SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

This study analyzed the results of the decentralized management of biliary atresia (BA) in France, where an improved collaboration between centers has been promoted since 1997. Results were compared to those obtained in England and Wales, where BA patients have been centralized in three designated centers since 1999. According to their birth dates, BA patients were divided into two cohorts: cohort A, with patients born between 1986 and 1996, had 472 patients; and cohort B, with patients born between 1997 and 2002, had 271 patients. Survival rates were calculated according to the Kaplan-Meier method and compared by using the log rank test and the Cox model. Four-year overall BA patient survival was 73.6% (95% CI 69.5%-77.7%) and 87.1% (CI 82.6%-91.6%) in cohorts A and B, respectively (P < .001). Median age at time of the Kasai operation was 61 and 57 days in cohorts A and B, respectively (NS). Four-year survival with native liver after the Kasai operation was 40.1% and 42.7% in cohorts A and B, respectively (NS): 33.9% (cohort A) and 33.4% (cohort B) in the centers with two or fewer caseloads a year, 30.9% (cohort A) and 44.5% (cohort B) in the centers with 3-5 cases/year, 47.8% (cohort A) and 47.7% (cohort B) in the center with more than 20 caseloads a year. In cohorts A and B, 74 (15.7%) and 19 (7%) patients, respectively, died without liver transplantation (LT). Four-year survival after LT was 75.1% and 88.8% in cohorts A and B, respectively (P = .006). In conclusion, BA patients currently have the same chance of survival in France as in England and Wales. The early success rate of the Kasai operation remains inferior in the centers with limited caseloads in France, leading to a greater need for LTs in infancy and early childhood. (HEPATOLOGY 2006;44:75–84.)

Biliary atresia is an obstructive cholangiopathy of unknown origin that occurs in the perinatal period. Bile duct obliteration variably affects extra- and intrahepatic bile ducts, leading to severe cholestasis, liver fibrosis, and cirrhosis.1 The current treatment for biliary atresia is sequential2–4: in the perinatal period, the Kasai operation5, 6 or its technical variants aim to restore the biliary flow toward the intestine. Secondary LT is needed if jaundice fails to clear and /or when complications of biliary cirrhosis occur.

In the last decades, several prognostic factors for BA have been identified. Some are related to characteristics of the disease (and cannot be altered): existence of a polysplenia syndrome,7–9 anatomical pattern of the extrahepatic biliary remnant,7, 10–12 histological lesions of the biliary remnant,13, 14 and degree of liver fibrosis at the time of the Kasai operation.15–20 Other prognostic factors are related to the management of BA patients and are therefore improvable: age at Kasai operation,7, 10, 12, 21 accessibility to liver transplantation (LT),22 experience of the treating center in the management of BA patients, especially in the performance of the Kasai operation and pediatric liver transplantation.7, 23, 24

In the United Kingdom (population 60 million and incidence of BA 40-45 patients per year), two studies showed that the results of treatment were infraoptimal in centers with limited patient caseloads.23, 24 These findings led to the centralization of BA patients from England and Wales in three designated pediatric liver units, able to manage children from diagnosis to LT. After 3 years, a new survey including all BA patients born between 1999 and 2002 showed that this centralization policy allowed standardization of the results at the highest level, nationwide.25

In France, where the population, the incidence of BA, and the level of economic development are similar to those of the United Kingdom, a first national study also showed a high discrepancy in results according to center caseload: in this survey, which included all BA patients born between 1986 and 1996, 5-year overall patient survival decreased from 78% at the center with a caseload of more than 20 new BA patients a year to 56% at centers with a caseload of less than 3 new BA patients a year.7 The impact of age at Kasai operation was also confirmed, and mortality without LT was 15.2%, partly because of deaths that occurred while waiting for a liver graft. National results could therefore be improved by standardizing the practices and results of the centers, reducing the age at which the Kasai operation is performed, and promoting surgical techniques that increase the availability of pediatric liver grafts (split liver grafts, liver donation by related living donors). Contrary to the United Kingdom experience, decentralization of patients was maintained, but collaboration between centers nationwide was reinforced. A national observatory of BA was created to promote and evaluate this collaborative policy.

The aims of the present study were: (1) to determine the current prognosis of BA in France, (2) to analyze the evolution of the prognostic factors of BA since the first national study, and (3) to compare the results of the British centralized and the French decentralized policies of management of BA patients.

