SEARCH

SEARCH BY CITATION

Keywords:

  • DCD;
  • graft inflow modulation;
  • hepatic artery thrombosis;
  • LDLT;
  • liver flows;
  • liver transplantation;
  • systemic and hepatic hemodynamics;
  • portal hypertension

Abstract

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

The interaction of systemic hemodynamics with hepatic flows at the time of liver transplantation (LT) has not been studied in a prospective uniform way for different types of grafts. We prospectively evaluated intraoperative hemodynamics of 103 whole and partial LT. Liver graft hemodynamics were measured using the ultrasound transit time method to obtain portal (PVF) and arterial (HAF) hepatic flow. Measurements were recorded on the native liver, the portocaval shunt, following reperfusion and after biliary anastomosis. After LT HAF and PVF do not immediately return to normal values. Increased PVF was observed after graft implantation. Living donor LT showed the highest compliance to portal hyperperfusion. The amount of liver perfusion seemed to be related to the quality of the graft. A positive correlation for HAF, PVF and total hepatic blood flow with cardiac output was found (p = 0.001). Portal hypertension, macrosteatosis >30%, warm ischemia time and cardiac output, independently influence the hepatic flows. These results highlight the role of systemic hemodynamic management in LT to optimize hepatic perfusion, particularly in LDLT and split LT, where the highest flows were registered.


Introduction

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

The systemic hemodynamics in patients undergoing liver transplantation (LT) is abnormal (1–3). As well as a high resistance from the cirrhotic liver, the augmented splanchnic volume increases portal pressure progressively with the development of collateral circulation, diverting splanchnic blood to the systemic circulation. Initial systemic vasodilatation is followed by a decrease in central volume causing relative hypovolemia, which leads to sodium retention and plasma volume expansion, resulting in increased cardiac output (CO) (4–7). High CO together with decreased peripheral vascular resistance and arterial pressure characterize this hyperdynamic circulation, worsening the initial endothelial stress and closing the circuit. LT replaces the cirrhotic liver with a normal liver, relieving the mechanical component of portal hypertension but without immediately restoring the systemic or the splanchnic circulation to normal (8–11). At the same time, the relative central hypovolemia accentuates the influence of vasoactive drug administration and adequate volume management during LT, because tissue hypoperfusion during surgery has been shown to be a cause of poor outcome (2,12,13). The normal liver has no active role in the physiological regulation of hepatic inflow. Hence, the liver is a passive recipient of fluctuating amounts of blood flow, which can encompass a wide range of flow (7,14). Moreover, the newly grafted liver has to comply with the high splanchnic volume of the recipient. The presence of a hepatic arterial buffer response (HABR) operates to compensate for portal vein flow (PVF) changes (15–17). Indeed, the extreme increase of PVF observed after living donor LT (LDLT) together with the HABR are responsible for the reduced hepatic artery flow (HAF) usually encountered in this setting (18,19). Exposition of grafts (mainly partial grafts, but not exclusively) to excessive portal perfusion could determine specific problems such as prolonged cholestasis, ascites and increased vascular thrombosis rates, events characterizing the small-for-size syndrome (SFSS) and potentially leading to graft loss (20–22). To overcome the SFSS and to reduce postreperfusion graft flow imbalance, the concept of graft inflow modulation (GIM) has been proposed with beneficial influence in optimization of HAF and improved outcome in LDLT (23–25).

To our knowledge, no prospective evaluation on direct liver flow measurement in different types and qualities of grafts is currently available. To fill this gap, we designed a prospective protocol for intraoperative hepatic hemodynamic data collection to evaluate systemic and regional hemodynamics of LT with different graft types, measuring intraoperative HAF and PVF. The aim of this report is to explore the influence of hyperdynamic flows in different graft types and assess variables that could potentially influence hepatic hemodynamics.

Patients and Methods

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

Between January 2007 and December 2009, 151 consecutive LTs were performed in 142 adult recipients at the Ghent University Hospital.

Study protocol

After approval by the local IRB, informed consent was obtained from all adult liver transplant candidates when eligibility for transplantation was confirmed. According to this prospective study, all type of indications (acute and chronic end-stage liver diseases), donors (deceased donor, living donor, donation after cardiac death) and type of graft (whole and partial) were included. Patients with preexisting pulmonary or hepato-pulmonary syndrome and patients with cardiac diseases, apart from those with common symptoms of end-stage liver dysfunction (26), were excluded.

Among 142 adult patients, 7 (4.9%) who had pulmonary hypertension were excluded from analysis because of their known increased hemodynamic instability and worse outcome (27,28). Two other patients (1.4%) who underwent sequential aortic valve replacement and LT, as well as 30 patients (21.1%) with incomplete systemic hemodynamic monitoring or hepatic flow measurements were excluded from the study analysis.

All donors after cardiac death (DCD) grafts, as well as those with a combination of known risk factors such as donor age above 70 years old, higher than three times normal transaminases or bilirubin levels, macrosteatosis above 30% or presence of initial portal fibrosis in HCV positive donors (29,30), were considered extended criteria donor (ECD) grafts. Portal hypertension status was assessed prior to LT and defined as presence of TIPSS, esophageal varices grade 2–3 and refractory ascites (31). The presence of these parameters is in most cases compatible with the presence of clinically significant portal hypertension (32–34).

