Intraoperative blood loss is a major concern in liver surgery. In classical, open, hepatic resection, several surgical and anaesthesiological measures are known to limit intraoperative blood loss. The use of radiofrequency energy has been described to perform liver resection safely and with minimal blood loss. However, these results are based on few studies with small patient series, in which only open surgical procedures were performed [17, 27, 32, 37]. With improvements in technology and hepatobiliary surgery, laparoscopic liver resection has become feasible and safe. Although only few reports are available on laparoscopic liver resection (LLR), this type of resection may offer the advantages of minimally invasive surgery, provided there is adequate patient selection, surgical expertise, technological equipment, and infrastructure [1–11, 13, 14, 18, 21, 29, 35].

No data are available on the potential contribution of RF energy to the limitation of blood loss during laparoscopic liver resection. The aim of the present study was to investigate the role of radiofrequency energy to reduce intraoperative blood loss in patients undergoing laparoscopic liver resection.

Patients and methods

From October 2002 until September 2005, forty-five consecutive patients were enrolled in a prospective, nonrandomized study to undergo laparoscopic partial hepatectomy. The male/female ratio was 22/23, with a median age of 57 years (range 26–80 years). A laparoscopic procedure was considered in patients without severe systemic disease according to the American Society of Anaesthesiology (ASA score < 3). Patient selection was based on the judgment of the hepatobiliary surgeon. Radiofrequency energy was applied in 20 patients, whereas 25 patients underwent LLR without RF assistance. The surgeon decided at random whether or not to use RF for parenchymal transection. Fifty-eight hepatic tumors were resected (47 malignant, 11 benign), of which the majority were metastases from colorectal cancer (Table 1).

Table 1. Pathology of hepatic tumors
Full size table

Two hepatobiliary surgeons performed the laparoscopic procedures. In order to prevent the occurrence of clinically significant gas embolism, the intraabdominal pressure (CO2 pneumoperitoneum) was kept as low as possible, aiming for 6 to 8 mmHg during hepatic parenchymal transection. All procedures were assisted by laparoscopic ultrasound. In the RF-assisted LLR, the transection plane was precoagulated using radiofrequency energy. Parenchymal transection was performed through this precoagulated plane, moving ahead step by step, and allowing the direction of transection to be altered when necessary. The radiofrequency energy was applied during 1 to 2 minutes every 2 cm using a monopolar RF generator and a single cool-tip electrode (Tyco Healthcare, Radionics Europe NV). Parenchymal transection of the precoagulated liver tissue was accomplished with the harmonic scalpel and the ultrasonic aspirator. Blood vessels measuring less than 4 mm in diameter were controlled with bipolar diathermy. Major portal and hepatic veins were controlled with endoclips or endoscopic stapling devices. The resected specimen was extracted through a suprapubic incision by means of an endobag.

Patient, tumor, and surgery characteristics were comparable in patients undergoing hepatic resection with or without RF-assistance (p ≥ 0.1370), that is, gender, age, number and type of tumors (malignant vs benign), maximum tumor diameter, tumor location in right/left liver, presence/absence of cirrhosis, preoperative chemotherapy (yes/no), duration of hepatic vascular occlusion, other additional surgical procedures (yes/no), surgeon, and duration of surgery (Table 2). At the time of surgery, macroscopic invasion of the diaphragm (2) or a hepatic vein (2) was encountered equally in both groups. Cirrhosis was present in 5 patients who underwent LLR (segmental hepatectomy 4, left lateral lobectomy 1) for a hepatocellular carcinoma (HCC) with a median maximum diameter of 37 (20–45) mm. Preoperative chemotherapy was given in 12 out of 19 patients with colorectal liver metastases.

Table 2. Patient and tumor characteristics
Full size table

Any hepatic segment involved with a tumor was considered for the laparoscopic approach. The location of the liver tumors is shown in Table 3. Thirty-seven minor hepatectomies (resection of 2 segments or less) were performed, of which 22 were segmental resections (segmentectomy/metastasectomy) and 15 were left lateral lobectomies (LLL). Major hepatectomy (resection of 3 segments or more) consisted of 6 right hemihepatectomies (RHH), 2 left hemihepatectomies (LHH), and 1 central hepatectomy (Sg4-5-8). Eighteen patients underwent additional surgery simultaneously with laparoscopic liver resection (Table 3).

Table 3. Type of surgical procedure
Full size table

Analysis of the statistical significance of differences between groups of data was performed using two-tailed Fisher exact test or Wilcoxon / Kruskal-Wallis test, as appropriate. A p-value of ≤0.05 was considered statistically significant. All analyses were performed with the statistical software package JMP (version 5.1.2).