Patients and Methods

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

The charts of all children with BA or suspected BA treated in France since LT became available were reviewed by the study investigators, who visited the participating centers and analyzed the charts locally. All 45 pediatric centers (including medical, surgical, and LT units) involved in the management of BA patients in France contributed to this survey, and therefore data collection was exhaustive.

Inclusion criteria were: (1) having a diagnosis of BA made by clinical, biochemical, and radiological data, surgical findings, and liver histology, all consistent with BA, with other causes of neonatal cholestasis ruled out26 (results of histological examination of biliary tract remnants were available for 584 patients [78%]; diagnostic criteria were checked by the investigators in each patient before inclusion to the study); and (2) living in France and born between January 1986 and December 2002. The Kasai operation and LT were available for every child during this period, and the medical costs were covered by the national medicosocial insurance system.

Registered data included: identity, birth date, date of first symptoms, whether the Kasai operation had been performed, date of Kasai operation, team in charge, anatomical pattern of BA, evolution of bilirubin with time, whether LT was performed, date of LT, indication for LT, prothrombin time before LT, team in charge, type of liver graft, outcome after LT, whether there had been retransplantation, final outcome, date of last follow-up, cause of death, and, in some cases, reasons why the Kasai operation and/or LT was not performed. Macroscopic classification was based on the anatomical pattern of the extrahepatic biliary remnant: type 1 if atresia was limited to the common bile duct; type 2 if there was a biliary cyst in the liver hilum communicating with dystrophic intrahepatic bile ducts; type 3 if the gallbladder, cystic duct, and common bile duct were patent; or type 4 if there was complete extrahepatic atresia. Because of the retrospective nature of the study, data collection in some cases was less than 100%.

Children were divided into two cohorts according to birth date: cohort A if born between 1986 and 1996 and cohort B if born between 1997 and 2002. Early success of the Kasai operation was defined by complete clearance of jaundice and bilirubin of ≤20 μmol/L. The level of experience of the centers in the management of BA patients was determined according to: (1) number of Kasai operations performed each year for analysis of survival with native liver; (2) number of pediatric LTs performed each year for analysis of survival after LT; and (3) number of new BA patients treated each year for analysis of overall patient survival. If a child had been treated in several centers, the center to which the child was finally referred was considered the center in charge for the overall outcome analysis.

Median follow-up in survivors was 7 years (range 0.2-18.1 years). This study received the agreement of French authorities (Commission Nationale de l'Informatique et des Libertés, authorization number 997085). The registry was strictly confidential. Each family received an information letter indicating the aims and methods of the study and had the possibility to check or correct the data recorded for their child.

Statistical Analysis.

Survival rates were estimated according to the Kaplan-Meier method: (1) survival with native liver (with liver transplantation or death as the end points), (2) survival after liver transplantation (with death as the end point), and (3) overall patient survival (with death as the end point). Results were expressed as calculated survival rate with 95% confidence interval. Univariate analyses used the log-rank test, and multivariate analyses, when appropriate, were done with the Cox regression model. The proportional hazard ratio was verified graphically. Categorical data were compared with the χ2 test, using Yate's correction if indicated. All significance tests were two tailed with alpha fixed at .05. All statistical analyses were done with SAS software (version 8.1; SAS Institute Inc., Cary, NC).

Results

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Patients Included.

A total of 743 children were included in the study, 472 in cohort A and 271 in cohort B. The ratio of males to females was 0.82. Six hundred and fifty children (87.5%) were born in metropolitan France, 67 (9.0%) in the overseas territories, and 26 (3.5%) abroad. The incidence of BA in metropolitan France was 5.12 (4.63-5.61) per 100,000 live births between 1986 and 1996 (cohort A) and 5.09 (4.45-5.73) per 100,000 live births between 1997 and 2002 (cohort B). The difference in incidence between the two cohorts was not significant.

Eleven additional patients who had been misdiagnosed as having BA and who underwent a Kasai operation (1.6%) were secondarily identified (and excluded from this study). The corrected diagnoses were Alagille's syndrome for five patients, alpha 1 antitrypsin deficiency for two patients, cystic fibrosis for two patients, choledocal cyst for one patient, and ischemic cholangiopathy for one patient.

Medical History of Patients.

