Hepatic resection using ultrasonic surgical aspirator

The purpose of this study was to examine, retrospectively, results of hepatic resection performed by the first author from April 1999 to March 2013 since the introduction of CUSA into our institution. Data collection included demographic data, extent of hepatic resection, duration of operation, blood transfusion, final pathological reports, postoperative complications, and perioperative mortality. Factors associated with perioperative mortality were also analyzed.

This study was approved by the Institutional Review Board of the Faculty of Medicine, Chulalongkorn University.

During the study period, the indications for hepatic resection were: (1) hepatocellular carcinoma, (2) intrahepatic and hilar cholangiocarcinoma, (3) metastatic cancer to the liver, (4) carcinoma of the gall bladder, (5) other benign and malignant tumors of the liver, (6) hepatic necrosis following hepatic trauma, and (7) recurrent cholangitis from choledochal cysts or other benign strictures of the bile duct. With the exception of 3 patients who underwent emergency hepatic resection, patients underwent hepatic resection on an elective basis when the following criteria were fulfilled: (1) acceptable anesthetic risks, (2) appropriate indication for surgery, and (3) estimated adequacy of hepatic reserve after resection by preoperative CT scan. No patient in this study underwent preoperative portal vein embolization. In patients who had hilar cholangiocarcinoma with marked jaundice, preoperative biliary drainage was performed by either an endoscopic or percutaneous method to achieve a serum bilirubin level of <2 mg/dL. The three emergency hepatic resections were performed in patients with a ruptured hepatocellular carcinoma resulting in severe ongoing hemorrhage. Operations were conducted after careful preoperative evaluation of the computed tomography (CT) and/or magnetic resonance imaging (MRI), and conferencing between surgeons and radiologists regarding extent of the lesions, their resectability, and the adequacy of the hepatic remnant. Patients who underwent hepatic resection during emergency laparotomy for trauma were excluded from this study.

Operations were performed in the same manner in most cases and included the following steps: (1) bilateral subcostal incision with midline extension to the xiphoid cartilage, (2) extensive mobilization of the targeted hepatic lobe from the diaphragm and the posterior abdominal wall, (3) cholecystectomy, (4) dissection of hepatic hilum, isolation of the left or right hepatic artery and left or right branch of the portal vein, (5) ligation of the left or right hepatic artery supplying the resected hepatic segments, (6) ligation of left or right branch of the portal vein supplying the targeted hepatic segments, (7) identification of line of demarcation on the liver surface, (8) parenchymal transection along the line of demarcation by using CUSA and intermittent hepatic inflow occlusion (Pringle maneuver), (9) ligation and division of the right or middle or left hepatic vein according to type and extent of hepatic resection, (10) division of left or right hepatic duct of the hepatic segments to be removed, (11) complete hemostasis with electrocoagulation on the raw surface of the hepatic remnant, and (12) placement of 2 vacuum drains in the space between the raw surfaces of the resected hepatic segments, the diaphragm, and the posterior abdominal wall.

The above-mentioned steps in hepatic resection were used in right hepatectomy (removal of segment V, VI, VII, and VIII) or left hepatectomy (removal of segment II, III, and IV). These techniques were modified as necessary. With very large tumors, performing right or left hepatectomy were changed owing to anatomic distortion of the bulging tumor making early mobilization dangerous. Extensive mobilization of the liver was not conducted as an early step; instead, ligation of the corresponding hepatic artery and portal vein branches were conducted first to decrease blood supply to the tumor causing shrinkage of the tumor and the respective hepatic segments. Subsequent mobilization of the tumor and the liver could then be performed more easily after deprivation of the blood supply to the liver mass.

Operative procedures were classified as follows: wedge resection (nonanatomic resection of limited normal hepatic parenchyma surrounding the lesion), segmentectomy (removal of one or more hepatic segments from segment I to VIII), right hepatectomy (removal of segments V, VI, VII, and VIII), left hepatectomy (removal of segments II, III, and IV), and right lobectomy or extended right hepatectomy (removal of segments IV, V, VI, VII, and VIII).

