eg, septal defects), absent inferior vena cava, preduodenal portal vein, polysplenia, and malrotation. The type of reconstruction varied widely (Fig 2). Roux-en-Y portoenterostomy with double-barrelled exteriorization was chosen most often (38.3%) and a simple Roux-en-Y without exteriorization (Kasai I) was used in 23%. Other reconstructions were less popular--double-Y (Kasai II, 4.5%), simple intestinal conduit (Sawaguchi, 5.3%)-or had selected indications-choledochoenterostomy, 5.5%, used for correctable types and portocholecystostomy (gallbladder Kasai, 7.7%) where the distal bile ducts were patent. The early complication rate was low, with eight intraabdominal abscesses, 14 instances of small bowel obstruction, seven episodes of postoperative hemorrhage, and a total of four early deaths (0.4% operative mortality). The majority of patients achieved bile drainage after the Kasai operation, but 140 children never drained or had only temporary bile drainage. Bile drainage was verified by radionucleide scans (rose bengal or iminodiacetic acid), a decrease in serum bilirubin by more than half, bile bilirubin concentration greater than serum bilirubin, bile pigment in the stool, or other means in those who drained bile. Cholangitis was reported in 5 1% of infants during the
1077
first year postoperatively and in 44% the second year. Usually this consisted of one to three episodes per year. One quarter of patients had one or more revisions of the portoenterostomy. Of 4 15 reconstructions in which exteriorization was done, the stoma remains in 94 patients. When internalized, 25% were closed before 6 months, 39% between 6 and 12 months, 30% between 1 and 3 years, and 6% after 3 years. The late complication rate was high. Of long-term survivors (>3 years), varices were present in 49%, ascites in 63%, rickets in 32%. and developmental retardation in 65%. Subjects in the Registry were born between 1967 and 1989 (mean, 1977). Follow-up was available in 670 with average length of follow-up of 5 years (range, 1 to 15). The actuarial survival curves are shown in Fig 3. Long-term survival rate was for the entire group approximately 25%. For the subset of patients having no corrective operation (or exploratory laparotomy only), survival was limited to 3 years. Of those who died, the primary cause of death was liver failure in 64%, infections in 16.4%, gastrointestinal bleeding in 9%, and other causes in 10.9%. Prognostic Factors Liver biopsy.
The amount of fibrosis or cirrhosis on liver biopsy was found to be most strongly correlated with the patient age at the time of the operation. As the operative age increased, there was a greater amount of fibrosis or cirrhosis on liver biopsy (P < .OOl). There was no relationship between liver biopsy findings and
KASAI
23%
I
f..-B & T
DOUBLE ENTEROSTOMY 38.3%
II 4.5%
KASAI
F
. 6 SAW;GFHZ
NO OPERATION OR EXPLORATION ONLY '.,i, 1, f- F
11.9%
CHOLEOOCHOENTERO:TOMY FOR “COR;E;TABLE TYPE .%
GALLBLADDER KASAI 7.7%
Fig 2. Types of reconstruction and percentages of each. Not shown are 35 (3.8%) coded as other types of reconstruction.
1078
KARRER ET AL
80 4 f
60
6 K
40
3 -dQ
20 0
0
1
2
3
4
5
6
7
8
g 10 11 12 13 14 15 16
YEARS Fig 3. Actuarial survival curves for those undergoing operative correction of biliary atresia (solid line) and for those not having operative intervention or exploratory laparotomy only (shaded line). The groups are statistically different (P = 6001). Note that the extended survival in the “No Kasai” group represents one patient reported to be alive 11 years after exploratory laparotomy in infancy.