The medical history of the patients is summarized in Fig. 1. A total of 695 (93.5%) children underwent the Kasai operation or a derived procedure. Complete clearance of jaundice was achieved in 242 of 668 patients (36.2%) in whom follow-up bilirubin levels were available: 34.4% (144/418) in cohort A and 39.5% (98/250) in cohort B (NS).

thumbnail image

Figure 1. Outcomes of 743 infants with BA. All patients with BA living in France, born between January 1986 and December 2002: cohort A, patients born between 1986 and 1996; cohort B, patients born between 1997 and 2002.

Download figure to PowerPoint

Three hundred and eighty-six patients underwent 456 LTs, 62 had a second transplant, and 8 a third transplant. The median age at first transplantation was 18.6 months (range: 5.2-172.1).

In June 2004, 566 patients (76.2%) were alive: 328 (69.5%) in cohort A and 238 (87.8%) in cohort B. Two hundred and sixty-four children (46.6%) were alive with their native liver: 141 (43.0%) in cohort A and 123 (51.7%) in cohort B. Three hundred and two (53.4%) had been transplanted: 187 (57.0%) in cohort A and 115 (48.3%) in cohort B. One hundred and seventy seven children (23.8%) had died: 12 without surgical treatment, all of them (2.5%) in cohort A; 81 (11.6%) after Kasai operation without LT: 62 (14.1%) in cohort A and 19 (7.5%) in cohort B (P = .009) ; and 84 after liver transplantation: 70 (27.2%) in cohort A and 14/129 (10.9%) in cohort B (P < .001).

Patients Without Kasai Operation.

Forty-eight (6.5%) children did not undergo the Kasai operation nor any of its technical variants (32 in cohort A, 16 in cohort B, NS). The reason was delayed diagnosis in 30, contraindication to curative surgery in 2 (polymalformation syndromes), and miscellaneous or unknown in 16. Survival with native liver in these patients was 58.4% at 1 year (28 patients), 28.4% at 2 years (13 patients), 10.9% at 3 years (5 patients), 4.4% at 4 years (2 patients), and 2.2% at 5 years (1 patient). Forty-six patients without contraindications for curative surgery were oriented to de novo LT (without previous Kasai operation), 35 patients were transplanted and one was waiting for LT at last follow-up. Four-year overall survival of the patients oriented to de novo LT was 66.2%: 53.3% in cohort A and 93.3% in cohort B (P = .01). Four-year survival after liver transplantation was 82.6%: 75.0% in cohort A and 93.3% in cohort B (NS).

Survival With Native Liver (Table 1 and Figs. 1 and 2).

Four-year survival with native liver after Kasai operation was 40.6%: 40.1% and 42.7% in cohorts A and B, respectively (NS). Median age at operation was 60 days (12-180): 61 days in cohort A and 57 days in cohort B (NS). The Kasai operation was performed in 30 centers in cohort A and 22 centers in cohort B.

Table 1. Survival With Native Liver
Univariate AnalysisCohort A (N = 439)*Cohort B (N = 253)*A Versus B
Number of Data Sets Available for Analysis4-Year Survival With Native Liver (Number of Patients Alive With Native Liver at 4 Years)PValueNumber of Data Sets Available for Analysis4-Year Survival With Native Liver (Number of Patients Alive with Native Liver at 4 Years)PValuePValue
  • *

    Three French patients who underwent a Kasai operation abroad were excluded from this analysis.

All patients who underwent a Kasai operation*[95% confidence interval]43840.1% (168) 25242.7% (47) NS
  [35.5-44.7]  [35.8-49.6] 
Age at Kasai operation       
≤ 45 days11246.5% (50).0097751.2 (19).08NS
> 45 days32137.5% (115) 17239.8 (28) NS
Anatomical pattern       
1+24771.7% (32)< .00011966.2% (4).15NS
3+438336.0% (132) 22143.2% (43) NS
Polysplenia syndrome       
 Yes3518.4% (6).00181919.2% (2).05NS
 No38242.1% (154) 22345.0% (45) NS
Level of center experience (number of Kasai operations performed/year)       
 ≥ 20208 (1 center)47.8% (97)< .0001123 (1 center)47.7% (20).19NS
 [3-5]68 (2 centers)30.9% (20) 57 (3 centers)44.5% (13) .04
 <2162 (27 centers)33.9% (51) 72 (18 centers)33.4% (14) NS
        