The liver hemodynamic measurements in live donors provide an estimate of normal hepatic inflow in the intraoperative setting. Recordings in 80 out of 97 live liver donors from our historical cohort of a median HAF 20.5 mL/min/100 g LW (range 7–86) and a median PVF of 90 mL/min/100 g LW (range 34–208), were used as reference values. Liver transplant recipients can present or not portal hypertension according to the state of their disease. Likewise, the hyperdynamic status can be established or not. According to these two parameters, four different clinical situations regarding the systemic and regional hemodynamics can be identified: type 1, patients with absence of portal hypertension or systemic hyperdynamic status (e.g. HCC patients with compensated liver function); type 2, portal hypertension alone (e.g. initial stages of portal hypertension); type 3, systemic hyperdynamic status alone (e.g. cirrhotic patients with large collateral vascularization); type 4, presence of both portal hypertension and systemic hyperdynamic status (e.g. advanced liver disease).

Cardiopulmonary monitoring

Anesthetic management was standardized. A radial artery catheter to measure systemic arterial pressure was placed. An 8.0 F pulmonary artery catheter (Swan-Ganz Continuous Cardiac Output Thermodilution Catheter; CCOmbo CCO/SvO2 catheter 777HF8; Edwards Lifesciences, Irvine, CA, USA) was introduced into the right internal jugular vein using an 8.5 F introducer (Edwards Lifesciences) and connected to a Vigilance monitor (Edwards Lifescience) for measurement of stroke volume and continuous CO. All patients were supine and the zero reference was the mid-axillary line.

Intraoperative hepatic flow measurement

At predetermined time points, serial measurements for assessing hepatic hemodynamics were performed in living donors and liver transplant recipients using ultrasound transit-time flow measurement (TTFM) (Medi-Stim AS, Oslo, Norway). This method has been validated, showing a great accuracy, reliability and reproducibility in patients undergoing vascular surgery (35). Readings of HAF and PVF were recorded as follows: on the native liver prior to total hepatectomy (M0), on the temporary portacaval shunt during the anhepatic phase (M1), following portal and arterial reperfusion (M2) and after biliary anastomosis (M3) (Figure 1). The average flow values are expressed in mL/min and were taken under stable clinical conditions. When multiple arteries were present, mean values of the added single measurements were used for analysis.

image

Figure 1. Liver perfusion and timeline of hemodynamics measurement protocol. M0, recipient flows; M1, temporary portocaval flows; M2, flows at reperfusion; M3, second measurement of hepatic flows after reperfusion. Boxes represent interquartile range, whiskers minimum and maximum values.

Download figure to PowerPoint

To avoid misinterpretation of data and possible slant deriving from different sizes and types of grafts, we defined liver perfusion as the amount of flow normalized by liver weight expressed in mL/min/100 g LW. For M0 perfusion calculation, the weight of the hepatectomy specimen was used; for M1–M3 the weight of the liver graft was used. In case of HAF inferior to 100 mL/min without evident arterial kinking, topical application of papaverine was done for 5 min to rule out vasospasm. In case of persisting low flows and to exclude anastomotic flow-limiting factors, a test clamping of the portal vein was performed and GIM considered when a marked improvement of the HAF was observed.

Surgical technique

The caval vein was preserved in all cases. An end-to-side temporary portocaval shunt (PCS) was routinely applied to drain splanchnic blood during total hepatectomy and taken off during engraftment after caval anastomosis completion. All the grafts were implanted using the end-to-side cavo-caval anastomosis technique for all types of grafts (36–38). An end-to-end arterial reconstruction was performed using a branch patch at the level of the recipient gastroduodenal artery in deceased donor LT or at the level of the proper left or right hepatic artery in LDLT. Biliary continuity was restored by end-to-end anastomosis or using a Roux-en-Y loop. Technical variations in case of LDLT and split LT (SLT) have been previously described (24,25,39,40).

Statistical analysis

Matched pairs Wilcoxon was used to contrast differences between registered flows and perfusions (HAF and PVF) at different phases of LT. Comparisons of systemic and hepatic hemodynamics between graft type groups after reperfusion were performed using Student's test, Fisher's test or Kruskal–Wallis when appropriate. Comparisons with live donor perfusions were assessed analogously. Pearson correlation coefficient was calculated to assess correlations between systemic (CO) and regional hemodynamics (HAF, PVF) after reperfusion. Graft survival was estimated according to the Kaplan–Meier method. Forward fitting linear regression models (p value for inclusion 0.1) were used to explore the influence of different factors potentially influencing hepatic flows. Factors assessed at univariate analysis were those reported in Tables 2–5. Factors included in multivariate analyses were those that at univariate analysis showed a significance of at least 0.1 or those deemed clinically relevant. Because of high colinearity, CO and PHT were substituted with the four patient classification types according to systemic hemodynamic status and portal hypertension. The results of this subgroup analysis are reported in Table 9. Statistical analysis was performed using STATA 11 for Windows and GraphPad Prism 5.0c for Macintosh.

