Pediatr Transplantation 2014
© 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Pediatric Transplantation DOI: 10.1111/petr.12363
Risk factors for early and late biliary complications in pediatric liver transplantation L€ uthold SC, Kaseje N, Jannot A-S, Mentha G, Majno P, Toso C, Belli DC, McLin VA, Wildhaber BE. (2014) Risk factors for early and late biliary complications in pediatric liver transplantation. Pediatr Transplant, 00: 1–9. DOI: 10.1111/petr.12363. Abstract: BC are a common source of morbidity after pediatric LT. Knowledge about risk factors may help to reduce their incidence. Retrospective analysis of BC in 116 pediatric patients (123 LT) (single institution, 05/1990–12/2011, medium follow-up 7.9 yr). One-, five-, and 10-yr survival was 91.1%, no patient died of BC. Prevalence and risk factors for anastomotic and intrahepatic BC were examined. There were 29 BC in 123 LT (23.6%), with three main categories: 10 (8.1%) primary anastomotic strictures, eight (6.5%) anastomotic leaks, and three (2.4%) intrahepatic strictures. Significant risk factors for anastomotic leaks were total operation time (increase 1.26-fold) and early HAT (48 yr (increase 1.09-fold) and MELD score >30 (increase 1.2-fold). To avoid morbidity from anastomotic BC in pediatric LT, the preferred biliary anastomosis appears to be biliary-enteric. Operation time should be kept to a minimum, and HAT must by all means be prevented. Children with a high MELD score or receiving livers from older donors are at increased risk for intrahepatic strictures.
Pediatric LT is currently both the standard therapy for end-stage liver disease in children and an attractive therapeutic option for many metabolic diseases (1). BC have been long said to be the Achilles’ heel of LT (2). Until today, BC remain a frequent cause of morbidity, leading to longterm and repeated treatments, including percutaneous and surgical procedures (3). The incidence of BC in pediatric LT has been reported to fall in the range of 6–32%, the most frequent complications being anastomotic bile duct strictures and leaks (4–7). In adult LT-series, HAT, donor age, and prolonged cold ischemia time are among the
Abbreviations: BC, biliary complications; BE, biliaryenteric anastomosis; BMI, body mass index; CMV, cytomegalovirus; DD, duct-to-duct choledochal anastomosis; EBV, Epstein-Barr virus; HAT, hepatic artery thrombosis; HLA, Human leucocyte antigen; LT, liver transplantation; MELD, Model of end stage liver disease.
Samuel C. L€uthold1, Neema Kaseje1, Anne-Sophie Jannot2, Gilles Mentha3†, Pietro Majno3, Christian Toso3, Dominique C. Belli4, Valerie A. McLin4 and Barbara E. Wildhaber1 1
Division of Pediatric Surgery, Geneva University Hospitals, Geneva, Switzerland, 2Clinical Research, Geneva University Hospitals, Geneva, Switzerland, 3 Divsion of Transplantation, Geneva University Hospitals, Geneva, Switzerland, 4Pediatric Gastroenterology Unit, Geneva University Hospitals, Geneva, Switzerland
Key words: pediatric liver transplantation – biliary complications – risk factors Barbara E. Wildhaber, H^opitaux Universitaires de Geneve, Rue Willy-Donze 6, 1211 Geneve 14, Switzerland Tel.: +41 22 382 46 62 Fax: +41 22 382 46 66 E-mail: [email protected] †
Professor Mentha passed away in May 2014.
