Journal of Hepato-Biliary-Pancreatic Surgery

, 16:310

Laparoscopic liver resection facilitates salvage liver transplantation for hepatocellular carcinoma


  • Alexis Laurent
    • Liver Transplantation and Hepatobiliary Unit, Service de Chirurgie Digestive et HépatobiliaireHôpital Henri Mondor—Université Paris 12
  • Claude Tayar
    • Liver Transplantation and Hepatobiliary Unit, Service de Chirurgie Digestive et HépatobiliaireHôpital Henri Mondor—Université Paris 12
  • Marion Andréoletti
    • Service d’Anesthésie-Réanimation Chirurgicale-SAMU 94Hôpital Henri Mondor—Université Paris 12
  • Jean-Yves Lauzet
    • Service d’Anesthésie-Réanimation Chirurgicale-SAMU 94Hôpital Henri Mondor—Université Paris 12
  • Jean-Claude Merle
    • Service d’Anesthésie-Réanimation Chirurgicale-SAMU 94Hôpital Henri Mondor—Université Paris 12
    • Liver Transplantation and Hepatobiliary Unit, Service de Chirurgie Digestive et HépatobiliaireHôpital Henri Mondor—Université Paris 12
    • Department of Digestive SurgeryHôpital Henri Mondor
Original article

DOI: 10.1007/s00534-009-0063-0

Cite this article as:
Laurent, A., Tayar, C., Andréoletti, M. et al. J Hepatobiliary Pancreat Surg (2009) 16: 310. doi:10.1007/s00534-009-0063-0



In patients with hepatocellular carcinoma (HCC), a previous liver resection (LR) may compromise subsequent liver transplantation (LT) by creating adhesions and increasing surgical difficulty. Initial laparoscopic LR (LLR) may reduce such technical consequences, but its effect on subsequent LT has not been reported. We report the operative results of LT after laparoscopic or open liver resection (OLR).


Twenty-four LT were performed, 12 following prior LLR and 12 following prior OLR. The LT was performed using preservation of the inferior vein cava. Indication for the LT was recurrent HCC in 19 cases (salvage LT), while five patients were listed for LT and underwent resection as a neoadjuvant procedure (bridge resection).


In the LLR group, absence of adhesions was associated with straightforward access to the liver in all cases. In the OLR group, 11 patients required long and hemorrhagic dissection. Median durations of the hepatectomy phase and whole LT were 2.5 and 6.2 h, and 4.5 and 8.3 h in the LLR and OLR groups, respectively (P < 0.05). Median blood loss was 1200 ml and 2300 ml in the LLR and OLR groups, respectively (P < 0.05). Median transfusions of hepatectomy phase and whole LT were 0 and 3 U, and 2 and 6 U, respectively (P < 0.05). There were no postoperative deaths.


In our study, LLR facilitated the LT procedure as compared with OLR in terms of reduced operative time, blood loss and transfusion requirements. We conclude that LLR should be preferred over OLR when feasible in potential transplant candidates.


Hepatocelullar carcinomaLaparoscopic liver resectionLiver transplantationOpen liver resectionSavage liver transplantation


Liver transplantation (LT) is considered the treatment of choice for early hepatocellular carcinoma (HCC) complicating chronic liver disease (CLD). Current clinical strategies are to submit patients with CLD to screening for HCC by serial ultrasound. This diagnostic step enables an increasing number of patients with asymptomatic small HCCs to benefit from transplantation [15]. However, LT cannot be offered on a large scale to HCC patients because of donor shortage and, therefore, only a very small portion of these patients can be listed. Of those listed, many may drop out because of tumor progression due to long waiting times. Alternative options to LT in patients with small HCCs include percutaneous ablation and liver resection (LR). The latter option, although efficient, has been limited by two factors: (1) it is associated with significant postoperative morbidity and mortality in patients with CLD and (2) it is an extensive procedure which may compromise subsequent LT. In that respect, laparoscopic liver resection (LLR) could address these two limiting factors.

The aim of this study was to compare the intraoperative results of LT for HCC after laparoscopic or open liver resection (OLR).