Table 2.  Recipient characteristics
 Overall (n = 103)FS (n = 76)LDLT (n = 5)SLT (n = 13)DCD (n = 9)
  1. Results are given in number(percentage) or median(range). No significant differences between groups were found.

  2. FS = full size liver transplantation; LDLT = living donor liver transplantation; SLT = split liver transplantation; DCD = donation after cardiac death; PHT = portal hypertension; SBC = secondary biliary cirrhosis; AHF = acute hepatic failure; MELD = Model for End-stage Liver Disease; Type 1, absence of portal hypertension and systemic hyperdynamic status; Type 2, portal hypertension alone; Type 3, systemic hyperdynamic status alone; Type 4, presence of both portal hypertension and systemic hyperdynamic status.

Gender male73(71%)57(75%)3(60%)6(46%)7(78%)
Age(years)60(17–74)59(22–74)53(26–67)60(17–69)60(55–67)
PHT85(83%)65(86%)1(20%)11(85%)8(89%)
Indication to LT
 Alcohol41(40%)31(41%)04(31%)6(67%)
 Cirrhosis35(34%)25(33%)2(40%)7(54%)1(11%)
 SBC7(7%)6(8%)01(8%)0
 AHF7(7%)6(8%)01(8%)0
 Cholestatic4(4%)2(3%)1(20%)01(11%)
 Other9(9%)6(8%)2(40%)01(11%)
labMELD16(6–47)17(6–47)8(6–41)14(8–44)16(11–33)
 Creatinine1(1–4)1(1–4)1(1–4)1(1–4)1(1–1.7)
 INR1.5(1–5.6)1.5(1–5.6)1.1(1–2)1.4(1.2–3.3)1.6(1.1–4.1)
 Bilirubin2.9(1–54.8)3.4(1–54.8)1(1–38)2.3(1–30.8)2.8(1–15.9)
Hemodynamic classification
 Type 17(7%)3(4%)2(40%)1(8%)1(11%)
 Type 224(23%)20(26%)2(15%)2(22%)
 Type 311(11%)8(11%)2(40%)1(8%)
 Type 461(59%)45(59%)1(20%)9(69%)6(67%)
Table 3.  Donor characteristics
 Overall (n = 103)FS (n = 76)LDLT (n = 5)SLT (n = 13)DCD (n = 9)
  1. Results are given in number (percentage) or median (range). No significant differences between groups were found.

  2. FS = full size liver transplantation; LDLT = living donor liver transplantation; SLT = split liver transplantation; DCD = donation after cardiac death; ICU = intensive care unit; AST = aspartate aminotransferase; ALT = alanine aminotransferase; ECD = extended criteria donor; DRI = donor risk index; BMI = body max index; UW = University of Wisconsin flush and preservation solution; HTK = histidine–tryptophan–ketoglutarate flush and preservation solution; NA = not applicable.

Age(years)45(12–74)47(12–74)45(24–51)29(18–53)47(13–62)
 >60 years14(13%)12(11.6%)002(22%)
ICU>7d17(17%)8(11%)03(23%)6(67%)
AST/ALT x319(18%)15(20%)04(31%)0
Na+ >15021(20%)19(25%)02(15%)0
HD instability17(17%)15(20%)01(8%)1(11%)
ECD21(20%)11(14%)01(8%)9(100%)
DRI1.68(1.03–3.06)1.58(1.03–2.55)NA2.05(1.53–3.06)2.26(1.79–2.91)
Macrosteatosis4(0–40)4(0–40)5(0–10)0(0–5)2(0–25)
 >30%3(5%)3(6%)000
BMI24(17–38)24(17–33)24(21–27)23(20–30)23(19–38)
UW69(67%)60(79%)2(33%)7(54%)0
HTK34(33%)16(21%)3(66%)6(46%)9(100%)
Recovery of organs     
 Local23(22%)10(13%)5(100%)2(15%)6(67%)
 National45(44%)40(53%)03(23%)2(22%)
 International35(34%)26(34%)08(62%)1(11%)
Table 4.  Intraoperative data and graft characteristics
 Overall (n = 103)FS (n = 76)LDLT (n = 5)SLT (n = 13)DCD (n = 9)
  1. Results are given in median (range).

  2. FS = full size liver transplantation; LDLT = living donor liver transplantation; SLT = split liver transplantation; DCD = donation after cardiac death; CIT = cold ischemia time; WIT = warm ischemia time; TIT = total ischemia time; OT = operative time; GW = liver weight; GRWR = graft-to-recipient weight ratio; GV/SLV = graft volume-to-standard liver volume ratio; Time rep-M2 = time from reperfusion to M2 measurement; Time M2–M3 = time between measurements after reperfusion M2–M3.

  3. 1FS vs. LDLT, p < 0.05.

  4. 2FS vs. SLT, p < 0.05.

  5. 3FS vs. DCD, p < 0.05.