Multivariate Analysis412 Data Sets Available for AnalysisRisk Ratio [95% CI]PValue236 Data Sets Available for AnalysisRisk Ratio [95% CI]PValue 
Age at Kasai operation       
 ≤ 45 days 1.01 1.066 
 > 45 days 1.43 [1.07-1.91]  1.47 [0.97-2.23]  
Anatomical pattern       
 1+2 1< .0001 1.23 
 3+4 2.08 [1.82-5.20]  1.65 [0.72-3.79]  
Polysplenia syndrome       
 No 1.001 1.03 
 Yes 1.89 [1.29-2.77]  1.87 [1.05-3.34]  
Level of center experience (number of Kasai operations performed/year)       
 ≥ 20 1  1  
 [3-5] 2.08 [1.51-2.88]< .0001 0.71 [0.43-1.20].20 
 < 2 1.35 [1.04-1.76]0.02 1.25 [0.82-1.89].29 
thumbnail image

Figure 2. Outcomes of 472 patients in cohort A (1986-1996). Overall patient survival (blue curve) was 72.0% ± 4.1%, 70.2% ± 4.1%, and 67.9% ± 4.9%, at 5, 10 and 15 years, with 323, 197 and 46 patients alive, respectively. Survival with native liver (including the patients who did not ungergo the Kasai operation) (red curve) was 35.0% ± 4.3%, 29.7% ± 4.3%, and 25.7% ± 4.7%, at 5, 10 and 15 years, with 156, 83 and 18 patients alive with their native livers, respectively.

Download figure to PowerPoint

In cohort A, better survival with native liver was associated with early Kasai operation (before 45 days), “favorable” anatomical patterns types 1 and 2, nonsyndromic type of BA, and a high annual caseload in the center performing the Kasai operation. These factors were independent by multivariate analysis. In cohort B, the same prognostic factors were associated with survival with native liver, but statistical significance was reached only for polysplenia syndrome. The only difference observed between cohorts A and B concerned the results of the centers performing 3-5 Kasai operations per year: 4-year survival with native liver after Kasai operation in such centers increased from 30.9% in cohort A to 44.5% in cohort B (P = .04).

Survival After Liver Transplantation (Table 2).

Four-year survival after liver transplantation was 79.1% (74.9%-83.3%). Survival was significantly greater in cohort B than in cohort A, 88.8% versus 75.1%, respectively (P = .006). Nine centers performed pediatric LTs in cohort A, with 3 centers whose annual caseload was less than 3, five centers whose caseload was 3-5, and one center whose caseload was more than 20. Seven centers performed pediatric LTs in cohort B, with three centers whose annual caseload was less than 3, three whose caseload was 5-15, and one center whose caseload was more than 20.

Table 2. Survival After Liver Transplantation
Univariate AnalysisCohort A (N = 254)*Cohort B (N = 129)A Versus B
Number of Data Sets Available for Analysis4-Year Survival After LT (Number of Patients Alive 4 Years after LT)PValueNumber of Data Sets Available for Analysis4-Year Survival after LT (Number of Patients Alive 4 Years after LT)PValuePValue
  • Abbreviations: PT, Prothrombin time (% of normal value); NS, not significant.

  • *

    Three French patients transplanted abroad were excluded from this analysis.

  • *

    Multivariate analysis: not relevant.

All patients transplanted in France: 383*25075.1% (173) 12988.8% (22) .006
 [95% CI] [69.7-80.5]  [83.3-94.3]  
Age at LT       
 ≤ 1 year4355.8% (24).00054490.9% (10)NS.0007
 > 1 year20779.1% (149) 8587.6% (12) NS
Degree of liver failure       
 PT ≤ 50%6160.4% (35).00033287.5% (4)NS.0179
 PT > 50%15382.2% (114) 8791.6% (17) 0.09
Type of liver graft       
 Full-size8477.2% (62)NS2385.4% (5)NSNS
 Reduced size10973.4% (75) 4290.5% (10) .0318
 Split graft3573.9% (23) 4292.4% (4) .0567
 Living related donor1580.0% (10) 1888.9% (3) NS
Level of center experience       
 ≥ 5 pediatric LT/year163 (1 center)82.1% (125).0001119 (4 centers)89.6% (21)NSNS
 < 5 pediatric LT/year87 (8 centers)61.9% (48) 10 (3 centers)80.0% (1) NS
Multivariate Analysis214 Data Sets Available for AnalysisRisk RatioPValue119 Data Sets Available for AnalysisRisk RatioPValue 
Age at LT >1 year 1.20 **  
≤ 1 year 1.54 [0.79-2.98]     
Degree of liver failure    **  
 PT > 50% 1.0318    
 PT ≤ 50% 1.94 [1.06-3.55]     
Level of center experience    **  
 ≥ 5 pediatric LT/year 1.02    
 < 5 pediatric LT/year 1.92 [1.09-3.38]     

