Randomized clinical trial comparing infrahepatic inferior vena cava clamping with low central venous pressure in complex liver resections involving the Pringle manoeuvre

Abstract

Background

Control of bleeding remains key to successful hepatic resection. The present randomized clinical trial compared infrahepatic inferior vena cava (IVC) clamping with low central venous pressure (CVP) during complex hepatectomy using portal triad clamping (PTC).

Methods

Consecutive patients undergoing complex hepatectomy were allocated randomly to PTC combined with infrahepatic IVC clamping or to PTC with low CVP. Primary outcome was blood loss during parenchymal transection. Secondary outcomes were intraoperative surgical and haemodynamic parameters, postoperative recovery of liver and renal function, postoperative morbidity and mortality, and duration of hospital stay.

Results

Between January 2008 and September 2010, 192 patients were randomized. Compared with low CVP, infrahepatic IVC clamping significantly decreased blood loss during parenchymal transection (mean(s.e.m.) 243(158) versus 372(197) ml; P < 0·001), was associated with faster recovery of liver function, and caused less impairment in renal function and fewer haemodynamic changes. The degree of cirrhosis correlated positively with CVP (R² = 0·963, P = 0·019) and with infrahepatic IVC pressure (R² = 0·950, P = 0·025). For patients with moderate or severe cirrhosis, infrahepatic IVC clamping was more efficacious in controlling blood loss during parenchymal transection (mean(s.e.m.) 2·9(1·8) versus 6·1(2·4) ml/cm²; P < 0·001).

Conclusion

PTC combined with infrahepatic IVC clamping is more efficacious in controlling bleeding during complex hepatectomy than PTC with low CVP, especially in patients with moderate to severe cirrhosis. Registration number: NCT01355887 (http://www.clinicaltrials.gov).

Introduction

Excessive blood loss during liver resection and perioperative blood transfusion have a negative impact on perioperative morbidity and mortality, as well as on long-term survival^1–3. Using modern technology, hepatic parenchymal transection can be carried out with little blood loss. Many liver surgeons still prefer to use hepatic vascular inflow occlusion with, or without outflow occlusion during parenchymal transection, especially in complex liver resections^3–6, to limit blood loss while performing safe and expeditious resections. There are many ways to control hepatic vascular blood flow⁷. These include portal triad clamping (PTC or Pringle manoeuvre), selective vascular inflow occlusion (inflow occlusion of the liver lobe to be resected), total or selective hepatic vascular exclusion (THVE or SHVE, inflow and outflow occlusion of the whole liver or the lobe to be resected), or hepatic vascular exclusion with preservation of caval flow (HVEPC). PTC is sufficient in most situations to control bleeding from the hepatic artery or portal vein during hepatectomy². However, inflow clamping does not prevent backflow bleeding from hepatic veins, which can be massive and may result in death⁴. Control of bleeding from the hepatic veins may be a major problem during complex liver resections.

Techniques that involve hepatic vascular inflow and outflow exclusion, including THVE, SHVE and HVEPC, are recommended for use in patients with tumours in close proximity to major hepatic veins or the inferior vena cava (IVC)^4,8. A drawback of these techniques that has limited their use is the need to dissect the suprahepatic and infrahepatic IVC, or major hepatic veins. The present authors have developed a modified technique of hepatic vascular exclusion, and have found this to be safe and effective in reducing blood loss in complex liver surgery⁶. This method involves the use of PTC to control hepatic vascular inflow combined with infrahepatic IVC clamping (IIVCC) to control backflow bleeding. The present randomized clinical trial was designed to compare PTC combined with either IIVCC or low central venous pressure (CVP) in complex liver resections.