  6. 4LDLT vs. SLT, p < 0.05.

  7. 5LDLT vs. DCD, p < 0.05.

  8. 6SLT vs. DCD, p < 0.05.

CIT(min)475(79–905)495(184–905)1,3110(79–146)4591(450–755)6338(286–350)
WIT(min)50(23–83)45(23–70)175(60–76)439(29–83)54(32–75)
TIT(min)531(154–970)535(235–970)1,3178(154–222)4635(510–790)6386(333–425)
OT(min)540(285–980)540(300–840)885(615–980)570(360–840)503(285–660)
GW(mg)1456(321–2614)1500(790–2614)1,2595(321–686)51130(460–1780)61975(1300–2300)
GRWR1.85(0.53–4.06)1.96(1.01–4.06)1,20.83(0.53–0.86)51.55(0.74–2.03)62.62(1.53–3.25)
GV/SLV(%)109(23–182)110(72–182)49(23–59)82(31–115)129(100–181)
Time rep-M237(14–134)36(14–134)51(40–98)53(22–130)35(28–91)
Time M2–M360(10–193)52(10–148)94(71–147)52(17–193)72(33–110)
Table 5.  Intraoperative systemic and hepatic hemodynamics
  M0M1M2M3
  1. Values expressed as median(range).

  2. M0 = recipient flows; M1 = temporary portocaval flows; M2 = flows at reperfusion; M3 = second measurement of hepatic flows after reperfusion; CO = cardiac output; MAP = mean arterial pressure; HAF = hepatic arterial flow; PVF = portal vein flow; TBF = total hepatic blood flow; P/A ratio = portal-to-arterial flow ratio; NA = not applicable.

  3. 1M0 vs. M2, p < 0.05.

  4. 2M0 vs. M3, p < 0.05.

  5. 3M1 vs. M2, p < 0.05.

  6. 4M1 vs. M3, p < 0.05.

Systemic hemodynamics
CO(L/min)7.3(4.9–13.3)7.2(4.5–13.1)8.1(4.3–14.2)9(4.9–16.9)
MAP(mmHg)71(60–96)73(53–88)67(48–108)68(50–104)
Hepatic hemodynamics
HAF(mL/min)329(73–1031)1NA175(27–630)198(36–834)
(mL/min/100 g LW)18.3(3.5–63.3)NA11.9(1.2–64.2)14.2(2.7–65.8)
PVF(mL/min)943(18–2780)1,2958(312 – 2597)3,41912(453–5420)1618(446–3654)
(mL/min/100 g LW)50.4(1.4–534.6)1,279.1(13.4 – 492)3,4125.6(26.2–433.6)121.5(30–383.1)
THF(mL/min)1020.5(445–3109)1,2NA2056(531–5635)1837(607–3975)
(mL/min/100 g LW)54.6(21.7–598)1,2NA140(30.2–450.8)130.9(43.7–396.9)
P/A ratio 3.8(0.03–12.1)1,2NA9.8(1.6–57.1)7.9(1.2–42)
Table 9.  Linear regression analysis of variables influencing hepatic flows
 PVF (mL/min)HAF (mL/min)
Coefficient95% CIpCoefficient95% CIp
LowerUpperLowerUpper
  1. Subgroup analysis based on liver transplant recipient classification according to systemic hemodynamics and portal hypertension.

  2. PVF = portal vein flow; HAF = hepatic artery flow; Type 1, absence of portal hypertension and systemic hyperdynamic status (reference value); Type 2, portal hypertension alone; Type 3, systemic hyperdynamic status alone; Type 4, presence of both portal hypertension and systemic hyperdynamic status; WIT, warm ischemia time.

Univariate
 Type 2446.9−398.11292.10.29−59.2−248.3129.80.53
 Type 3377.9−597.91353.70.44148.2−70366.50.18
 Type 41017.7227.418080.0125.5−170.3182.30.95
Multivariate
 Type 4604.2187.11021.30.006102.2−3.84208.20.058
 WIT−17.9−34−1.80.03    

Results

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

A hundred and three transplantations in 100 patients with complete intraoperative hemodynamic assessment were considered for analysis. The median follow-up was of 14 months (range 0.03–34). Patient and donor characteristics are shown in Tables 2 and 3. Portal hypertension was present in 83% of indications. Measurement of spleen diameter in the PHT patients showed splenomegaly (>13 cm) and was significantly bigger compared to non-PHT patients (14.88 ± 2.94 vs. 11.62 ± 1.41, p < 0.001). All graft type groups were comparable to each other for recipient and donor characteristics (Tables 2 and 3). The cold ischemia time (CIT) and the total ischemia time of LDLT and DCD transplants were shorter compared to full size grafts (FS) and SLT (p < 0.001). A shorter CIT is inherent to the LDLT procedure and in the case of DCD because all efforts were made to reduce the ischemic times. In the SLT group the longer CIT reflects the fact that most procedures were carried out ex situ. A longer warm ischemia time (WIT) was observed in the LDLT group compared with FS and SLT groups (p = 0.001) (Table 4). Donor WIT prior to recovery of DCD grafts was not included in the analysis.