In cohort A, worse survival after LT was associated with early LT (before the age of 1 year), liver failure before LT (prothrombin time < 50%), and a center caseload of pediatric LTs of less than 5/year. By multivariate analysis, only preoperative liver failure and the level of center experience appeared as independent prognostic factors. Survival after LT did not depend on the type of liver graft used for transplantation. In cohort B, none of these prognostic factors persisted. Survival of patients transplanted with pre-operative liver failure, as well as survival of patients transplanted before the age of 1 year, dramatically improved in cohort B compared to cohort A. Survival after LT increased from 82.1% to 89.6% in the centers performing more than 5 LTs per year and increased from 61.9% to 80.0% in centers with smaller caseloads.

Deaths of Children With Biliary Atresia (Table 3).

Actual mortality of BA patients decreased from 30.5% in cohort A to 12.2% in cohort B. This improvement was found in every step of the history of the patients: mortality without any surgical treatment, mortality after the Kasai operation without LT (including mortality on waiting list for LT), and mortality after LT.

Table 3. Causes of Death
 Cohort ACohort BP Value
  • **

    Polymalformation syndromes.

All Patients With BA472271 
Number of deaths144 (30.5%)33 (12.2%) 
Deaths without LT74 (15.6%)19 (7.0%) 
— contraindications to curative surgery2*0 
— deaths from complications of the Kasai operation72 
— contraindications to LT214 
ratio contraindications/transplantations21/257 (8.2%)4/129 (3.1%).07
— deaths while waiting for LT243 
ratio deaths while waiting for LT/transplantations24/257 (8.5%)3/129 (2.1%).02
— miscellaneous or unknown cause of death2010 
Deaths after LT70/257 (27.2%)14/129 (10.8%) 
Patients with at Least 2 Years Follow-Up464222 
Number of deaths in the first 2 years of life96 (20.7%)26 (11.7%).005
Deaths without LT in the first 2 years of life58 (12.5%)17 (7.6%).06

Overall Patient Survival (Table 4 and Figs. 2 and 3).

Four-year overall survival of all patients with BA from diagnosis was 78.1% and significantly increased from 73.6% in cohort A to 87.3% in cohort B (P < .0001).

Table 4. Overall BA Patients Survival
Univariate AnalysisCohort A (N = 472)Cohort B (N = 271)A versus B
Number of Data Sets Available for Analysis4-year Overall Patient Survival (Number of Patients Alive at 4 Years)PValueNumber of Data Sets Available for Analysis4-Year Overall Patient Survival (Number of Patients Alive at 4 Years)PValuePValue
  • *

    Early success of the Kasai operation defined as postoperative clearance of jaundice with bilirubin level ≤ 20 μmol/L.

  • *, **

    When a child was treated in several centers, the center to which the child was finally referred was considered the treating center.

  • ***

    Multivariate analysis not relevant.

All patients treated in France46573.6% (331) 27087.1% (97) <.0001
[95% CI] [69.6-77.6]  [82.7-91.5]  
Kasai operation       
Early success of the Kasai operation*14397.8% (132)<.000198100% (38)<.0001NS
Failure of the Kasai operation26865.2% (172) 15279.4% (51) .008
No Kasai operation3250.0% (16) 1693.3% (6) .006
Experience of the center       
Number of new BA patients/ year**       
 ≥20277 (1 center)81.2% (222)<.0001163 (1 center)89.6% (54)NS.058
 3-575 (2 centers)60.8% (44) 56 (3 centers)87.3% (24) .0007
 ≤2113 (23 centers)63.2% (65) 51 (14 centers)79.3% (19) .055
Multivariate Analysis447 Sets Available for AnalysisRisk RatioPValue226 Sets Available for AnalysisRisk RatioPValue 
Kasai operation       
Early success* 1  ***  
Failure 4.9 [2.9-8.3]<.0001    
Not performed 6.9 [3.5-13.6]<.0001    
Experience of the center       
Number of new BA patients/year**       
 ≥20 1  ***  
 3-5 2.0 [1.3-3.1].002    
 ≤2 2.4 [1.6-3.4]<.0001    
thumbnail image