Methods

Between January 2008 and September 2010, all patients undergoing liver resection at Tongji Hospital were eligible for this study. Inclusion criteria were: a high risk of backflow bleeding from major hepatic veins and IVC, based on the size and location of the tumour on preoperative imaging (tumour at least 5 cm in diameter, involving liver segments VII, VIII and/or the cranial portion of segment IV; tumour in close proximity or compressing, but not actually invading, major hepatic veins or IVC); Child–Pugh grade A; indocyanine green retention rate at 15 min (ICG-R15) below 10 per cent; acceptable clotting profile (platelet count at least 50 × 10⁹/l and prothrombin activity 60 per cent or more); and no previous liver resection. Exclusion criteria were: extrahepatic metastases in patients with malignancy and refusal to take part in the study. This study was approved by the ethics committee of Tongji Hospital. Informed consent was obtained from each patient before operation.

Outcomes

The primary outcome was blood loss during parenchymal transection. Secondary outcomes were intraoperative surgical data, intraoperative haemodynamic parameters, postoperative recovery of liver and renal function, postoperative morbidity and mortality, and duration of hospital stay.

Randomization

Randomization was done during operation, after abdominal exploration and intraoperative ultrasonography had confirmed that the patient was eligible for the study. Sealed and opaque envelopes containing random numbers were used for randomization, and were opened by a nurse who was not involved in the study.

Preoperative assessment

Routine preoperative assessment included liver function tests, full blood count, coagulation tests, hepatitis B serology, measurement of serum urea, creatinine, electrolytes, α-fetoprotein and carcinoembryonic antigen, ICG-R15, electrocardiography and chest X-ray. Abdominal ultrasonography, abdominal computed tomography and/or magnetic resonance imaging were used to assess tumour resectability.

Operative technique

Surgery was performed through a right subcostal incision. The abdominal cavity was explored carefully to determine the extent of local disease, extrahepatic metastases and peritoneal seeding. Intraoperative ultrasonography was used in all patients to assess the number and size of the tumours, and their relation to vascular structures.

For patients in the IIVCC group, the infrahepatic IVC was dissected and taped with a vessel loop. Aberrant hepatic arteries, if present, were clamped. For patients in the low CVP group, the CVP was maintained at 2–5 cmH₂O. A low CVP was achieved mainly by the following methods: vasodilatation (0·5–6 mg/h nitroglycerin) assisted if necessary by the reverse Trendelenburg position, forced diuresis (10–20 mg furosemide) and fluid restriction (100 ml/h). Continuous vascular occlusion of the portal triad for up to 30 min was used for non-cirrhotic livers. Intermittent vascular occlusion with clamp/unclamp times of 10/5 min was used for cirrhotic livers.

Parenchymal transection was performed using Kelly forceps. Small vessels (less than 2 mm) were coagulated with an irrigated bipolar forceps set at 120 W, whereas larger vascular structures and intrahepatic bile ducts were ligated or clipped.

After parenchymal transection, the hepatic circulation was restored by unclamping and increasing the CVP back to normal (10–15 cmH₂O). Haemostasis was secured using 4/0 polypropylene sutures, and by coagulation to control oozing on the raw liver surface. Fibrin glue was used routinely on the raw liver surface. Surgery was carried out under general anaesthesia with tracheal intubation and controlled ventilation. Epidural anaesthesia was not used.

Data collection

Preoperative demographic and clinical data, and details of the surgical procedure, pathological diagnosis, postoperative course and complications were collected prospectively. Transection time was defined as the interval between the beginning and end of hepatic parenchymal transection. Time to haemostasis was defined as the interval between release of the portal triad clamp and completion of haemostasis. Blood loss was measured from the start of the operation to the beginning of parenchymal transection, during transection, and from the end of transection to the end of operation. Blood loss was estimated, taking into account suction volume minus rinsing fluids, and the weight of swabs used during transection (assuming that 1 ml of blood weighs 1g). Blood loss was also calculated relative to the transection area (ml/cm²). A sponge was put on the raw surface and the liver transection surface was drawn on a sheet of A4 paper (80 g/m²). The drawn area was cut out and weighed to calculate the surface area based on the paper density. The speed of transection was calculated by dividing transection area by transection time (cm²/min).