Patients and methods


From March 1995 to May 2007, 186 patients underwent LT for HCC complicating CLD. Among these, 24 (13%) of the LT patients had prior liver resection involving 12 LLR and 12 OLR.

Liver resections

Initial liver resections were performed in patients with compensated CLD (i.e. Child’s class A). Resections were performed either with a curative intent followed by LT in the case of recurrence (salvage LT) or as a neoadjuvant treatment while waiting for LT (bridge LR).

All resections were performed by laparotomy prior to 1998. Laparoscopic resection for HCC was introduced into clinic practice at our hospital in 1998, and since that date 61 procedures have been performed. Since the introduction of LLR, the choice of open or laparoscopic approach for resection of HCC is decided upon on an individual patient basis, with LLR being proposed primarily in patients with peripheral tumors <5 cm. The surgical techniques for laparoscopic and open resections have been described previously [6, 7].

Briefly, open resections were performed through a right subcostal incision or a bilateral subcostal incision with upper midline extension. The LLR were performed through five upper abdominal ports of 5–12 mm each and the specimen retrieved through a 4- to 8-cm-long lower abdominal incision (supra pubic or right iliac fossa).

Liver resections were intended to be anatomical (i.e. resection of one or several anatomical segments) in order to resect the portal territory of the tumor. Segmentectomies were defined by their numbers and performed according to selective clamping, external segmental borders and use of intraoperative ultrasound, as previously described [7]. However, in small peripheral lesions developed in patients with portal hypertension and hepatic anatomic deformation (atrophy–hypertrophy), non-anatomical wedge resections were performed, especially in the case of bridge resections. These consisted of resection of the tumor and an intended 1-cm tumor-free margin.

Liver transplantation

In all cases, a bilateral subcostal incision with an upper midline extension was performed. In the OLR group, the previous incision was reused and usually extended. A fresh incision was created in the LLR group.

Liver transplantation was performed using preservation of the inferior vein cava (IVC) and temporary portacaval anastomosis [8] except when IVC resection was necessary. Reconstruction was performed in the piggyback fashion with end-to-side cava-caval anastomosis on the joined stump of the three main hepatic veins. Portal, arterial and biliary reconstructions were performed using standard techniques.

Studied criteria

Main outcome measures included intraoperative data (duration of surgery, blood loss and blood transfusions). Secondary outcome measures included mortality and morbidity.

In terms of reducing intra-abdominal adhesions, a previous laparoscopy is considered to be the preferred procedure compared to a previous laparotomy. Therefore, we paid special attention to the study of the initial dissection phase of the LT procedure (hepatectomy phase) and as well as comparing the whole procedure, we also compared the hepatectomy phase, corresponding to the period between the incision and the end of the hepatectomy.


Categorical variables were compared using the chi-square test, and the Mann–Whitney U test was used for noncategorical variables. Survival rates were estimated using actuarial methods.


The patient cohort consisted of 18 men and six women. Cirrhosis was the result of hepatitis B or C chronic infection in 14 cases and alcohol abuse or steatohepatitis in ten cases.

Hepatic resection

The types of laparoscopic and open liver resections are summarized in Table 1.
Table 1

Types of laparoscopic and open resections

Types of LR



Segmentectomy 6


Bisegmentectomy 6 + 7


Left lateral sectionectomy


2 (2)a

Central hepatectomy


Right hepatectomy



Wedge resection


5 (2)

LR Liver resection, OLR open LR, LLR laparoscopic LR

aNumbers in parentheses in the open group are patients operated on before 1998 who would have been considered for laparoscopy after that date

Median age at resection was 54.3 years (range 40–66 years). Resections were performed in curative intention in 19 cases and as a bridge to LT in five cases. All but one patient were Child A (MELD < 15). There were ten nonanatomical resections and 14 anatomical resections, including six left lateral sectionectomies, one bisegmentectomy, two monosegmentectomies, four right hepatectomies and one central hepatectomy. In the LLR group, all resections were completed laparoscopically without conversion. None of the patients in the open group originated from a previously converted laparoscopy. There were no bile leaks or subphrenic abscesses after initial open or laparoscopic resection. In addition, there was a lower production of ascites in the LLR group than in the OLR group (one vs. four).