Hemodynamics of liver transplantation

Stable baseline readings with a normal morphology of the pulse wave were obtained requiring at most 8 min in all cases. In five patients, portal measurements in the native liver (M0) were not available due to portal vein thrombosis; after thrombectomy flow measurements were collected from M1 to M3 either trough the native vein or through a mesenteric-portal venous jump. PVFs increased significantly from M0 to M2, remaining altered until skin closure, whereas HAF displayed low values. A median 2.5-fold increase in P/A ratio was recorded following reperfusion with a highest value of 57.14 at M2. In M3, 1.5 hours after reperfusion, we recorded a decrease of PVF with a concomitant HAF increase compared to M2 without reaching statistical significance (Table 5; Figure 1). A different portal and arterial perfusion responses to portal hyperflow according to the type of graft was observed. LDLT showed a higher PVF compared to DCD and FS grafts (p = 0.009 and 0.051, respectively). When compared to healthy donor flows, PVF after reperfusion were higher (FS, p = 0.01; LDLT, p = 0.004; SLT, p = 0.04; DCD, p = 0.2) and, conversely, the HAF of the different graft types were lower (FS, p < 0.001; LDLT, p = 0.3; SLT, p = 0.04; DCD, p = 0.005). Consequently, the P/A ratio of all graft types was significantly increased (FS, p < 0.001; LDLT, p = 0.007; SLT, p = 0.003; DCD, p = 0.002) (Table 6; Figure 2). Flows according to the four types of liver transplant patients are given in Table 7.

Table 6.  Intraoperative hepatic flows after reperfusion according to graft type
 FS (n = 76)LDLT (n = 5)SLT (n = 13)DCD (n = 9)
  1. Values expressed as median (range). All hepatic hemodynamic values are indexed to the graft weight (mL/min/100 g LW) to allow comparisons between different graft types.

  2. FS = full size liver transplantation; LDLT = living donor liver transplantation; SLT = split liver transplantation; DCD = donation after cardiac death; HAF = hepatic arterial flow; PVF = portal vein flow; TBF = total hepatic blood flow; P/A ratio = portal-to-arterial flow ratio.

  3. 1FS vs. LDLT, p = 0.009.

  4. 2LDLT vs. DCD, p = 0.051.

  5. Compared to healthy live donor flows (HAF 25 mL/min/100 g LW) and PVF (90 mL/min/100 g LW); PVF were higher (FS, p = 0.01; LDLT = p = 0.004; SLT = p = 0.04; DCD = p = 0.2) and = conversely = the HAF of the different graft types were lower (FS = p < 0.001; LDLT = p = 0.3; SLT = p = 0.04; DCD = p = 0.005). Consequently, the P/A ratio of all graft types was significantly increased (FS, p < 0.001; LDLT, p = 0.007; SLT, p = 0.003; DCD, p = 0.002).

HAF11.6(1.2–64.2)12.4(4.5–60)13.2(4.1–37.2)9.47(3.9–16.8)
PVF113.9(26.5–433.6)1179.3(163–291.7)2147.1(44.6–342.5)110.4(26.2–217.8)
THF130.6(42.5–450.8)201.4(174.8–311.6)157.6(63.5–350.7)120.7(30.2–229.7)
P/A ratio8.3(1.6–57.1)14.6(2.9–38.4)14.1(2.3–41.9)11.4(6.6–27.1)
image

Figure 2. Liver perfusion according to graft type: full size (FS), living donor liver transplantation (LDLT), split liver transplantation (SLT), donation after cardiac death (DCD). Lines represent median hepatic artery and portal vein flows registered from a historical cohort of living donors used as reference values for comparison (dashed line: HAF 20.5 mL/min/100 g LW, solid line: PVF 90 mL/min/100 g LW). Boxes represent interquartile range, whiskers minimum and maximum values.

Download figure to PowerPoint

Table 7.  Intraoperative hepatic flows after reperfusion in recipient classification types according to systemic hemodynamics and portal hypertension
 Type 1 (n = 7)Type 2 (n = 24)Type 3 (n = 11)Type 4 (n = 61)
  1. Values expressed as median (range). All hepatic hemodynamic values are indexed to the graft weight (mL/min/100 g LW) to allow comparisons between different graft types.

  2. Type 1, absence of portal hypertension and systemic hyperdynamic status; Type 2, portal hypertension alone; Type 3, systemic hyperdynamic status alone; Type 4, presence of both portal hypertension and systemic hyperdynamic status; HAF = hepatic arterial flow; PVF = portal vein flow; TBF = total hepatic blood flow; P/A ratio = portal-to-arterial flow ratio.

  3. 1Type 2 vs. type 3, p = 0.048.

  4. 2Type 2 vs. type 4, p = 0.008.