Accepted for publication 14 August 2014
most frequent risk factors for the development of BC (8–10). There are few studies in children examining risk factors for BC. Therefore, in the present study, we analyzed the incidence of BC after LT in our pediatric cohort and aimed to identify potential risk factors, based on those reported in adult recipients. Patients and methods One hundred and sixteen consecutive pediatric patients (age zero to 16 yr at time of LT), who underwent 123 LT from May 1990 to December 2011 at the University Children’s Hospital Geneva, were included in this retrospective study. Children transplanted at another institution and followed up in Geneva were excluded from the analysis. Ethical approval for this study was obtained from the institutional ethics committee (CER 11-01OR/MATPED 11-004R). Follow-up ended on September 30, 2012, or earlier if the patient died or was lost to follow-up (n = 1). Since the use of the MELD score was not implemented until 2006, calculated MELD scores were obtained
1
L€ uthold et al. retrospectively using laboratory values recorded immediately prior to LT (11). Post-operative patient care varied slightly over time, but generally followed the same basic principles. Prior to 2000, immunosuppression consisted of ciclosporine, azathioprine, and prednisone. Thereafter, the immunosuppression protocol was tacrolimus based, and except for a few cases, in combination with steroids and basiliximab. Before 2000, most left lateral segments were reduced livers (ex situ reduction); after the year 2000, they were mostly derived from in situ splitting. The LT operative procedure was similar throughout the study period, and there was very little turnover among the surgical team. For duct-to-duct choledochal anastomosis, two precautions were taken: the dissection of the remaining bile duct in the recipient was minimal in order to preserve adequate blood supply, and the gastroduodenal artery was not divided. For both duct-to-duct anastomoses and Roux-en-Y hepaticojejunostomy (i.e., biliary-enteric anastomosis), one of two approaches was used: either circumferential interrupted sutures or continuous sutures for the posterior wall and interrupted sutures for the anterior wall. Suture material for the biliary anastomoses was 7-0 or 6-0 MaxonTM polyglyconate synthetic absorbable monofilament, or PDSTM polydioxanone synthetic absorbable monofilament. Nonabsorbable sutures, such as ProleneTM, were not used. There was no internal or external biliary drainage or use of stents in either type of anastomosis. The following variables were extracted from patient records: (i) recipient characteristics: patient age and weight, MELD score, weight, and BMI; (ii) donor characteristics: donor age, donor BMI, peak donor serum sodium; (iii) medical parameters: CMV mismatch between donor and recipient (defined as positive donor and negative recipient serology), EBV mismatch (defined as with CMV), HLA mismatch, ABO compatibility; and (iv) surgical parameters: type of liver graft (deceased vs. living donor, split vs. reduced vs. whole liver), cold ischemia time (defined as the time from when the organ is cooled to 4 °C during organ procurement to the time of placement in the recipient abdomen), warm ischemia time (defined as the time from when the organ is placed in the recipient abdomen to time of venous reperfusion), total ischemia time, total operation time (defined as time from incision to skin closure), type of suture material (see above), occurrence of early HAT (within 30 days following LT), and type of biliary anastomosis (duct-to-duct vs. entero-biliary). BC were diagnosed clinically, biochemically (increase of bilirubin and/or gamma-glutamyl transpeptidase above normal levels), and/or radiologically, most often by ultrasound, abnormal visibility of bile ducts leading to cholangio-MRI and in any doubt to cholangiography. In case of the presence of pathologic liver function tests, organ rejection was excluded by liver biopsy, and if histological signs of cholangiopathy were observed, further imaging was performed (cholangio-MRI or direct cholangiography). Secondary anastomotic strictures after a primary anastomotic leak were excluded from the analysis of risk factors for anastomotic strictures. Intrahepatic biliary strictures were analyzed separately from anastomotic strictures, as they are recognized to have a different pathophysiology (10). We focused our risk factors analysis on the most frequent BC (i.e., anastomotic and intrahepatic strictures and anastomotic leak), in order to avoid confounding bias regarding the different types of BC. Thus, rapidly resolving, small bile collections due to cut surface leaks in reduced/split-liver LT and following percutaneous drainage were counted as BC but were not included in the risk factor analysis.