Figure 3. Survival of 271 patients in cohort B (1997-2002). Overall patient survival (blue curve) was 87.9% ± 4.1% and 87.1% ± 4.3%, at 3 and 5 years, with 140 and 48 patients alive, respectively. Survival with native liver (including the patients who did not undergo the Kasai operation) (red curve) was 43.6% ± 6.5% and 35.1% ± 7.6%, at 3 and 5 years, with 73 and 19 patients alive with their native livers, respectively.

Download figure to PowerPoint

In cohort A, 4-year overall survival was 97.8% for patients who had had a successful Kasai operation, 65.2% for patients who had had a failed Kasai operation, and 50% for patients without a Kasai operation (P < .0001). In cohort B, 4-year overall survival was 100%, 79.4%, and 93.3%, respectively, for the same subgroups (P < .0001). Four-year overall survival of all patients by Kasai operation status (excluding 2 patients with contraindications to curative surgery) was 75.3% for those who had had a Kasai operation versus 53.3% for those who had not in cohort A (P = .01) and 86.6% versus 93.3% in cohort B (P = NS).

Overall patient survival increased from cohort A to cohort B whatever the center caseload: from 63.2% to 79.3% in the centers treating 2 or fewer new BA patients per year, from 60.8% to 87.3% in the centers treating 3-5 new BA patients per year, and from 81.2% to 89.6% in the center treating more than 20 new BA patients per year.

Discussion

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Survival of BA patients with their native liver relies on the success of the Kasai operation. In the cases where the Kasai operation has not been performed or fails to restore the biliary flow, cirrhosis progresses rapidly, leading to LT or death. In the present study, survival of 48 patients with native liver who did not have the Kasai operation was 58%, 28%, 11%, and 2% at 1, 2, 3, and 4 years, respectively. These findings are similar to those of the American BA registry (from 1976 to 1989), in which the survival of 44 patients who did not have the Kasai operation was 10% at 3 years.12 On the other hand, when jaundice can be cleared by the Kasai operation, survival with native liver is possible until adulthood, although most of these patients have liver fibrosis or cirrhosis.27, 28

Several prognostic factors of the Kasai operation are related to the characteristics of the disease in an individual patient and cannot be changed, such as the presence of a polysplenia syndrome,7–9 the anatomical pattern of the extrahepatic biliary remnant,7, 10–12 the histological lesions of the biliary remnant,13, 14 and the degree of liver fibrosis at the time of the Kasai operation.15–20 Other prognostic factors, which depend on the organization of the care to BA patients, have been shown to have a major impact on outcome; they include age at time of the Kasai operation,7, 10, 12, 21 accessibility to LT,22 and level of experience of the treating center in the management of BA patients.7, 23, 24 Despite the efforts in France to diagnose BA earlier, mainly by providing repeated information to pediatricians and general practitioners on the importance of immediately investigating every neonate with jaundice that lasts more than 2 weeks,29, 30 the age when the Kasai operation was performed did not significantly decrease between cohorts A (1986-96) and B (1997-2002), with a median age of 61 versus 57 days, respectively. The present study has confirmed that the chance of prolonged survival with native liver decreases when the age at time of the Kasai operation increases (Table 1). Further efforts for earlier diagnosis could focus on the families and nonmedical health personnel, for instance, by adding items on the normal and abnormal colors of stools to the personal health leaflet of each neonate (an action currently being implemented in France), a simple system inspired by the stool color card used in Japan.31