HAF9.9(3.9–37.2)10(6.2–39.4)119.8(8.2–60)12.2(1.2–64.2)
PVF86.8(26.2–174.1)80.1(26.5–182.4)2131.6(53.3–291.7)134.1(44.6–433.6)
THF124(30.2–178.7)89.8(42.5–192.4)2139.9(64.8–311.6)149.4(56.5–450.8)
P/A ratio7.2(2.3–38.4)7.8(1.6–18.2)5.5(2.9–16)10.9(2.3–57.1)

GIM was applied in five recipients presenting high PVF (260.4 mL/min/100 g LW, range 172.1–986.9), low HAF (7.14 mL/min/100 g LW, range 5.4–16) and high P/A ratio (33.6, range 18.4–66). Two were FS grafts, two LDLT and one SLT. One hemiportocaval shunt and four splenic artery ligations were performed reducing PVF (130 mL/min/100 g LW, range 95.2–205.8) (p = 0.01), increasing HAF (15.5 mL/min/100 g LW, range 11.9–18.7) (p = 0.08) and reducing the P/A ratio (10.5, range 6.1–16.5) (p = 0.027). In another patient, an increased upstroke and low systolic flow with hepatofugal diastolic flow, suggested an arterial obliteration. After surgical revision of the anastomosis, measured flows showed normal values and morphology. Omentoplasty increased the recorded flows in another patient with multiple arteries.

CO was slightly correlated with hepatic flows: a positive correlation for the HAF (r = 0.24, p = 0.04), PVF (r = 0.35, p = 0.002) and total hepatic blood flow (TBF) (r = 0.38, p = 0.001) were found (Figure 3).

image

Figure 3. Systemic and hepatic interrelationship: cardiac output (CO) and total hepatic blood flow (THF).

Download figure to PowerPoint

Among the characteristics studied in recipient and donors, as well as intraoperative, potentially influencing hepatic flows, PHT, hypernatremia, WIT and CO showed an association with PVF, whereas hypernatremia, CO and liver weight showed an association with HAF. Variables independently associated to hepatic flows were: portal hypertension, steatosis >30%, WIT and CO (Table 8). Regarding clinical classification of the liver transplant recipient in terms of PHT and hyperdynamic status, only type 4 was independently associated with hepatic flows (Table 9).

Table 8.  Linear regression analysis of variables influencing hepatic flows
 PVF (mL/min)HAF (mL/min)
Coefficient95% CIpCoefficient95% CIp
LowerUpperLowerUpper
  1. HAF = hepatic artery flow; PVF = portal vein flow; PHT = portal hypertension; MELD = model for end-stage liver disease; donor Na>150 = natriemia in mmol/L; DRI = donor risk index; ECD = extended criteria donor; HTK = histidine–tryptophan–ketoglutarate flush and preservation solution; CIT = cold ischemia time; WIT = warm ischemia time; GW = graft weight; CO = cardiac output in L/min.

Univariate
 PHT631.59220.471042.710.003−59.41−143.2224.390.1
 MELD5.01−12.1522.180.561.86−1.535.250.28
  Creatinine34.17−222.9291.30.79−22.95−74.628.690.38
  INR2.09−174.7178.90.9816.1−19.2751.480.37
  Bilirubin−4.84−20.8711.20.552.12−0.734.980.14
 Donor age128.41−409.6666.40.64−39.47−147.768.780.47
 Donor Na>150−497.43−881.24−113.620.0190.119.2169.980.013
 Macrosteatosis2.14−21.8926.180.86−0.98−5.783.820.68
 Steatosis >30%169.66−719.871059.190.3−89.8−226.6847.070.19
 DRI175.81−169.54521.160.31−53.06−121.7115.560.12
 ECD233.4−160.2627.10.24−48−127.331.120.23
 HTK−65.05−412.4282.20.71−22.63−90.9545.680.51
 CIT (min)0.45−0.0551.460.370.11−0.0840.320.25
 WIT (min)−15.6−26−50.004−0.45−2.61.70.68
 GW (mg)0.27−0.080.630.10.120.050.180.001
 CO (L/min)97.624.8170.40.00916.70.8532.60.039
Multivariate
 CO (L/min)153.649258.30.00644.825.664<0.001
 PHT    −263.06−501.5−24.60.032
 Steatosis >30%    −332.06−576.07−88.030.009
 WIT (min)−21.1−38−4.30.016    

Outcome

One-year graft survival was of 90.5%. Early hepatic artery thrombosis (HAT) occurred in two patients (1.9%); no late HAT was observed during the follow-up. No other vascular complications were observed. The first HAT occurred in postoperative day 21, in a patient with normal hepatic flows at transplantation. The control tests revealed a positive lupus anticoagulans, a well-known risk factor for HAT (41,42). Other causes of graft loss were: n = 2 primary non function, n = 1 ischemic type biliary lesions (DCD graft), n = 1 chronic rejection at 11 months and n = 1 severe ischemia–reperfusion injury, giving a total incidence of 6.8%. All patients with graft failure but one were retransplanted (n = 6, 5.8%).

Discussion

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

Interpretation of normal liver flows during LT is complex and is influenced by several factors such as the quality of the organ, ischemia–reperfusion damage, clinical status of the recipient, systemic hemodynamics and anesthesiology management. The heterogeneity of methods used in literature, type of patients and reporting of flows or perfusions add further complexity to the interpretation of an adequate inflow (Table 1). In these exploratory analyses, we present a prospective data collection with uniform methods in over 100 liver transplants.

