Ischemia Time and Liver Transplantation, Today U. Maggia,*, G. Fornonia, L. Centonzea, E. Meladaa, G. Contea, and G. Rossib a
U.O. Chirurgia Generale e Trapianti di Fegato, Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico Milano, Milan, Italy; and General Surgery and Liver Transplantation Unit, Department of Pathophysiology and Transplantation, University of Milan and Fondazione IRCCS Ca’ Granda, Ospedale Maggiore Policlinico, Milan, Italy
b
ABSTRACT The aim of our study was to retrospectively evaluate the impact of ischemia time and other clinical factors on the development of liver allograft primary nonfunction (PNF). We enrolled 531 consecutive liver transplantations from 1998 to 2013, identifying 10 PNF (1.9%). PNF was found to be statistically related to 4 different variables: donor age >60 years (P ¼ .01), female donor gender (P ¼ .01), total ischemia time >10 hours (P ¼ .03) and infusion of more than 30 fresh frozen plasma units during surgery (P ¼ .02). The study focused on the clinical impact of total ischemia time. We grouped total ischemia time into 4 groups (Group 1: 7.5 hours; Group 2: between 7.5 and 10 hours; Group 3: between 10 and 12 hours; Group 4: >12 hours) and 2 groups (assigning a cut-off value of 10 hours): both these grouping systems significantly influenced the development of PNF and 1-year graft survival, with limited impact on long-term survival. We split total ischemia time in a “technical time,” “hepatectomy time,” and “warm ischemia time.” Only the first 2 components were found to be statistically related to PNF development with P ¼ .02 and P ¼ .003, respectively. Further studies should focus on these aspects of PNF.
I
SCHEMIA time has always been one of the most complex variables in liver transplantation (LT). Such a parameter is directly related to many aspects of the transplant procedure as a whole: the use of different perfusion solutions could affect variable degrees of graft preservation; achieving longer “safe” ischemia time could change the management of liver transplantation, transforming an urgent procedure in an elective one; the ischemia time itself is directly influenced by the geographical distance of the retrieval, and could affect many outcomes of the transplant procedure as well, such as the development of primary nonfunction (PNF), early graft survival and long-term survival. In recent years, donor selection criteria have been extended, new perfusion solutions for organ preservation have been introduced, and the impact of such changes on ischemia time is still under investigation. It is generally accepted that very long ischemia times are detrimental for grafts, but clear evidence of the influence of ischemia times 30%
Grafts (whole/partial) Perfusion sol. (UW, Celsior, HTK) CE (n packs) FFP (n packs) Platelets (n packs) Hepatectomy time (recipient) (min)
Recipient
Age (y) Gender (M/F) Disease (hepatocellular, cholestatic, tumoral, RETX) Blood group (O, A, B, AB) Virus (yes/no) HBV/HCV/HBVþHCV Portal thrombosis (yes/no) Creatinine at Tx (mg/dL) Bilirubin at Tx (mg/dL) INR at Tx MELD at Tx GGT at Tx (IU/L) AST (IU/L) at Tx (IU/L) ALT (UI/L) at Tx (IU/L) Albumine (g/dL) at Tx Platelets (n*1000/mm3) at Tx UNOS Status (1, 2a, 2b, 3) RETX (previous LT) (yes/no)
y, years; m, male; f, female; ICU, intensive care unit; GGT, gamma-glutamyltransferase; Na, sodium; mEq, milliequivalents; BMI, body mass index; min, minutes; UW, University of Wisconsin Solution; HTK, Histidine-tryptophan-ketoglutarate Solution; CE, concentrated erythrocytes; FFP, fresh frozen plasma; RETX, retransplantation; HCV, hepatitis C virus; HBV, hepatitis B virus; Tx, transplantation; INR, International Normalized Ratio; MELD, Model for End-Stage Liver Disease; AST, aspartate aminotransferase; ALT, alanine aminotransferase; UNOS, United Network for Organ Sharing.