patient sex, race, cholangiographic findings, or bile duct histology at the liver hilus. Cholangiographicfindings. The older the infant at the time of surgery, the more often nonpatent ducts were found (P = .018). Distal patency and proximal correctable-type lesions were found more often in younger infants. Boys were significantly more likely to have distal patency of the gallbladder and common bile duct (boys 28%, girls 18%; P = .OOl). Cholangitis. The frequency of cholangitis during the first year and subsequently was significantly influenced by the type of reconstruction. Those having a “gallbladder Kasai,” portocholecystostomy, had the lowest incidence of cholangitis; 35% versus roughly 55% for the other reconstructions (P = .02). The presence of distal bile duct patency, permitting portocholecystostomy, also correlated with a lower incidence of cholangitis. The degree of hepatic fibrosis also influenced the rate of cholangitis; increasing fibrosis correlated with more cholangitis. Race had an effect as well; the incidence of cholangitis was 37% for blacks, 54% for Caucasians, 54% for asians, and 6 1% for Hispanics (P = .05). There was no correlation with patient sex, age at operation, or survival. Hepatic decompensation. The development of varices and ascites was found to correlate strongly with the amount of fibrosis on operative liver biopsy and with the cholangiographic findings. Those with more hepatic fibrosis/cirrhosis, and those with nonpatent extrahepatic ducts had significantly higher incidence of both varices and ascites. Cholangitis was also associated with the development of varices and ascites. Mortality. The lo-year actuarial survival for children having a Kasai procedure was over 30%, com-
pared with only 10% at age 3 years for those not undergoing any drainage procedure (Fig 3). Many factors influenced patient survival. The most unexpected finding was a significantly decreased survival among Caucasians versus other races (Table 1). The finding could not be explained by difference between the racial groups in terms of cholangiographic findings, liver biopsy, bile duct histology, type of reconstruction used, or age at operation. There was a lower incidence of cholangitis among blacks but cholangitis did not correlate with mortality (vide supra). Follow-up was greater for Caucasians, asians, and Hispanics (81%, 87%, 92%, respectively) than for blacks (65%). The age of the child at the time of operation had an important impact on bile drainage and on long-term survival. Infants under 30 days of age had the best chance of achieving bile drainage and for survival (Table 2). The beneficial impact of younger age seemed to plateau over 100 days of age. The operative age was also related to the bile duct size; with increasing age, total fibrosis of the bile duct remnant was more common (P = .05). The findings on exploration were also predictive of ultimate outcome. Cholangiographic findings strongly influenced survival; proximal duct patency (correctable) had the best prognosis, followed by patent distal ducts, and totally obliterated extrahepatic ducts had the worst prognosis (5-year survival rates, 72%, 62%, and 38%, respectively). Bile duct size similarly predicted outcome with mortality rates for those with a totally fibrotic porta of 77%, versus only 45% if sections at the hilum showed a grossly patent, visible duct. Inflammation of the extrahepatic ducts also had no impact on long-term outcome. On liver biopsy, minimal fibrosis at the initial operation was found to be a favorable finding, with increasing fibrosis yielding lower survival and, paradoxically, normal or near normal hepatic architecture on biopsy falling in between. The presence of associated malformations did not influence longevity (30.1% with, 35.6% without). The type of reconstruction had no effect on outcome when separated from the cholangiographic findings. That is, although the gallbladder Kasai and choledo-
Table 1. Racial Influence on Survival
Race Caucasian
Total Patients
Survival
355
32.1%
Black
89
42.7%
Hispanic
71
62.0%
Asian
26
50.0%
NOTE: Only those patients undergoing some type of reconstruction and having follow-up information were included. P -z ,001.
BILIARY ATRESIA REGISTRY
1079
Table 2. influence of Operetive Age
at Operation WYS)
Age on Bile Drainage end Survival
BileDrainage (%)
SurvivalRate (%I
530
66.7
31-60
69
6 l-90
62.3
39.5
91-120
53.3
28.6
>120
62.5 *
35 P = .004
43.6
28.8 P = .023
choenterostomy groups had improved survival, they were only applicable to certain types of biliary anatomy, distal patency and proximal patency, respectively. The other reconstructive procedures, when applied to infants with totally obliterated extrahepatic ducts, had equal survival rates and the frequency of achieving bile drainage with each of the reconstructions was not significantly different. However, the vented procedures (Roux-en-Y with exteriorization, Kasai II, and Sawaguchi) were associated with a higher frequency of cholangitis. Patients with no, or only temporary bile drainage had a poor prognosis with 3-year survival rate of only 15%. Postoperative development of varices, ascites, rickets, or developmental retardation also reduced chances of long-term survival. DISCUSSION
The original intent of the Biliary Atresia Registry was to determine the feasibility of the Kasai procedure. Following Kasai’s report in English in 1968,4 the portoenterostomy operation began to be used in centers in the United States. Published articles reported widely disparate, even contradictory results, with Campbell et al’ reporting no bile drainage in a series of 12 patients, and Lilly and Altman’ finding more than 50% successful bile drainage. This, and the absence of long-term studies, led to the establishment of a nationwide registry. By accumulation of a large number of cases from multiple centers, it was hoped that more valid conclusions about the long-term survival would result. In that regard, the Registry data quite clearly shows a 30% IO-year survival rate with Kasai’s operation, and a less than 10% 3-year survival rate without treatment. The reported poorer operative results in the United States (versus Japan),’ have been blamed on racial influence. The Registry data show that Caucasians have significantly reduced survival compared with nonwhite races (Table 1). The racial mix of patients in the Registry was dissimilar from the racial distribution in the general population of the United States, ie, fewer Caucasians (68 v 80%) and more nonwhite races. Whether this is merely a sampling error or represents a true difference in racial incidence of biliary atresia is