2
Statistical analysis The Cox proportional hazard survival model was used to analyze the relationship between anastomotic strictures and leaks and between patient and surgical characteristics. Associations between intrahepatic strictures and patient and surgery characteristics were assessed using a logistic regression, owing to the small numbers of events. Comparison between groups of the time at BC diagnosis was performed using Mann–Whitney–Wilcoxon test. A significant p-value was defined as
© 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
Pediatric Transplantation DOI: 10.1111/petr.12363
Risk factors for early and late biliary complications in pediatric liver transplantation L€ uthold SC, Kaseje N, Jannot A-S, Mentha G, Majno P, Toso C, Belli DC, McLin VA, Wildhaber BE. (2014) Risk factors for early and late biliary complications in pediatric liver transplantation. Pediatr Transplant, 00: 1–9. DOI: 10.1111/petr.12363. Abstract: BC are a common source of morbidity after pediatric LT. Knowledge about risk factors may help to reduce their incidence. Retrospective analysis of BC in 116 pediatric patients (123 LT) (single institution, 05/1990–12/2011, medium follow-up 7.9 yr). One-, five-, and 10-yr survival was 91.1%, no patient died of BC. Prevalence and risk factors for anastomotic and intrahepatic BC were examined. There were 29 BC in 123 LT (23.6%), with three main categories: 10 (8.1%) primary anastomotic strictures, eight (6.5%) anastomotic leaks, and three (2.4%) intrahepatic strictures. Significant risk factors for anastomotic leaks were total operation time (increase 1.26-fold) and early HAT (48 yr (increase 1.09-fold) and MELD score >30 (increase 1.2-fold). To avoid morbidity from anastomotic BC in pediatric LT, the preferred biliary anastomosis appears to be biliary-enteric. Operation time should be kept to a minimum, and HAT must by all means be prevented. Children with a high MELD score or receiving livers from older donors are at increased risk for intrahepatic strictures.
Pediatric LT is currently both the standard therapy for end-stage liver disease in children and an attractive therapeutic option for many metabolic diseases (1). BC have been long said to be the Achilles’ heel of LT (2). Until today, BC remain a frequent cause of morbidity, leading to longterm and repeated treatments, including percutaneous and surgical procedures (3). The incidence of BC in pediatric LT has been reported to fall in the range of 6–32%, the most frequent complications being anastomotic bile duct strictures and leaks (4–7). In adult LT-series, HAT, donor age, and prolonged cold ischemia time are among the
Abbreviations: BC, biliary complications; BE, biliaryenteric anastomosis; BMI, body mass index; CMV, cytomegalovirus; DD, duct-to-duct choledochal anastomosis; EBV, Epstein-Barr virus; HAT, hepatic artery thrombosis; HLA, Human leucocyte antigen; LT, liver transplantation; MELD, Model of end stage liver disease.
Samuel C. L€uthold1, Neema Kaseje1, Anne-Sophie Jannot2, Gilles Mentha3†, Pietro Majno3, Christian Toso3, Dominique C. Belli4, Valerie A. McLin4 and Barbara E. Wildhaber1 1
Division of Pediatric Surgery, Geneva University Hospitals, Geneva, Switzerland, 2Clinical Research, Geneva University Hospitals, Geneva, Switzerland, 3 Divsion of Transplantation, Geneva University Hospitals, Geneva, Switzerland, 4Pediatric Gastroenterology Unit, Geneva University Hospitals, Geneva, Switzerland
Key words: pediatric liver transplantation – biliary complications – risk factors Barbara E. Wildhaber, H^opitaux Universitaires de Geneve, Rue Willy-Donze 6, 1211 Geneve 14, Switzerland Tel.: +41 22 382 46 62 Fax: +41 22 382 46 66 E-mail: [email protected] †
Professor Mentha passed away in May 2014.