When needed, continuous data were shifted to categorical based on receiver operating characteristic (ROC) curves. The ROC analysis of CIT related to PNF failure is presented in Fig 1. The ROC analysis shows a relation between the occurrence of PNF and cold ischemia times. The AUC is 0.717 (P ¼ .018; CI ¼ 0.515e0.919), and the coordinates of the curve allow determination of the optimal cutoff for both specificity and sensitivity rate: 582.5 minutes (spec ¼ 0.714, sens ¼ 0.700). For clinical reasons we have chosen 600 minutes (10 hours) as the optimal cutoff for the analysis. The ROC analysis of CIT >10 hours
vs 10 Cold ischemia (min) 500 >500 Warm ischemia (min) 50 >50 Hepatectomy time (hr) 4 60 years (P ¼ .01), female donor gender (P ¼ .01), total ischemia time >10 hours (P ¼ .03), and infusion of >30 fresh frozen plasma units during transplant procedure (P ¼ .02). We further subdivided total ischemia times into 4 groups (Group 1: 7.5 hours; Group 2: between 7.5 and 10 hours; Group 3: between 10 and 12 hours; Group 4: >12 hours) and analyzed their correlation with PNF. Rates of PNF were 1.3% (2/154), 0.7% (2/272), 5.7% (5/87), and 5.6% (1/18), respectively. There was no statistically significant difference between PNF ratio among groups 1 and 2 (P ¼ .562) or 3 and 4 (P ¼ .975), whereas significant differences were highlighted among all the 4 groups as a whole (P ¼ .014), among the first 3 groups (P ¼ .007), among Group 1 and Group 3 (P ¼ .048), and among Group 2 and Group 3 (P ¼ .003). Table 4. Logistic Regression Model for Primary Nonfunction After Transplantation Variable
Donor Age >60 years ICU days >3 Na >150 mEq/L Gender (F) Intervention Total ischemia time >10 hours Hepatectomy time >4 hours Recipient Viral disease FFP >30 packs
Regression Coefficient
P Value
Confidence Interval (95%)
9.264 4.089 4.023 11.675
.0113 .0755 .0823 .0150
1.655e51.839 0.865e19.321 0.837e19.342 1.613e84.494
5.707 1.786
.0311 1.171e27.808 .4833 0.353e9.047
6.459 6.025
.1051 0.677e61.644 .0282 1.212e29.961
2298
We split total ischemia time in its 3 main components, looking for a possible relationship of these independent time lapses and PNF development. We labeled (a) the time from cross-clamping in the donor to the start of the hepatectomy in the recipient as “technical time”; (b) the time of hepatectomy in the recipient as “hepatectomy time”; and (c) the time of venous anastomoses in the recipient as “warm ischemia time.” A multivariate analysis was performed and only (a) “technical time” and (b) “hepatectomy time” retained a statistically significant correlation with PNF, with P ¼ .02 and P ¼ .003, respectively. One-year graft survival was 82.9% for Group 1, 85.7% for Group 2, 68.8% for Group 3, and 47.1% for Group 4. Statistical differences were similar to those with PNF, with a P ¼ .001 between Groups 2 and 3. Long-term graft survival curves obtained for 4 groups of total ischemia times are reported in Fig 3: statistical differences among the groups persist (P ¼ .0001), but statistical significance among the first 3 groups (P ¼ .077) and between Groups 2 and 3 (P ¼ .05) is lower. EAD occurred in 52 grafts (24.1%) among 213 transplanted during the last 5 years (2008e2012). Only 1 graft with a total ischemia time >12 hours was identified among those 213 patients, so it was excluded from the analysis. EAD occurred in 7/ 50 (14%) of grafts in Group 1 and in 45/162 (27.7%) of Groups 2 and 3 (P ¼ .048). No statistical relation was found between EAD and 6-month graft survival (P ¼ .175). DISCUSSION
Shifting LT from an urgent procedure to an elective or semielective one has always been an attractive target. Even though improvements have been made in different areas such as surgical technique, immunosuppression, selection of recipients, ischemia time, and liver preservation have always been a main issue in LT. In the history of LT, Euro-Collins (EC) solution [3] was soon replaced by University of Wisconsin (UW) solution [4]: early results [5] on average preservation times comparing UW-preserved livers and EC-preserved livers highlighted UW group preservation time to be twice as long as that of EC group, showed less PNF in the UW group, and a higher survival rate for UW-preserved grafts, though without a statistical difference; the mean cold ischemia time was 10.1 5.0 (SD) hours for the UW-preserved livers, and the mean donor age was 36.5 20.6 years [5]. The influence of ischemia time on PNF and liver enzymes using UW solution was further investigated by other authors [6], showing no PNF in any of the 2 groups of UW-preserved livers with >10 hours or 12 hours [7].