Recurrence after hepatic resection

Nineteen patients developed intrahepatic recurrence, none in the vicinity of the initial resection. Median time to recurrence after liver resection was 16 months (range 2–88 months). These patients were listed for salvage LT. The remaining five patients were listed before liver resection, which was performed as a neoadjuvant treatment because of the expected long waiting times and a high risk of tumor progression and drop out. These procedures were referred to as bridge resections.

Orthotopic liver transplantation

Median age at transplantation was 55.6 years (range 43–69 years). Twenty patients received a full graft and four patients received a partial graft (two right lobes from split cadaveric grafts and two right livers from living donors). For the 19 patients transplanted for HCC recurrence (salvage transplantation), the median delay between HCC recurrence and LT was 10 months (range 3–91 months); for the five remaining patients (bridge resection), the median delay between liver resection and LT was 18 months (range 7–40 months). Table 2 presents the data on patients undergoing liver transplantation.
Table 2

Liver transplantation features





Delay between LR and LT (months)


29 ± 62

38 ± 28


  Median (range)

28 (7–94)

27 (14–104)





Graft: split/LDLT/full graft




LT (overall)

  Duration (h)


6.4 ± 1.7

7.9 ± 1.8


    Median (range)

6.2 (4.1–9)

8.5 (4.5–9.4)

  Blood loss (ml)


1,500 ± 760

2,100 ± 740


    Median (range)

1,200 (400–2700)

2,300 (1,000–3,200)

  Units of red cells transfused


2.9 ± 3.5

6.1 ± 6.9


    Median (range)

3 (0–11)


LT phase 1

  Duration (min)



150 ± 52

247 ± 71


    Median (range)

150 (80–250)

270 (60–315)

  Units of red cells transfused


0.1 ± 0.3

3.4 ± 3.5


    Median (range)

0 (0–1)

2 (0–11)

Early surgical post op complications






  Early biliary stenosis



  Incisional hernia



LT Liver transplantation, LDLT living donor LT, phase 1 hepatectomy phase of LT, ns not significant

The median LT duration was 6.2 (range 4.1–9) and 8.5 h (range 4.5–9.4 h) in the LLR and OLR groups, respectively (P < 0.05).

The median blood loss was 1200 (range 400–2700) and 2300 ml (range 1000–3200 ml) in the LLR and in the OLR groups, respectively (P < 0.05). Median total transfusion of red cells transfusion were 3 (range 0–11) and 6 U (range 0–21 U) in the LLR and OLR groups, respectively (P < 0.05).

In the LLR group, access to the liver was straightforward in all cases, meaning that there were no adhesions between intra-abdominal organs and abdominal wall. In contrast, all but one patient in the OLR group had developed significant peritoneal adhesions below and around the incision. These adhesions required significant dissection, often tedious and hemorrhagic, increased by the presence of portal hypertension. Adhesions between the cut surface and the surrounding structures were present in all cases. These adhesions concerned mainly the omentum, the diaphragm, the colon and/or the stomach, depending on the type and side of resection. In the LLR group, such adhesions were much easier to dissect than comparable open procedures, except in one case of dense hypervascular omental adhesions after a segmentectomy 5.

The median duration of the hepatectomy phase was 150 (range 80–250) and 270 min (range 60–315 min) in the LLR and OLR groups, respectively (P < 0.05). During the hepatectomy phase, the median number of red cell units transfused were 0 (range 0–1) and 2 U (range 0–11 U) in the LLR and OLR groups, respectively (P < 0.05). The IVC was preserved in ten of 12 cases in the LLR group, and it was resected by principle in one case because the recurrent tumor was close to the IVC. In the OLR group, IVC preservation was performed in ten of 12 cases; in the two other cases, it was not possible to preserve the IVC because of adhesions due to previous liver resection.