Mean arterial pressure (MAP), heart rate and CVP during surgery were monitored using a Medex TranStar® 84 in Single Monitoring Kit (Smiths Medical, St Paul, Minnesota, USA) connected to a Solar® 8000M Modular Patient Monitor (GE Medical Systems, Milwaukee, Wisconsin, USA). All these parameters were recorded at different time points during the operation by the anaesthetist. A 21-G needle connected to a monitoring kit was inserted into the infrahepatic IVC to measure pressure (IIVCP). Indications for red blood cell transfusion were blood loss exceeding 1000 ml or a haemoglobin level below 5·6 mmol/l during, or within 48 h of surgery.

The degree of liver cirrhosis was evaluated based on the diameter of cirrhotic nodules in the resected specimens⁹: grade 1, no cirrhosis; grade 2, slight cirrhosis—nodule diameter less than 0·4 cm; grade 3, moderate cirrhosis—nodule diameter 0·4–0·8 cm; and grade 4, severe cirrhosis—nodule diameter more than 0·8 cm. The tumours were diagnosed histopathologically.

Serum total bilirubin, aspartate aminotransferase (AST), alanine aminotransferase (ALT) and prothrombin time were measured on days 1, 3, 5, 7 and 14 after surgery. Postoperative morbidity and mortality were assessed according to the Clavien–Dindo classification¹⁰.

Statistical analysis

The objective of the study was to assess the non-inferiority of IIVCC compared with low CVP in complex liver resections involving the Pringle manoeuvre. The sample size calculation was based on a margin of non-inferiority for blood loss during transection set at 50 ml. Previous studies by the authors' group showed that mean(s.e.m.) blood loss during transection with PTC combined with IIVCC or low CVP was 350(110) ml⁶ and 330(230) ml respectively. A sample size of 80 per group was calculated for an α risk of 0·05 and a β risk of 0·10. Assuming a postrandomization dropout rate of up to 20 per cent, the sample size required for this study was estimated to be 96 patients in each study group.

All analyses were performed on an intention-to-treat basis. Summary statistics are expressed as mean(s.e.m.) or median (range), as appropriate. Comparisons between groups were done using Student's t test for continuous data and χ² test for categorical data. The area under the curve (AUC) was determined for the first 5 days after surgery. Integration of the AUC was performed using GraphPad Prism® software (GraphPad Software, San Diego, California, USA). The reference area, defined by the minimum values, was subtracted from the total integrated AUC. Two-variable correlation analysis was performed by a linear correlation method. Multiple linear regression analysis was carried out to find the key factor influencing IIVCP. All statistical tests were two-sided, and P < 0·050 was considered significant. SPSS® 13.0 statistical software (SPSS, Chicago, Illinois, USA) was used for statistical analysis.

Results

During the study period, 1874 liver resections were performed at the authors' institution. Of these, 215 were assessed for eligibility and 192 patients were randomized to one of the two groups (Fig. 1). There were no significant differences in demographic parameters (sex, age, body mass index), preoperative laboratory tests (haemoglobin, platelet count, prothrombin time, ALT, AST, bilirubin, creatinine, hepatitis B and C virus serology, and ICG-R15), proportion of malignant and benign diseases, size of tumours, types of hepatectomy, or degree of liver cirrhosis between the two groups.

Fig. 1

CONSORT diagram for the trial. IIVCC, infrahepatic inferior vena cava clamping; CVP, central venous pressure

Blood loss

Total blood loss in the IIVCC and the low CVP groups was 315(192) and 439(229) ml respectively (P < 0·001), and blood loss during hepatic transection was 243(158) versus 372(197) ml (P < 0·001). After standardization for transection surface area, the mean transection-related blood loss in the IIVCC group was also significantly lower than that in the low CVP group (3·0 versus 5·5 ml/cm²; P < 0·001).