MAGGI, FORNONI, CENTONZE ET AL
In a review from 1999, the medically acceptable cold ischemia time for the liver was limited to 12 hours [8]. The increasing number of so-called extended criteria grafts was probably the main driver that changed previous data and results of initial times of cold-storage preservation. Focusing on donor age in the OPTN database [9], it is noteworthy how, from 1988 to 2013, the percentage of patients aged 50e64 years changed from 2.4% to 26% and those aged >65 years changed from 0.1% to 7.5%, but those aged 18e34 years dropped from 42.3 to 29.6%. Probably because of extended-criteria grafts, recent reports have suggested that the actual risk for PNF and EAD could be further stratified even for short ischemia times (ie,
U.O. Chirurgia Generale e Trapianti di Fegato, Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico Milano, Milan, Italy; and General Surgery and Liver Transplantation Unit, Department of Pathophysiology and Transplantation, University of Milan and Fondazione IRCCS Ca’ Granda, Ospedale Maggiore Policlinico, Milan, Italy
b
ABSTRACT The aim of our study was to retrospectively evaluate the impact of ischemia time and other clinical factors on the development of liver allograft primary nonfunction (PNF). We enrolled 531 consecutive liver transplantations from 1998 to 2013, identifying 10 PNF (1.9%). PNF was found to be statistically related to 4 different variables: donor age >60 years (P ¼ .01), female donor gender (P ¼ .01), total ischemia time >10 hours (P ¼ .03) and infusion of more than 30 fresh frozen plasma units during surgery (P ¼ .02). The study focused on the clinical impact of total ischemia time. We grouped total ischemia time into 4 groups (Group 1: 7.5 hours; Group 2: between 7.5 and 10 hours; Group 3: between 10 and 12 hours; Group 4: >12 hours) and 2 groups (assigning a cut-off value of 10 hours): both these grouping systems significantly influenced the development of PNF and 1-year graft survival, with limited impact on long-term survival. We split total ischemia time in a “technical time,” “hepatectomy time,” and “warm ischemia time.” Only the first 2 components were found to be statistically related to PNF development with P ¼ .02 and P ¼ .003, respectively. Further studies should focus on these aspects of PNF.
I
SCHEMIA time has always been one of the most complex variables in liver transplantation (LT). Such a parameter is directly related to many aspects of the transplant procedure as a whole: the use of different perfusion solutions could affect variable degrees of graft preservation; achieving longer “safe” ischemia time could change the management of liver transplantation, transforming an urgent procedure in an elective one; the ischemia time itself is directly influenced by the geographical distance of the retrieval, and could affect many outcomes of the transplant procedure as well, such as the development of primary nonfunction (PNF), early graft survival and long-term survival. In recent years, donor selection criteria have been extended, new perfusion solutions for organ preservation have been introduced, and the impact of such changes on ischemia time is still under investigation. It is generally accepted that very long ischemia times are detrimental for grafts, but clear evidence of the influence of ischemia times 30%
Grafts (whole/partial) Perfusion sol. (UW, Celsior, HTK) CE (n packs) FFP (n packs) Platelets (n packs) Hepatectomy time (recipient) (min)
Recipient
Age (y) Gender (M/F) Disease (hepatocellular, cholestatic, tumoral, RETX) Blood group (O, A, B, AB) Virus (yes/no) HBV/HCV/HBVþHCV Portal thrombosis (yes/no) Creatinine at Tx (mg/dL) Bilirubin at Tx (mg/dL) INR at Tx MELD at Tx GGT at Tx (IU/L) AST (IU/L) at Tx (IU/L) ALT (UI/L) at Tx (IU/L) Albumine (g/dL) at Tx Platelets (n*1000/mm3) at Tx UNOS Status (1, 2a, 2b, 3) RETX (previous LT) (yes/no)
y, years; m, male; f, female; ICU, intensive care unit; GGT, gamma-glutamyltransferase; Na, sodium; mEq, milliequivalents; BMI, body mass index; min, minutes; UW, University of Wisconsin Solution; HTK, Histidine-tryptophan-ketoglutarate Solution; CE, concentrated erythrocytes; FFP, fresh frozen plasma; RETX, retransplantation; HCV, hepatitis C virus; HBV, hepatitis B virus; Tx, transplantation; INR, International Normalized Ratio; MELD, Model for End-Stage Liver Disease; AST, aspartate aminotransferase; ALT, alanine aminotransferase; UNOS, United Network for Organ Sharing.