Table 1.  Intraoperative direct measurement of hepatic flow
AuthorYearGraft typenTypeHAFPVFTBFP/A ratio
  1. All reported studies are retrospective except the present series. The values are expressed as mean perfusion values (mL/min/100 g LV) unless specified, values between parenthesis represent flow values (mL/min).

  2. HAF = hepatic arterial flow; PVF = portal vein flow; TBF = total hepatic blood flow; P/A ratio = portal-to-arterial flow ratio; TTFM = transit time flow measurement.

  3. 1, median values calculated from published data; 2, GV/SLV≤40% grafts; 3, median values; 4, GRWR≤0.8 in 9 (37%) patients; 5, full size liver transplantation, living donor liver transplantation, split liver transplantation, donation after cardiac death (for a detailed description of the flows observed in the different graft types, please see Table 6).

Paulsen (8)1992FS282Electromagnetic32.5 (571.8)137.2 (2348.3)169.9 (2920.3) 4.1
Henderson (9)1992FS34TTFM(268)(1808)(2091) 6.7
Margarit (44)1999FS45TTFM(184)(1590)(1744) 8.6
Marcos (18)2000LDLT16Electromagnetic(112.5)1(1909)12030118.2
Shimamura (52)2001LDLT220TTFM21.3373.4384.717.5
García-Valdecasas(66)2003LDLT22TTFM16 (121)3243 (1970)3259 (2091)315.2
Troisi (54)2003LDLT424TTFM12.8 (104)318 (2100)330.8 (2155)24.8
Troisi (24)2005LDLT13TTFM20.5469489.522.8
Gontarczyk (67)2007FS15TTFM16.2 (158)1127 (1700)1143.2 (2180)1 7.8
Hashimoto (68)2010FS234TTFM171151321 6.7
Present series2010All types5103TTFM14.5 (198)3121.5 (1618)3130.9 (1837)3 7.9

Native flows through the cirrhotic liver are characterized by low PVF and concomitant increase of HAF (8,43,44), which is confirmed by the results of our study (Table 5; Figure 1). After transplantation PVF increases to double the flows observed in healthy subjects (43,45,46). In half of the patients, PVF accounted for 93% of total liver flow after reperfusion. Portal vein perfusion values were higher in all graft types, the HAF values were inferior and the P/A ratio at least doubled by comparison to the reference flows from healthy live donors, suggesting a still functioning hepatic artery buffer response (Figure 2). In contrast, the native liver had a low portal to arterial ratio confirming a chronically active HABR with increased HAF in response to the reduced PVF, registering an inferior range as low as 0.03 (19).

WIT intensifies the ischemia–reperfusion damage of the hepatic sinusoids causing structural alterations and impaired microcirculation, and has been identified as an independent risk factor for graft failure (47). An association between the severity of steatosis and the degree of microvascular impairment has been previously shown (48–51). In our study, a negative trend for PVF and a positive trend for HAF did not reach significance. When grouped, macrosteatosis >30%, present in 5% of grafts, showed a reduced HAF in multivariate linear regression model, even though confidence intervals are wide.

As expected, graft weight (GW) and graft-to-recipient weight ratio (GRWR) of LDLT and SLT groups were smaller than FS and DCD groups. The use of perfusion values (mL/min/100 g LW), as presented, allows for a better understanding of the stress that an augmented portal flow poses on the liver graft by correcting the measured flows for the amount of liver that has to accommodate it, and minimizes the influence of GW on the measured flows. Indeed, the median GW of the SLT doubled the weight of the LDLT group but the flows were comparable (Figure 2). In clinical practice, the ‘ideal’ target PVF for LDLT has been regarded as twice the perfusion observed in the FS graft (250 mL/min/100 g LW) (8,52,53), or as twice the flows observed in the healthy donor (180 mL/min/100 g LW) (25). In LDLT, in spite of the smaller GW, and consequently reduced hepatic vascular bed, the highest recorded PVF were observed (Figure 2), doubling the reference values measured in living donors. We could hypothesize that, granted a good hepatic outflow as described in our technique, the optimal quality of the organ allows for a higher compliance of the portal hyperperfusion when safe GRWR and PVF are respected (24,25,52,54). The opposite may hold true for DCD grafts, where we observed the highest GW and GRWR and still, the TBF was low after reperfusion consistent with reports of edema and decreased compliance after cold preservation and warm ischemia (Tables 4 and 6; Figure 2) (55–58).

This study also suggests a direct effect of hyperdynamic systemic circulation on hepatic hemodynamics during LT. The greater the CO the higher the hepatic flows (Figure 3). Total hepatic blood flow after reperfusion represented in median 23% of CO. As previously described (9,16,19,59), our results seem to confirm the presence of an active intrahepatic buffer response, while both HAF and PVF are influenced by systemic hemodynamics. According to multivariate analysis, the effect on hepatic flows of CO seems to affect to a greater extent the HAF (Table 8).