Results

No intra- or postoperative mortality occurred. The median duration of surgery was 115 (45–360) minutes. The median intraoperative blood loss was 200 ml (range 5–4000 ml) and was similar in patients undergoing hepatic resection with or without RF assistance (Table 4). The type of surgical procedure was a determinant for the amount of intraoperative blood loss (p = 0.0002). Laparoscopic right/left hemihepatectomy was associated with a median blood loss of 2000 ml, left lateral lobectomy with 200 ml, and segmental hepatectomy with 100 ml (Table 5). Bleeding was the only intraoperative complication and was observed in 3/25 (12%) of patients in the RF– group and in 7/20 (35%) patients in the RF+ group. In these 10 (22.2%) patients, bleeding occurred from large hepatic vessels and was encountered in 6 out of 9 (66.7%) patients who underwent a major hepatic resection, and in 4 out of 36 (11.1%) patients undergoing minor LLR (Table 6). No significant (Spearman) correlations were found among patient, tumor, and surgery-associated variables and the amount of intraoperative blood loss.

Table 4. Clinical outcome of laparoscopic liver resection in 45 patients
Full size table
Table 5. Type of surgical procedure and intraoperative blood loss
Full size table
Table 6. Analysis of intraoperative hemorrhage
Full size table

Median intraoperative blood loss in patients with cirrhosis and/or preoperative chemotherapy (RF+ in 8 vs RF– in 9 patients) was 150 (5–3000) ml as compared to 200 (10–4000) ml in patients without cirrhosis and/or preoperative chemotherapy (RF+ in 12 vs RF– in 16 patients). This difference was not statistically significant (p = 0.2175).

Pringle’s maneuver was performed in 6 patients for a median duration of 18 (7–40) minutes, that is, in one LLL, in one LLL combined with a segmental resection in Sg4-5, and in 4 patients undergoing segmental hepatectomy (respectively in Sg4-5, Sg5, Sg5-6 + Sg6-7, and Sg5 + Sg6-7). Segmental hepatectomy for a tumor involving Sg7 or Sg8 was performed in 3 and 2 patients, respectively. In one patient (Sg5+Sg6-7), Pringle’s maneuver was performed to control the hemorrhage. Median blood loss in these patients was 100 (10–1400) ml. Microscopic resection margins [8 (1–10) mm] were free of tumor in all patients.

In patients who underwent a right (n = 6) or left (2) hemihepatectomy, the inflow vessels were controlled selectively prior to parenchymal transection, without a Pringle’s maneuver. The draining hepatic veins were divided with an endoscopic vascular stapling device after completing the parenchymal transection. The site of significant hemorrhage in patients undergoing a major LLR was from a hepatic vein in 4 patients, diaphragmatic vein in 1, and portal vein in 1. During minor LLR, hemorrhage occurred from a hepatic vein in 3 and from a portal vein in 1 patient.

Conversion from a laparoscopic procedure to an open hepatic resection was necessary in 3 patients because of uncontrollable intraoperative bleeding from the right hepatic vein (at RHH), the middle hepatic vein (at segmental hepatectomy Sg4a), and the right portal vein (at RHH) (Table 6).

Postoperative complications were observed in 2 patients, that is, 1 biliary leakage with intraabdominal abscess formation and 1 pulmonary infection. The median length of hospital stay (LOS) was 7 days following LLR with and without RF assistance. In all resection specimens, the margins were free of tumor on macroscopic and pathologic examination. The magnitude of the tumor-free margin [median 10 (1–30) mm] was comparable after LLR with or without RF-assistance, less than a centimeter in 16 and greater than a centimeter in 29 patients.

Discussion

In open liver surgery, major blood loss implicates the risk of higher morbidity and mortality as well as shorter long-term survival [17, 27, 32, 37]. In order to reduce parenchymal bleeding during the actual transection phase in open liver surgery, several measures have been described, including radiofrequency energy to create a “zone of dessication” prior to performing hepatic transection. The potential advantages of radiofrequency-assisted hepatectomy might be a “virtually bloodless” procedure, shorter operation time, and reduced morbidity [17, 27, 32, 37]. The RF technique cannot be used close to the liver hilum, since heat might traumatize biliary structures with subsequent bile leakage and/or abscess formation [27, 28, 32, 37]. The RF technique is also unable to control blood vessels measuring larger than 4 mm in diameter, such as hepatic or portal vein branches [24]. The major shortcomings of previously published data are that these studies primarily cover small patient series, and often information on the number or size of the liver tumors, as well as the exact type of hepatic resection, is omitted or is not clearly specified [17, 32, 37].

The implementation of radiofrequency-assisted hepatectomy in laparoscopy was assumed possibly to represent a considerable step in making laparoscopic liver resection safer and more feasible [30, 37, 38]. However, this hypothesis is supported by few case reports [9, 38]. The present study is the first to report laparoscopic liver resection prospectively in a substantial number of patients over a relatively short period of time, and to evaluate the contribution of radiofrequency energy to the amount of blood loss during minor and major laparoscopic hepatectomy.

In the present study, laparoscopic liver resection indeed appears safe and feasible, without mortality, and with a morbidity rate of 24% (11/45). With a median blood loss of 200 ml, an operative duration of 113 minutes, and a length of hospital stay of 7 days, the short-term results concur with those reported in literature [1, 2, 4, 13, 15, 18, 25, 29, 31, 35, 38].