Postoperative course

No patient died. Six patients developed seven postoperative surgical complications (Table 2). There were no differences between LLR and OLR patients in terms of hospital stay, medical complications and postoperative blood transfusion.


This is the first series reporting LT after LLR for HCC in CLD. In an earlier publication, we reported that LLR for HCC in CLD can be performed without mortality in selected patients [6]. The results from the present study confirm the safety of liver transplantation after open or laparoscopic liver resection, with no mortality in 24 consecutive patients, and demonstrate that the expected technical facilitation of subsequent LT is an additional advantage of LLR over OLR [6, 9, 10].

Liver transplantation after a previous laparotomy for upper abdominal procedure is often much more challenging and associated with longer operating times and a higher blood loss [1113]. The main data gained from our study is that the hepatectomy phase of the transplant procedure was facilitated in the LLR group in whom the absence of adhesions gave straightforward access to the liver in all cases. In contrast, all but one patient in the OLR group required tedious and hemorrhagic dissection of adhesions to gain access to the operative site. As a consequence, the median duration of the hepatectomy phase duration was almost 1 h shorter in the LLR group compared to the OLR group, and median transfusions during the hepatectomy phase were significantly lower in the former than in the latter (0 vs 2 U). These differences obtained during the hepatectomy phase influenced the whole procedure, with reduced overall operative time, blood loss and transfusion requirements in the LLR group compared to the OLR group.

Another interesting feature of this study is the absence of postoperative mortality of LT patients after both open and laparoscopic resections and identical morbidity in both groups, supporting liver resection as an important option in HCC even in potential LT candidates. Until recently, most authors have been reluctant to propose liver resection for HCC in potential LT candidates, mostly because it would compromise subsequent LT. However, donor shortage and improved results of resection have led to regained interest in liver resection [1416]. The results from our study confirm the value of initial resection with a special mention to LLR where applicable.

Current management of HCC in patients with CLD must be addressed with the intention of treatment depending on whether of not LT is an option. While LT is the best theoretical treatment for patients with early HCC (Milan criteria), donor shortage leads to increased waiting times and tumor progression, which in turn may lead to potential recipients dropping out the waiting list. Therefore, most authors recommend some kind of neoadjuvant therapy to control tumor progression and reduce the risk of drop out. Possible treatments include transcatheter arterial chemoembolization (TACE), radiofrequency ablation (RFA) and liver resection. In our unit, we consider these options complementary rather than competitive. Transcatheter arterial chemoembolization is proposed for multiple lesions or those inaccessible to resection or ablation, while RFA is proposed for lesions located deep within the liver; liver resection is considered the best option for peripheral lesions. In this latter situation, the laparoscopic approach has been associated with excellent safety and, as shown in our present study, has no technical impact on subsequent LT. Our selection criteria for LLR depend on the location of the lesions, with peripheral lesions being the best candidates [6]. In addition, liver resection offers the best tumor control. In our series, we observed no recurrence during the first 6 months, which is in contrast with the high rates of presence of residual or recurrent active tumor in RFA-treated patients, as shown on hepatectomy specimens from transplanted patients [17].

In other cases, LT may not be the first option chosen, particularly for patients with a solitary tumor >5 cm. In such cases, we favor liver resection with a curative intent. In some of these patients, initial resection may be considered a downstaging procedure, allowing salvage transplantation in case of delayed and limited recurrences within the Milan criteria.

An additional advantage of liver resection is the possibility of a full pathologic study of the specimen, which permits the assessment of prognostic factors of HCC, such as microvascular invasion, satellite nodules, tumor differentiation, and molecular biology markers of recurrence, all of which may participate in the transplant decision-making process [1821] and lead to the concept of bridge resection and preemptive LT [10, 21, 22].

In conclusion, this study demonstrates that initial LLR facilitates subsequent LT as compared with OLR, with reduced operative time, blood loss and transfusion requirements. Our results also shows the safety of salvage liver transplantation after both open and laparoscopic resection. There are now several arguments in favor of liver resection prior to LT. In this setting, our results support the use of LLR whenever possible.

Copyright information

© Springer 2009