Details of surgical procedure

The IIVCC group had a significantly shorter transection time than the low CVP group (11·6 versus 14·2 min; P = 0·003). The hepatic transection area was comparable in the two groups (74·8 versus 71·6 cm² respectively; P = 0·387). Thus, a significantly greater transection speed was achieved in the IIVCC group (6·9 versus 6·0 cm²/ min; P = 0·018). The number of clips, sutures and ligatures used per hepatic transection surface area was not significantly different between the two groups (0·368 versus 0·359 per cm²;P = 0·524). There was no difference in time to haemostasis between the two groups (17·3 versus 17·4 min; P = 0·950). Less crystalloid was infused in the low CVP group than in the IIVCC group (1280 versus 1476 ml; P < 0·001). A total of 148 patients (77·1 per cent) received a colloid infusion during the operation, 72 in the IIVCC group and 76 in the low CVP group (P = 0·492). The mean amount of colloid infused was 627 versus 555 ml respectively (P = 0·133). Thirty-six patients (18·8 per cent) received blood transfusions within the first 48 h of operation, 18 in each group.

Intraoperative haemodynamic parameters

During IIVCC or the low CVP phase, CVP, IIVCP and heart rate were lower in the IIVCC group than in the low CVP group, whereas MAP was significantly higher. After unclamping the portal triad, and unclamping the infrahepatic IVC or restoration of CVP to normal, CVP, IIVCP and heart rate were significantly lower in the IIVCC group (Table 1).

Table 1

Intraoperative data

	IIVCC (n = 96)	Low CVP (n = 96)	P†
Duration of operation (min)	161·8(36·1)	172·0(46·2)	0·091
Ischaemia time (min)	10·5(3·6)	12·4(5·0)	0·003
Time of IIVCC or maintenance of low CVP (min)	10·5(4·0)	21·8(3·8)	< 0·001
No. of patients requiring			0·122‡
blood transfusion*
1 unit	2	0
2 units	10	6
3 units	0	4
4 units	6	8
CVP (cmH2O)
Phase 1	10·5(1·3)	10·8(1·2)	0·165
Phase 2	4·3(0·9)	4·7(0·5)	< 0·001
Phase 3	9·9(1·1)	10·5(1·1)	0·001
IIVCP (cmH2O)
Phase 1	9·5(1·3)	9·8(1·1)	0·116
Phase 2	3·4(0·9)	3·7(0·5)	0·001
Phase 3	8·9(1·1)	9·5(1·1)	< 0·001
MAP (mmHg)
Phase 1	91·5(8·6)	90·3(6·8)	0·278
Phase 2	69·8(10·5)	65·1(7·9)	0·001
Phase 3	90·2(8·0)	89·4(8·1)	0·472
Heart rate (per min)
Phase 1	76·9(8·3)	77·4(9·1)	0·680
Phase 2	94·0(11·8)	100·6(11·3)	< 0·001
Phase 3	78·4(8·9)	81·3(8·0)	0·017

Values are mean(s.e.m.) unless indicated otherwise. Phase 1, before parenchymal transection; phase 2, parenchymal transection; phase 3, after parenchymal transection.

*

1 unit of blood is 200 ml. IIVCC, infrahepatic inferior vena cava clamping; CVP, central venous pressure; IIVCP, infrahepatic inferior vena cava pressure; MAP, mean arterial pressure.

†

Student's t test, except

‡

χ² test.

Postoperative hepatic and renal function

Peak values of AST and ALT (Table 2) occurred on the first day after surgery. In most patients, AST and ALT levels returned to normal within 7 days (data not shown). Peak values of total bilirubin in the IIVCC group were not statistically different from those in the low CVP group. However, on days 3, 5 and 7 after operation, total bilirubin levels in the low CVP group were higher than those in the IIVCC group (18·4 versus 14·7 µmol/l, P = 0·034; 17·7 versus 14·2 µmol/l, P = 0·031; 15·8 versus 11·7 µmol/l, P = 0·004). On the first day after surgery, creatinine levels were significantly higher in the IIVCC group than in the low CVP group. Thereafter creatinine levels between the two groups were not significantly different.