When needed, continuous data were shifted to categorical based on receiver operating characteristic (ROC) curves. The ROC analysis of CIT related to PNF failure is presented in Fig 1. The ROC analysis shows a relation between the occurrence of PNF and cold ischemia times. The AUC is 0.717 (P ¼ .018; CI ¼ 0.515e0.919), and the coordinates of the curve allow determination of the optimal cutoff for both specificity and sensitivity rate: 582.5 minutes (spec ¼ 0.714, sens ¼ 0.700). For clinical reasons we have chosen 600 minutes (10 hours) as the optimal cutoff for the analysis. The ROC analysis of CIT >10 hours
vs 10 Cold ischemia (min) 500 >500 Warm ischemia (min) 50 >50 Hepatectomy time (hr) 4 60 years (P ¼ .01), female donor gender (P ¼ .01), total ischemia time >10 hours (P ¼ .03), and infusion of >30 fresh frozen plasma units during transplant procedure (P ¼ .02). We further subdivided total ischemia times into 4 groups (Group 1: 7.5 hours; Group 2: between 7.5 and 10 hours; Group 3: between 10 and 12 hours; Group 4: >12 hours) and analyzed their correlation with PNF. Rates of PNF were 1.3% (2/154), 0.7% (2/272), 5.7% (5/87), and 5.6% (1/18), respectively. There was no statistically significant difference between PNF ratio among groups 1 and 2 (P ¼ .562) or 3 and 4 (P ¼ .975), whereas significant differences were highlighted among all the 4 groups as a whole (P ¼ .014), among the first 3 groups (P ¼ .007), among Group 1 and Group 3 (P ¼ .048), and among Group 2 and Group 3 (P ¼ .003). Table 4. Logistic Regression Model for Primary Nonfunction After Transplantation Variable
Donor Age >60 years ICU days >3 Na >150 mEq/L Gender (F) Intervention Total ischemia time >10 hours Hepatectomy time >4 hours Recipient Viral disease FFP >30 packs
Regression Coefficient
P Value
Confidence Interval (95%)
9.264 4.089 4.023 11.675
.0113 .0755 .0823 .0150
1.655e51.839 0.865e19.321 0.837e19.342 1.613e84.494
5.707 1.786
.0311 1.171e27.808 .4833 0.353e9.047
6.459 6.025
.1051 0.677e61.644 .0282 1.212e29.961
2298
We split total ischemia time in its 3 main components, looking for a possible relationship of these independent time lapses and PNF development. We labeled (a) the time from cross-clamping in the donor to the start of the hepatectomy in the recipient as “technical time”; (b) the time of hepatectomy in the recipient as “hepatectomy time”; and (c) the time of venous anastomoses in the recipient as “warm ischemia time.” A multivariate analysis was performed and only (a) “technical time” and (b) “hepatectomy time” retained a statistically significant correlation with PNF, with P ¼ .02 and P ¼ .003, respectively. One-year graft survival was 82.9% for Group 1, 85.7% for Group 2, 68.8% for Group 3, and 47.1% for Group 4. Statistical differences were similar to those with PNF, with a P ¼ .001 between Groups 2 and 3. Long-term graft survival curves obtained for 4 groups of total ischemia times are reported in Fig 3: statistical differences among the groups persist (P ¼ .0001), but statistical significance among the first 3 groups (P ¼ .077) and between Groups 2 and 3 (P ¼ .05) is lower. EAD occurred in 52 grafts (24.1%) among 213 transplanted during the last 5 years (2008e2012). Only 1 graft with a total ischemia time >12 hours was identified among those 213 patients, so it was excluded from the analysis. EAD occurred in 7/ 50 (14%) of grafts in Group 1 and in 45/162 (27.7%) of Groups 2 and 3 (P ¼ .048). No statistical relation was found between EAD and 6-month graft survival (P ¼ .175). DISCUSSION
Shifting LT from an urgent procedure to an elective or semielective one has always been an attractive target. Even though improvements have been made in different areas such as surgical technique, immunosuppression, selection of recipients, ischemia time, and liver preservation have always been a main issue in LT. In the history of LT, Euro-Collins (EC) solution [3] was soon replaced by University of Wisconsin (UW) solution [4]: early results [5] on average preservation times comparing UW-preserved livers and EC-preserved livers highlighted UW group preservation time to be twice as long as that of EC group, showed less PNF in the UW group, and a higher survival rate for UW-preserved grafts, though without a statistical difference; the mean cold ischemia time was 10.1 5.0 (SD) hours for the UW-preserved livers, and the mean donor age was 36.5 20.6 years [5]. The influence of ischemia time on PNF and liver enzymes using UW solution was further investigated by other authors [6], showing no PNF in any of the 2 groups of UW-preserved livers with >10 hours or 12 hours [7].
MAGGI, FORNONI, CENTONZE ET AL
In a review from 1999, the medically acceptable cold ischemia time for the liver was limited to 12 hours [8]. The increasing number of so-called extended criteria grafts was probably the main driver that changed previous data and results of initial times of cold-storage preservation. Focusing on donor age in the OPTN database [9], it is noteworthy how, from 1988 to 2013, the percentage of patients aged 50e64 years changed from 2.4% to 26% and those aged >65 years changed from 0.1% to 7.5%, but those aged 18e34 years dropped from 42.3 to 29.6%. Probably because of extended-criteria grafts, recent reports have suggested that the actual risk for PNF and EAD could be further stratified even for short ischemia times (ie,