undetermined.6 The survival differences could not be explained by differences of age at operation, operative findings (cholangiographic, bile duct size, hepatic and porta histology), or type of reconstruction, but they were less striking when patients reported from Mexico and the Orient were eliminated. We believe the racial differences to be genuine but unexplained. There was no difference in survival between the various types of reconstruction with the exception of choledochoenterostomy for the correctable type, and a slight but not statistically significant improvement with the gallbladder Kasai for patients with distal duct patency. The implication from these data is that the most important technical factor of any reconstruction is the anastomosis at the porta hepatis, not the various loops, anastomoses, and conduits. If bile drainage is achieved, regardless of the reconstruction, survival is markedly improved; 8.5% without and 56.3% with drainage. The other point of interest is that the vented procedures had higher reported incidences of cholangitis, the complication they were designed to reduce. Possible explanations include (1) increased reporting of cholangitis in the vented group because of improved accuracy of diagosis (bile drainage visible); and (2) exteriorization, which may allow more colonization. However, cholangitis does not correlate with reduced survival (even after those with bile drainage failure are selected out [vi& infru] ). Patient age at operation was strongly correlated with bile drainage and survival (Table 2) but also with the degree of hepatic fibrosis, the type of biliary anatomy on cholangiography, and the histological findings at the porta hepatis. Increasing age was accompanied by increasing amounts of fibrosis on liver biopsy and the older patients tended to have a higher incidence of a totally obliterated tree on attempted cholangiography. There also was a trend toward smaller or absent biliary ducts in the porta hepatis in older children, supporting the concept of biliary atresia being a dynamic process with progression of disease.’ Although hepatic pathological changes varied in progression and severity, only a few patients older than 3 months were cured by corrective surgery. Therefore, prompt diagnosis with early reconstruction optimized success. Separate analysis of bile duct histology showed a strong association with survival, larger true ducts being associated with the best results and the smallest ductules and glands having reduced bile drainage and longevity. However, care must be exercised in analysis of this type of data because the histological criteria between different pathologists and surgeons may vary. Independent observations by a single pathologist or
1080
KARRER ET AL
group would be preferrable. Nevertheless, the correlation has been corroborated by a number of investigators who reported a correlation between bile duct size and survival.*-‘* Significant negative factors noted in this study included the lack of correlation between survival and (1) cholangitis; (2) associated malformations; and (3) amount of portal hilar inflammation. Cholangitis was reported to be a negative prognostic sign,13 which we were unable to confirm. The aberration could be due to variation in criteria used for diagnosis between centers, but the correlation was so weak (P = 571) that it is doubtful small adjustments would result in significance. The more likely explanation is that cholangitis only occurs in patients with bile drainage and, therefore, improved survival, whereas those with less cholangitis have been mongrelized with patients with no cholangitis due to absent postoperative bile drainage. Associated malformations, especially those connected with the polysplenia syndrome (absent inferior vena cava, situs inversus, malrotation, preduodenal portal vein, bilobed midline liver, polysplenia) are well-recognized,*49*S and occurred in 10% to 15% of patients. There is an unsubstantiated belief that this group had an unfavorable prognosis, but data failed to support that idea. In many instances, examination of the resected tissue at the porta hepatis showed a marked inflammatory infiltrate. Hitch et al* found no correlation between the degree of portal hilar inflammation and survival. The Biliary Atresia Registry analysis confirms that finding. The degree of inflammation also bore no relationship to the biliary anatomy at opera-
tion, the size of the biliary ducts, or the degree of hepatic fibrosis in this study; therefore, it appears to be of uncertain importance. Biliary atresia continues to be a subject of considerable debate. However, the premise has now changed from Kasai versus no Kasai, to Kasai versus transplantation. The Registry data do not as yet address the transplantation issue. However, they do provide valuable information about the Kasai operation, the complications associated with it, and the long-term prognosis of children with biliary atresia. Although suffering from all of the limitations of any study performed on the basis of a survey, many significant prognostic indicators have been identified. The influence of age at operation makes continued early diagnosis and referral paramount. The unexpected racial influence on outcome deserves further investigation. The presence of totally obliterated extrahepatic ducts, grossly or microscopically, and hepatic cirrhosis forbode poor outcome, as does the development of varices or ascites. Unfortunately, none of these variables, solely or in combination, can accurately predict 100% failure. Further development of prognostic indicators may ultimately result in direct referral for transplantation, but at present the long-term survival with Kasai’s operation for biliary atresia supports our ongoing efforts. ACKNOWLEDGMENT The authors acknowledge the contributions of Drs R.P. Altman, D.C. Hitch, E.I. Smith, D.M. Hays, and J. Uceda, who are responsible for the development of the Registry and the initial patient survey.