The experience in the performance of the Kasai operation was shown to have a major impact on its results. In the United Kingdom, which has a population of about 60 million people and a BA incidence of about 40-45 new BA patients/year, the discrepancy of results between centers according to their experience in the management of such patients23, 24 led to the centralization of all English and Welsh patients in three designated pediatric liver units able to manage the patient from diagnosis to LT. This policy led to high-level results for the whole national cohort.25 In France, whose population and incidence of BA are very similar to those in the United Kingdom, the organization of care to BA patients was similarly shown to have a major impact on prognosis.7 A different policy, based on increased collaboration between centers, was adopted in order to standardize the results at the highest level nationwide. After feedback of the results of the first national study 1986-96 to all participating centers, a French observatory of BA was created, and centers with limited caseloads were encouraged to ask for advise or help on medical as well as surgical issues from centers with greater experience in the management of BA patients. Surgeons who rarely performed the Kasai operation were offered the option either of asking a surgeon with greater experience in BA to assist in their local centers or of referring the child to a center with higher caseloads and be invited to that center to operate on the child with the experienced team. What changes were observed in the management of BA patients during these years? The number of centers performing the Kasai operation decreased from 30 to 22, suggesting that eight centers with limited caseloads spontaneously stopped performing the occasional Kasai operation. These patients were mainly referred to the centers with intermediate caseloads: centers with limited caseloads (≤2 Kasai operations/year) performed 37.0% and 28.6% of all Kasai operations in cohorts A and B, respectively; centers with intermediate caseloads (3-5 Kasai operations/year) performed 15.5% and 22.6%; and the center with the highest caseload (>20 Kasai operations/year) performed 47.5% and 48.8% (Table 1). No improvement was seen in 4-year survival with native liver of the patients operated on in the centers performing at most two Kasai operations per year: 33.9% and 33.4% in cohorts A and B, respectively. These results are significantly worse than the European standards: 51% (42%-59%) in the British national cohort 1999-2002, 48% (38%-58%) in the French center performing more than 20 Kasai operations/year. Interestingly, the centers with intermediate caseloads showed results that significantly improved from the 1986-96 cohort to the 1997-2002 cohort: 4-year survival with native liver increased from 31.0% to 44.5%.

Deaths without liver transplantation significantly decreased between the two periods of the studies (Table 3). The reduction in the number of deaths without any surgery, and of contraindications for LT certainly reflects the improved evaluation and medical management of the patients in the recent cohort. Deaths while waiting for a liver graft significantly dropped, from 8.5% in the 1986-96 cohort to 2.1% in the 1997-2002 cohort (P = .02). This improvement is related to the development of innovative techniques that “generate” pediatric liver grafts: split livers accounted for 14.4% and 33.4% and living donations for 6.2% and 14.6% of all first LTs in cohorts A and B, respectively.

Pediatric liver transplantation was performed in nine centers in the first cohort and seven in the second. In the 1986-96 cohort, 35% of the patients were transplanted in centers with limited caseloads (≤ 5 pediatric LTs/year). In such centers the results were clearly suboptimal, with 4-year survival after LT of 62% when for the international standards at that time it was about 80%.32 In the 1997-2002 cohort, 92% of the patients were transplanted in centers performing more than 5 pediatric LTs/year, with 4-year survival after LT of 89.6%, at the level of the international standards.32 In 1999 a study of the French transplant network (Etablissement Français des Greffes)33 showed that the results of LT were positively correlated with the number of transplants performed by a center. In the United States, Edwards et al. showed that better survival rates after LT were obtained in centers performing more than 20 transplants per year.34 They also showed that when centers with limited caseloads were affiliated with larger centers, high-level results could be obtained in the small centers, pediatric LT programs being one example. A more recent study from the United States confirmed that liver transplant outcomes were better in centers with high caseloads.35

The overall outcome of BA patients closely depends on the success of the Kasai operation. In the first French national study, it was already shown that the better the results of the Kasai operation, the better the overall patient survival.7 The present study confirmed these findings: 4-year overall survival of patients with early success of the Kasai operation (bilirubin ≤ 20 μmol/L) was 97.8% and 100% in the 1986-1996 and 1997-2002 cohorts, respectively. Overall survival of patients with a failed Kasai operation progressed from 65.2% in the first cohort to 79.4% in the second, reflecting better accessibility to LT and progress in pre- and posttransplantation management. Nevertheless, even in the most experienced centers, the need for LT added risks and consequently impaired the prognosis. A successful Kasai operation still provides children with the best chance of survival, and therefore every effort should be made to improve the results of the Kasai operation.