Table 2

Postoperative laboratory test results and outcome data

	IIVCC (n = 96)	Low CVP (n = 96)	P†
Nadir haemoglobin (g/dl)	120·9(15·6)	114·5(15·3)	0·005
Nadir platelet count (×103/µl)	132·2(64·9)	123·4(39·7)	0·262
Nadir prothrombin time (%)	91·6(16·9)	82·1(17·7)	0·022
Total bilirubin
Peak (µmol/l)	17·8(10·1)	18·4(15·3)	0·722
AUC (µmol/l per day)	31·7(28·1)	37·1(63·8)	0·447
Aspartate
aminotransferase
Peak (units/l)	374·5(340·4)	383·8(332·0)	0·848
AUC (units/l per day)	973·2(762·8)	979·1(804·3)	0·958
Alanine aminotransferase
Peak (units/l)	384·0(348·0)	420·2(353·7)	0·476
AUC (units/l per day)	1309·0(1017·1)	1343·0(1147·8)	0·828
Creatinine
Peak (µmol/l)	71·3(18·9)	65·8(16·6)	0·033
AUC (µmol/l per day)	55·0(32·4)	44·4(28·2)	0·051
Overall complications*	35 (36)	37 (39)	0·766‡
Total minor complications (grade I + II)*	29 (30)	33 (34)	0·537‡
Total major complications (grade III–V)*	6 (6)	4 (4)	0·516‡
No. of patients admitted to ICU	2	4	0·439‡
ICU stay (days)	(1, 1)	(1, 1, 1, 2)
Hospital stay (days)	14·0(3·1)	13·9(3·7)	0·842

Values are mean(s.e.m.) unless indicated otherwise;

*

values in parentheses are percentages. Nadir values were minimum values measured within the first 5 days after surgery; AUC, area under the curve for the first 5 days after surgery. IIVCC, infrahepatic inferior vena cava clamping; CVP, central venous pressure; ICU, intensive care unit.

†

Student's t test, except

‡

χ² test.

Postoperative complications and hospital stay

There were no in-hospital deaths. Complication rates were similar in the two groups. There was no difference in the incidence of liver failure or intensive care unit admissions between groups (Table 2).

Influence of hepatic cirrhosis

A total of 158 patients (82·3 per cent) had liver cirrhosis. There was a positive correlation between transection-related blood loss and IIVCP (R = 0·145, P = 0·044). The degree of cirrhosis was positively related to IIVCP (R = 0·449, P < 0·001). There was no correlation between IIVCP and sex, age, height and weight. There were significant positive correlations between CVP (R² = 0·963, P = 0·019) or IIVCP (R² = 0·950, P = 0·025) and the degree of cirrhosis. For patients who had no or slight cirrhosis, there was no difference between the change in IIVCP (after IVC clamping or maintaining CVP at a low level) in either group (6·0 versus 5·7 cmH₂O; P = 0·203). However, for patients with moderate or severe cirrhosis, the change in IIVCP was significantly greater in the IIVCC group than in the low CVP group (6·9 versus 6·1 cmH₂O; P = 0·002).

The number of patients with moderate to severe hepatic cirrhosis in the low CVP and IIVCC groups was comparable: 42 (44 per cent) versus 34 (35 per cent) (P = 0·151). There were no significant differences in demographic parameters, tumour characteristics or laboratory tests between the two subgroups, except for a significantly lower platelet count in the low CVP group (138·0 versus 213·6 × 10³/µl; P < 0·001).