An increase of HAF and PVF with a concomitant CO increase has been previously described when comparing hemodynamic patterns in the same patients before and after LT (8). Previous experience has shown single correlation of CO with PVF and a negative trend with HAF (9), while our prospective study indicates a correlation with HAF, PVF and TBF. The elevated TBF probably represents a significant contribution to the elevated CO observed after reperfusion (Table 5; Figure 1), which in turn increases the hepatic flows (Figure 3). Indeed, persisting portosystemic collateral circulation has been described up to 23 months after transplantation (10,11). Surprisingly, PVF registered through the temporary porto-caval shunt showed inferior flows and perfusions compared to reperfusion values. Besides a possible technical factor regarding torsion of the portal vein or anastomotic stricture of the temporary PCS, the increased flow after reperfusion observed could be a direct effect of increased CO. Indeed, another possible explanation relates to the intraoperative management of fluid and drug administration in different phases of LT. Prior to the reperfusion phase, the hemodynamic status is optimized and as a result CO may increase.

Our study presents some limitations, from data collection to data analysis. Concerning the latter, it could be fittingly underlined that our analyses are mainly exploratory, because reaching an adequate sample size for this type of hypotheses is not feasible in any clinical setting. With regards to the former, because we did not collect prospectively data regarding volume filling during transplantation we cannot ascertain the effect of fluid load and vasopressor use. We assumed that patients that did not need vasoactive therapy while maintaining a diuresis above 1 mL/kg/h after reperfusion, were normovolemic. Our results highlight the crucial role of an adequate hemodynamic management during, and immediately after, LT to optimize hepatic hemodynamics, particularly in LDLT and SLT, where the highest flows were registered.

In spite of a higher PVF, type 3 (hyperdynamic patients alone) showed the highest HAF of all four types, and significantly higher than type 2 patients (PHT alone). This result is in line with the correlation found between CO and HAF. Type 4 patients showed the influence of combining PHT and hyperdynamic status on the hepatic flows, a significantly higher PVF compared to type 2, where a reduced flow can be expected. Furthermore, type 4 showed an independent association with increased hepatic flows, stressing the need for flow measurement in this type of patients (Tables 7 and 9).

With the use of flow measurement we were able to identify 7 patients (6.9%) with abnormal average flows requiring correction, an incidence congruent with previous experiences (59). Two of these patients with normal pulsatility at palpation presented altered flow measurements, allowing us to identify and correct technical failures that otherwise could have compromised outcome (59,60). The low rate of HAT (1.9%) compares favorably with available reports (61,62). It is well known that causes of thrombosis are always multifactorial (61–63). However, we might speculate an effect of the systematic graft flow measurement directing inflow modulation or correction of vascular anastomosis on this low incidence. GIM is applied only in case of transplantation with small-for-size grafts (64,65). A different situation is represented by the need of GIM in FS grafts. In a previous publication, hyperperfusion was defined as flow values four times above the reference values observed in living donors (≥360 mL/min/100 g LW). FS grafts transplanted into hyperdynamic patients with hyperperfusion flow measurements presented similar histological changes as those observed in small-for-size grafts with inferior graft survival (22). In the absence of sound evidence in either direction, we choose to improve the chance of recovery of all grafts in case of manifest hyperperfusion. On the contrary, patients with low PVF can present with spontaneous porto-systemic shunts (SPSS). In such case, a clamp test can be performed and, if PVF improve, the SPSS should be ligated.

We have described the evolution of hepatic flows according to donor factors such as macrosteatosis superior to 30% and technical factors such as WIT that showed a significant association with inferior hepatic flows. Other factors of suboptimal quality of grafts potentially influencing graft perfusion, either individually (donor age, CIT, DRI, steatosis), or when grouped (ECD), did not show an association with hepatic flows in this exploratory study. The question of whether attributing abnormal flows to the graft quality in marginal grafts remains open. The analysis of associations between systemic and regional hemodynamics showed a significant relationship between CO and HAF, PVF and TBF. Although an association between CO and HAF is intuitive, the association found between CO and PVF (PVF representing in median 93% of TBF), is a more complex one in need of further investigation.

In conclusion, hemodynamics of LT show a significant increase of PVF during the different phases when compared to native liver and reference flows recorded in living donors. LDLT and SLT showed a higher capability to comply with a superior hemodynamic stress. This higher perfusion must be anticipated and GIM considered, also in FS grafts. The association between systemic and hepatic hemodynamics was prospectively confirmed. Portal hypertension, macrosteatosis >30%, WIT, CO and type 4 recipients were identified as independent variables influencing hepatic hemodynamics. Notwithstanding the multiplicity of factors modifying hepatic hemodynamics, which hinder an identification of the ideal inflow, this experience adds a piece to the puzzle of liver hemodynamics during LT with an insight into different graft type perfusions.

Acknowledgments

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

The authors are deeply grateful to Mr. Hugo Claus for his continuous assistance in collecting the intraoperative flow measurements. The authors also thank Dr. Susana Torres Prieto for her revision of the paper.

References

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