Accepted for publication 14 August 2014
most frequent risk factors for the development of BC (8–10). There are few studies in children examining risk factors for BC. Therefore, in the present study, we analyzed the incidence of BC after LT in our pediatric cohort and aimed to identify potential risk factors, based on those reported in adult recipients. Patients and methods One hundred and sixteen consecutive pediatric patients (age zero to 16 yr at time of LT), who underwent 123 LT from May 1990 to December 2011 at the University Children’s Hospital Geneva, were included in this retrospective study. Children transplanted at another institution and followed up in Geneva were excluded from the analysis. Ethical approval for this study was obtained from the institutional ethics committee (CER 11-01OR/MATPED 11-004R). Follow-up ended on September 30, 2012, or earlier if the patient died or was lost to follow-up (n = 1). Since the use of the MELD score was not implemented until 2006, calculated MELD scores were obtained
1
L€ uthold et al. retrospectively using laboratory values recorded immediately prior to LT (11). Post-operative patient care varied slightly over time, but generally followed the same basic principles. Prior to 2000, immunosuppression consisted of ciclosporine, azathioprine, and prednisone. Thereafter, the immunosuppression protocol was tacrolimus based, and except for a few cases, in combination with steroids and basiliximab. Before 2000, most left lateral segments were reduced livers (ex situ reduction); after the year 2000, they were mostly derived from in situ splitting. The LT operative procedure was similar throughout the study period, and there was very little turnover among the surgical team. For duct-to-duct choledochal anastomosis, two precautions were taken: the dissection of the remaining bile duct in the recipient was minimal in order to preserve adequate blood supply, and the gastroduodenal artery was not divided. For both duct-to-duct anastomoses and Roux-en-Y hepaticojejunostomy (i.e., biliary-enteric anastomosis), one of two approaches was used: either circumferential interrupted sutures or continuous sutures for the posterior wall and interrupted sutures for the anterior wall. Suture material for the biliary anastomoses was 7-0 or 6-0 MaxonTM polyglyconate synthetic absorbable monofilament, or PDSTM polydioxanone synthetic absorbable monofilament. Nonabsorbable sutures, such as ProleneTM, were not used. There was no internal or external biliary drainage or use of stents in either type of anastomosis. The following variables were extracted from patient records: (i) recipient characteristics: patient age and weight, MELD score, weight, and BMI; (ii) donor characteristics: donor age, donor BMI, peak donor serum sodium; (iii) medical parameters: CMV mismatch between donor and recipient (defined as positive donor and negative recipient serology), EBV mismatch (defined as with CMV), HLA mismatch, ABO compatibility; and (iv) surgical parameters: type of liver graft (deceased vs. living donor, split vs. reduced vs. whole liver), cold ischemia time (defined as the time from when the organ is cooled to 4 °C during organ procurement to the time of placement in the recipient abdomen), warm ischemia time (defined as the time from when the organ is placed in the recipient abdomen to time of venous reperfusion), total ischemia time, total operation time (defined as time from incision to skin closure), type of suture material (see above), occurrence of early HAT (within 30 days following LT), and type of biliary anastomosis (duct-to-duct vs. entero-biliary). BC were diagnosed clinically, biochemically (increase of bilirubin and/or gamma-glutamyl transpeptidase above normal levels), and/or radiologically, most often by ultrasound, abnormal visibility of bile ducts leading to cholangio-MRI and in any doubt to cholangiography. In case of the presence of pathologic liver function tests, organ rejection was excluded by liver biopsy, and if histological signs of cholangiopathy were observed, further imaging was performed (cholangio-MRI or direct cholangiography). Secondary anastomotic strictures after a primary anastomotic leak were excluded from the analysis of risk factors for anastomotic strictures. Intrahepatic biliary strictures were analyzed separately from anastomotic strictures, as they are recognized to have a different pathophysiology (10). We focused our risk factors analysis on the most frequent BC (i.e., anastomotic and intrahepatic strictures and anastomotic leak), in order to avoid confounding bias regarding the different types of BC. Thus, rapidly resolving, small bile collections due to cut surface leaks in reduced/split-liver LT and following percutaneous drainage were counted as BC but were not included in the risk factor analysis.
2
Statistical analysis The Cox proportional hazard survival model was used to analyze the relationship between anastomotic strictures and leaks and between patient and surgical characteristics. Associations between intrahepatic strictures and patient and surgery characteristics were assessed using a logistic regression, owing to the small numbers of events. Comparison between groups of the time at BC diagnosis was performed using Mann–Whitney–Wilcoxon test. A significant p-value was defined as