Baseline values of IIVCP, MAP and heart rate were comparable between the two subgroups (Table 3). During IIVCC or the low CVP phase, CVP, IIVCP and heart rate were significantly lower in the IIVCC group, whereas MAP was significantly higher.

Table 3

Comparison of intraoperative parameters for patients with moderate and severe hepatic cirrhosis

	IIVCC (n = 34)	Low CVP (n = 42)	P*
Ischaemia time (min)	10·1(4·3)	11·9(5·5)	0·106
Time of IIVCC or maintenance of low CVP (min)	10·8(4·8)	22·5(4·3)	< 0·001
Transection speed (cm2/min)	6·2(2·4)	5·9(2·2)	0·582
Total blood loss (ml)	310(209)	479(294)	0·006
Blood loss during transection (ml)	246(178)	409(246)	0·002
Transection-related blood loss (ml/cm2)	2·9(1·8)	6·1(2·4)	< 0·001
CVP (cmH2O)
Phase 1	11·4(1·0)	11·0(0·8)	0·033
Phase 2	4·4(0·6)	4·9(0·5)	0·001
Phase 3	10·0(1·2)	10·7(0·8)	0·005
IIVCP (cmH2O)
Phase 1	10·4(1·0)	10·0(0·8)	0·061
Phase 2	3·5(0·6)	3·9(0·5)	0·003
Phase 3	9·0(1·2)	9·7(0·8)	0·005
MAP (mmHg)
Phase 1	91·1(11·6)	90·2(6·3)	0·661
Phase 2	69·1(9·8)	64·3(7·8)	0·020
Phase 3	89·5(10·6)	89·3(7·5)	0·921
Heart rate (per min)
Phase 1	76·4(10·2)	73·9(4·0)	0·148
Phase 2	92·2(13·4)	101·5(10·8)	0·001
Phase 3	78·1(10·7)	81·3(7·5)	0·128

Values are mean(s.e.m.). Phase 1, before parenchymal transection; phase 2, parenchymal transection; phase 3, after parenchymal transection. IIVCC, infrahepatic inferior vena cava clamping; CVP, central venous pressure; IIVCP, infrahepatic inferior vena cava pressure; MAP, mean arterial pressure.

*

Student's t test.

The mean transection-related blood loss in patients with moderate to severe cirrhosis was significantly lower in the IIVCC group than in the low CVP group (2·9 versus 6·1 ml/cm²; P < 0·001). There was no significant difference in ALT and AST levels between the two subgroups (data not shown). Postoperative changes in bilirubin, prothrombin activity and creatinine level are shown in Fig. 2. There was no significant difference in morbidity rate (P = 0·502) or hospital stay (P = 0·376) between the two subgroups.

Fig. 2

Comparison of a transection-related blood loss (TRBL), b postoperative total bilirubin, c creatinine and d prothrombin activity in patients with moderate or severe cirrhosis in the infrahepatic inferior vena cava clamping (IIVCC) and low central venous pressure (CVP) groups. In a values are presented as median (horizontal line within box), interquartile range (box) and range (error bars). Serial measurements in b–d are presented as mean(s.e.m.). *P < 0·050, $P<0\cdot 010$  versus low CVP (Student's t test)

Discussion

Control of bleeding remains key to successful hepatic resection. Bleeding, together with subsequent blood transfusion, increases postoperative morbidity and mortality rates^11–13, particularly in patients with cirrhosis^14–18.