Losing control on blood loss is obviously the most important concern for all authors reporting on LLR [1, 2, 4, 9, 13, 15, 19, 22, 25, 30, 35, 38]. The only intraoperative complication retained in the present study was bleeding that was observed in 66.7% (6/9) of the major LLR and in 11.1% (4/36) of the minor resections. Significant blood loss occurred from a hepatic vein in 7 patients, from a portal vein branch in 2, and from a diaphragmatic vein in 1 patient. It is clear that this type of bleeding can neither be prevented nor managed by using RF energy. On the other hand, the use of Pringle’s maneuver would only have been beneficial in the 2 cases with hemorrhage from the portal vein, since this maneuver does not control blood loss from the draining hepatic veins. Indeed, this high percentage of significant intraoperative bleeding indicates that laparoscopic major hepatectomy needs thorough evaluation before it can be implemented into daily practice.

The application of RF energy during parenchymal transection in LLR was not associated with reduced intraoperative blood loss, operation time, or complication rate. Only the type of surgical procedure appeared to determine significantly the amount of blood loss. No other significant correlations were found between the analyzed patient, tumor- and surgery-associated variables, and the amount of intraoperative blood loss. Intraoperative blood loss in patients with a “normal liver” was comparable to blood loss in patients with cirrhosis or with preoperative chemotherapy.

Patient selection, experience in hepatobiliary surgery, adequate technology, and infrastructure are vital in defining the safety, feasibility, and outcome of laparoscopic liver resection. In the present study, patient selection and the use of RF energy were based on the surgeon’s judgment in a nonrandomized fashion, suggesting a potential bias. However, it is a difficult task to randomize a large number of patients with similar characteristics, who are suffering from liver tumors, to undergo comparable surgical procedures. All LLRs were performed by surgeons with experience in both hepatobiliary and advanced laparoscopic surgery as well as in hepatic radiofrequency application. Although patients with severe systemic disease were excluded, any patient can be eligible for LLR as long as they can endure an advanced laparoscopic procedure. Anterior and left lateral hepatic segments and tumors smaller than 5 cm seem to be most suitable for LLR [1, 2, 4, 13, 15, 18, 25, 29, 35, 38]. However, tumor size and localization don’t have to be a contraindication for LLR as such, provided the tumor isn’t in close relationship to the portal bifurcation or the suprahepatic junction, and provided a sufficient tumor-free margin can be warranted. In the present study, any hepatic segment involved by tumor was considered for laparoscopic approach. Benign lesions were judged to be an indication for LLR when they where symptomatic, a suspected adenoma, or of uncertain histopathologic diagnosis. Malignant lesions eligible for LLR were HCC in cirrhosis and liver metastasis, mostly colorectal in origin. One gallbladder cancer in situ and 4 intrahepatic cholangiocarcinomas were resected. Five patients (11%) suffered from cirrhosis. Several reports on LLR suggest that this population could benefit from minimally invasive surgery, as it seems to mean significant reduction in perioperative blood loss and surgical trauma as compared to open surgery [1, 4, 7, 11, 15, 21, 26, 35]. In order to prevent cancer recurrence caused by technical failure, the same oncologic rules should be respected for LLR as for open hepatectomy [1, 2, 4, 11, 15, 18, 21, 25, 29, 34–36]. In the present study, the resection margins were tumor-free on macroscopic and pathologic examination in all resected specimens. The LLR was accomplished according to the “no touch” technique, without hand-assistance. The resected liver tissue was extracted using a specimen bag, avoiding direct contact of the tumor with the abdominal wall.

As far as postoperative outcome is concerned, the complications observed in the present study are comparable with those most frequently reported in literature [4, 10, 15, 20, 25, 29, 35]. The same remark can be made for the conversion rate, as the reported conversion rates vary between 0 and 15.4% [1, 2, 5, 7, 9, 15, 21, 25, 29]. The authors believe that conversion of the LLR to an open procedure must be acknowledged as a safety measure and not as a technical failure. Major intraoperative bleeding during LLR will always remain difficult to control laparoscopically, even with extensive surgical experience. Therefore, a low threshold for conversion to open hepatic resection is indicated.

Conclusions

Laparoscopic liver resection is feasible and safe, provided that adequate equipment and extensive experience of the surgical team in both open liver surgery and advanced laparoscopy are available. LLR can be performed with minimal amounts of intraoperative blood loss, which is determined by the type of hepatectomy. Significant intraoperative hemorrhage occurs from large hepatic vessels during major LLR. The high percentage of significant bleeding that is observed at major hepatectomy indicates that the role of laparoscopy in major liver resections needs further evaluation. Radiofrequency-assisted parenchymal transection in LLR does not seem to reduce blood loss, operation time, or perioperative morbidity. Therefore, RF-assisted LLR seems not to be advocated as the preferred surgical technique in laparoscopic hepatectomy, also taking into account the additional costs of the RF equipment. The present, prospective study, however, only analyzes the short-term outcome and has a nonrandomized nature. The impact of laparoscopic surgery on the long-term outcome of patients with hepatic tumors has yet to be evaluated.