REFERENCES 1. Campbell DP, Poley JR, Bhatia M, et al: Hepatic portoenterostomy-Is it indicated in the treatment of biliary atresia? J Pediatr Surg 9:329-333, 1974 2. Lilly JR, Altman RP: Hepatic portoenterostomy (the Kasai operation) for biliary atresia. Surgery 78:76-86, 1975 3. Arcari F: Discussion of Altman RP, Chandra RS, Lilly JR: Ongoing cirrhosis after successful porticoenterostomy in infants with biliary atresia. J Pediatr Surg 10~690,1975 4. Kasai M, Kimura S, Asakura Y, et al: Surgical treatment of biliary atresia. J Pediatr Surg 3:665-675,1968 5. Koop CE: Progressive extrahepatic biliary obstruction of the newborn. J Pediatr Surg 10:169-170. 1975 6. Shim WKT, Kasai M, Spense MA: Racial influence on the incidence of biliary atresia. Prog Pediatr Surg 6:53-62, 1974 7. Landing BH: Considerations of the pathogenesis of neonatal hepatitis, biliary atresia, and choledochal cyst-The concept of infantile obstructive cholangiopathy. Prog Pediatr Surg 6:113-139, 1974 8. Hitch DC, Shikes RH, Lilly JR Determinants of survival after Kasai’s operation for biliary atresia using actuarial analysis. J Pediatr Surg 14:310-314, 1979 9. Kasai M: Treatment of biliary atresia with special reference to
hepatic porto-enterostomy and its modifications. Prog Pediatr Surg 6:5-52,1974 10. Ohi R, Shikes RH, Stellin GP, et al: In biliary atresia duct histology correlates with bile flow. J Pediatr Surg 19:467-470, 1984 11. Miyano T, Suruga K, Tsuchiya H: A histopathological study of the remnant of extrahepatic bile duct in so-called uncorrectable biliaryatresia. J Pediatr Surg 12:19-25, 1977 12. Gautier M, Jehan P, Odievre M: Histologic study of biliary fibrosis remnants in 48 cases of extrahepatic biliary atresia: Correlation with postoperative bile flow restoration. J Pediatr 89:704-709, 1976 13. Schweizer P: Treatment of extrahepatic bile duct atresia: Results and long-term prognosis after hepatic portoenterostomy. Pediatr Surg Int 1:30-36, 1986 14. Miyamoto M, Kajimoto T: Associated anomalies in biliary atresia patients, in Kasai M (ed): Biliary Atresia and Its Related Disorders. Amsterdam, The Netherlands, Excerpta Medica, 1983, pp 13-19 15. Lilly JR, Chandra RS: Surgical hazards of co-existing anomalies in biliary atresia. Surg Gynecol Obstet 139:49-54, 1974
BILIARY ATRESIA REGISTRY
1081
Discussion H. Filston (Durham, NC): I think it is important to look at John Lilly’s series that was just presented at the American Surgical Society and published in a recent issue of Annals of Surgery in conjunction with the results of this survey. This survey involved a tremendous amount of work, but it represents a mix of surgeons. I’ve heard many surgeons state that they do not believe in this operation, and if you don’t believe in an operation, you may not do very well with it. Even though it appears that about 35% or 40% of children benefit from this surgery, it is very important to look at how well some of them benefit from it. Many of the children who get a really superb result from this operation end up growing in the 50th to 75th percentile. They are on no medications, they eat normally, they grow normally, they participate in every activity from sports to cheerleading to music to whatever. We have to be very careful that we do not put this
operation down because we do not save everybody with it; we must continue to look at it as an operation that will benefit a certain number of children extremely well. Those who get that benefit are going to be much better off than a child with a liver transplant. A liver transplantation is not precluded for the 60% or so who may not, in the long run, get a lifetime benefit from it. So I hope we don’t hear any more comments that this is an operation “of historical interest only.” F. Karrer (response): I agree with all of your comments, Dr Filston, especially your point about this registry representing a very heterogenous mix of surgeons and surgical techniques. In order to get more consistent data and answer some more specific questions, I think that if this study is to be continued, it would be optima1 to choose a few selected centers that have an interest in this problem, rather than continue a general nationwide survey.