It is generally considered advisable to use hepatic vascular occlusion manoeuvres in complex hepatectomies to reduce the risk of excessive intraoperative bleeding. Although most liver surgeons agree that PTC is not necessary as a routine, it is still used widely and allows many hepatic resections to be carried out without any blood transfusion. The maximum tolerated duration of continuous normothermic ischaemia is generally considered to be 60 min for non-cirrhotic livers. Intermittent vascular occlusion with clamp/unclamp cycles of 15/5 min is safe even in cirrhotic liver resection^16–18. To allow a bigger safety margin, intermittent vascular occlusion with cycles of 10/5 min for cirrhotic liver resection, and continuous PTC of up to 30 min for non-cirrhotic liver resection, have been used in the authors' institution. However, occlusion of the portal triad does not prevent backflow bleeding from hepatic veins. THVE, originally described by Heaney and colleagues¹⁹, effectively suppresses all inflow and backflow bleeding. However, it is associated with significant haemodynamic changes and requires close invasive monitoring and anaesthetic expertise. Because of the difficulties of dissecting the IVC and major hepatic veins, and haemodynamic intolerance in some patients, THVE is not used widely.

When inflow hepatic occlusion is performed, blood loss during liver transection comes primarily from the hepatic veins and the sinusoids during parenchymal transection. The rate of bleeding is influenced by two factors: the diameter of the vessels and the pressure gradient. The latter is determined mainly by the CVP level. Hepatectomy carried out under low CVP (0–5 cmH₂O) can decrease blood loss effectively, and it is used routinely for hepatectomy in many centres^20,21. Placing a vascular clamp on the infrahepatic IVC above the right renal vein also reduces the CVP²² because IIVCC and PTC prevent almost 70 per cent of venous return²³. A previous study from the authors' institution showed that PTC combined with IIVCC is safe, effective and feasible⁶. Dissecting and placing a tourniquet around the infrahepatic IVC is technically straightforward^24–26.

In this randomized clinical study, total blood loss and transection-related blood loss were significantly less in the IIVCC group than in the low CVP group. This can be explained in two ways. First, IIVCC decreases CVP and controls venous backflow bleeding. Second, there was a high proportion (82·3 per cent) of patients with cirrhosis in this study. Compared with low CVP, IIVCC caused fewer haemodynamic disturbances. In contrast to what can be expected when THVE is used^3,27, all patients, including those with severe hepatic cirrhosis, were able to tolerate PTC with IIVCC. A possible explanation is that some blood can still go into the liver from the subphrenic veins, adrenal veins and lumbar veins during vascular clamping. Another possibility is that adrenal hormones still enter the circulation as they obviously play an important role in controlling blood pressure²⁸.

In this study, liver function recovered more quickly in the IIVCC group than in the low CVP group, which could be explained by greater impairment of hepatic metabolism as a consequence of the longer period of low CVP in the latter group. The impairment of renal function due to IIVCC was temporary and slight, in accordance with the recovery of creatinine values from day 3 after surgery. Moreover, there was no morbidity or mortality related to IIVCC.

Multiple linear regression analysis showed the extent of hepatic cirrhosis was positively correlated with the level of CVP and IIVCP, which could be explained by water–sodium retention in cirrhosis. IIVCC was more effective than low CVP in controlling bleeding during complex hepatectomy in patients with moderate to severe cirrhosis probably because of the more effective control of IVC pressure using IIVCC. An unexpected and somewhat surprising result was that the low CVP technique had a negative effect on bilirubin metabolism and coagulation, which probably resulted from a longer hypoperfusion time. Although the time for Pringle manoeuvre was similar in both groups, the hypoperfusion period was longer in the low CVP group because of the time taken for induction and recovery of the low CVP.

The study showed that IVC clamping gave better results than low CVP for a subgroup of patients with moderate to severe cirrhosis, and it was useful in controlling bleeding during complex hepatectomy. Even if the amount of blood lost was only slightly less in absolute terms, the present study provided evidence that IVC clamping is a useful adjunct to the available techniques of reducing blood loss during liver resection.

Acknowledgements

The authors thank Professor Y. Ping for the statistical analysis. This study was funded by grants from the China National Key Projects for Infectious Diseases (2008ZX10002-025), the Chinese Ministry of Public Health for Key Clinical Projects ([2010]439-51) and Hubei Province for the Clinical Medicine Research Centre of Hepatic Surgery (2008). Disclosure: The authors declare no conflict of interest.