Thirty of 472 (6.4%) and 16 of 271 (5.9%) patients were referred for de novo LT in the 1986-96 and 1997-2002 cohorts, respectively, of whom 20 and 15, respectively, were transplanted. Four-year overall patient survival progressed from 53.3% to 93.3%. In the first cohort, the results in patients oriented to de novo LT were clearly worse than those of patients who underwent the sequential treatment. In the second cohort, overall survival of the patients was similar for both strategies (4-year overall patient survival 86.6% versus 93.3%, not a significant difference). In a recent study of 755 BA patients listed for LT in the United States between 1995 and 2003, patient survival from the time of listing was 91% at 6 months, 89% at 1 year, and 86% at 3 years (including patients still on the waiting list).36 Therefore, because of the persistent shortage of liver grafts and the ignorance of the very long-term effects of immunosuppression and outcome in late adulthood of transplanted patients, the sequential treatment by Kasai operation ± secondary LT remains the standard of treatment for BA.2

A comparison of the results of the decentralized management of BA patients in France (1997-2002 cohort) with the centralized policy in England and Wales (1999-2002 cohort)25 showed that 4-year overall patient survival was similar in both countries: 87.1% (82.6%-91.6%) versus 89% (82%-94%), respectively. The results of the Kasai operation were worse in France than in England and Wales, with 4-year survival with native liver after the Kasai operation of 42.7% (35.6%-49.8%) versus 51% (42%-59%). Although this difference did not reach statistical significance, it led to a higher need for LTs in infancy and early childhood in France than in England and Wales, as the percentage of transplanted patients among patients alive at age 4 years was 51% and 43%, respectively. This was a direct consequence of the infraoptimal results of the Kasai operation in the centers with very limited caseloads (≤2/year) in France and raises the question of whether the current decentralized management of BA patients should be continued or replaced by more centralized planning of care like that across the Channel.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

We thank the pediatricians and surgeons of 45 participating centers—Amiens: Lenaerts C, Canarelli JP; Angers: Ginies JL, Coupris L; Besançon: Plouvier E, Aubert D; Bondy: Gaudelus J, Grapin C; Bordeaux: Lamireau T, Vergnes P; Brest: de Parscau L, Fenoll B; Caen: Duhamel JF, Petit T; Clermont-Ferrand: Meyer M, Scheye T; Colmar: Meyer Gast D, Geiss S; Dijon: Huet F, Sapin E; Grenoble: Chouraqui JP, Dyon JF; Le Havre: Le Luyer B, Menard M; Lille: Gottrand F, Bonnevalle m, Sfeir R, Besson R; Limoges: De Lumley L, Alain JL; Lyon Debrousse: Guibaud P, Mouriquand P; Lyon Edouard Herriot: Lachaux A, Dubois R, Rivet C, Boillot O; Marseille Nord: Alessandrini P; Marseille Timone: Sarles J, Roquelaure B, Delarue A; Metz: Pierre E, Marchal C; Montpellier: Fournier-Favre S, Galifer RB; Nancy: Morali A, Schmitt M; Nantes: Héloury Y; Nice l'Archet: Triolo V, Kurzenne JY; Nice Lenval: Descos B, Valla JS; Paris Béclère-Paul Brousse: Labrune P, Bismuth H; Paris Bicêtre-Cochin: Bernard O, Debray D, Habes D, Martelli H, Branchereau S, Devictor D, Gauthier F, Valayer J, Dousset B, Soubrane O, Houssin D; Paris Robert Debré: Amédée-Manesme O, Aigrain Y; Paris Necker enfants malades: Lacaille F, Revillon Y; Paris Saint Vincent de Paul: Dupont C, Bargy F; Paris Trousseau: Girardet JP, Hélardot P; Pau: Choulaud JJ, Gauriau L; Poitiers: Cardona J, Levard G; Reims: Digeon B, Poli-Merol ML; Rennes: Dabadie A, Fremond B; Rouen: Mouterde O, Bachy B; Saint Etienne: Lavocat MP, Varlet F; Strasbourg: Becmeur F, Fischbach M, Wolf P; Toulouse: Broué P, Vaysse P; Tours: Maurage C, Robert M; Guadeloupe: Gerry F; Guyanne: Dr Bertsch, Dr Delattre; Martinique: Colombani JF; Nouvelle Calédonie: Ménager C; Polynésie: Gestas P, Pasche J, Besnard M, Papouin Rauzy M; Réunion: Graber D, Lesure JF, de Napoli-Cocci S, Renouil M.

We also thank Dr Michael Chilcott for revising the English text.

References

  1. Top of page
  2. Abstract
  3. Patients and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References