Postoperatively, patients were admitted in our surgical intensive care unit (SICU) for 24 hours and then to the ward after stabilizing in the SICU. Special attention was paid to postoperative hemorrhage, which was associated with high mortality if not diagnosed early. Immediate reoperation was conducted when postoperative intra-abdominal hemorrhage was strongly suspected or diagnosed.

Perioperative mortality was defined as death within the hospital after hepatic resection or death resulting from postoperative complications. Causes of perioperative mortality were analyzed. Results were described as range, median, and mean ± SD. Statistical analyses of significance of variables associated with mortality were made using a Student t test and a Chi square test when appropriate. P < 0.05 was considered significant.

During the 14-year retrospective period, 101 patients with 104 hepatectomies were included into the study. Three patients had 2 hepatectomies performed at different times. Fifty-nine patients were men and 42 were women. Their age ranged from 28 to 81 years (median 56 years, mean 55.1 ± 11.5 years). Patients’ pathological diagnoses were hepatocellular carcinoma 32, hilar cholangiocarcinoma 25, peripheral cholangiocarcinoma 17, metastatic carcinoma to the liver 11, carcinoma of the gall bladder 5, cystadenocarcinoma of the liver 1, and other benign pathologies 10 (Table 1). Seventy-two resections (69%) were composed of tumors ranging from 1 to 25 cm in diameter (median 6 cm, mean 7.02 ± 5.52 cm). In the remaining 32 resections (31%), the size of lesions were not reported because of the infiltrative nature of the tumors, nonmass forming pathology, or incomplete data in some patients. Of the 32 patients who had hepatocellular carcinoma (34 resections), 21 (66%) had cirrhosis, 10 (31%) were carriers of hepatitis B virus, and 1 (3%) was a carrier of hepatitis C virus. Two patients, aged 75 and 80 years old, had hilar cholangiocarcinoma and cirrhotic livers. Both underwent right hepatectomy with biliary-enteric reconstruction by Roux-en-Y left hepaticojejunostomy. One (aged 75) recovered well, the other (aged 80) died from multisystem organ failure 12 days postoperatively.

Table 1

Diagnoses

Operations were conducted on an elective basis in all except 3 patients, who had emergency hepatic resection because of a ruptured hepatocellular carcinoma. Right hepatectomy was the most common operative procedure, followed in order by left hepatectomy, right hepatectomy and segment I resection, right lobectomy, segment II and III resection, segment VI resection, wedge resection, and other combinations of resections (Table 2). Biliary-enteric reconstruction was performed after right or left hepatectomy or right lobectomy in 28 patients. The biliary-enteric reconstruction was performed using a 40–60 cm long Roux-en-Y limb of the jejunum anastomosed to the left or right hepatic duct after hepatic resection. In 25 patients who had hilar cholangiocarcinoma, additional resection of the extrahepatic bile duct and lymphadenectomy of the hepatoduodenal ligament were performed concurrently with the hepatic resection before reconstruction of the biliary-enteric anastomosis.

Table 2

Operative procedures

The operative time ranged from 90 to 720 min (median 300 min, mean 327 ± 149 min). Operative blood transfusion ranged from 0 to 17 units (median 3 units, mean 3.9 ± 3.6 units). Twenty-one hepatectomies (20%) were conducted without blood transfusion. Intermittent hepatic inflow occlusion (the Pringle maneuver) was performed during hepatic parenchymal transection in 93 resections (89%). The ischemic time during intermittent hepatic inflow occlusion ranged from 5 to 105 min (median 40 min, mean 43.29 ± 20.72 min). CUSA was used during hepatic parenchymal transection in 90 resections (87%). In 8 resections (8%), the hepatic parenchymal transection was performed by using water jet. No assisting device for hepatic parenchymal transection was used in 6 resections (6%).

Thirty-four complications occurred in 30 patients (29%). Details of complications are shown in Table 3. Reoperation was performed in 9 patients (reoperation rate 9%). The causes for reoperation were intra-abdominal bleeding 5, bleeding duodenal ulcer 1, small bowel obstruction 1, intra-abdominal sepsis 1, and biliary obstruction 1 (Table 3). Six patients in this study died (Table 4). The perioperative mortality was 6%. Univariate analyses of factors associated with mortality conducted using a Student t test or a Chi square test revealed a significant difference in age, operative time, operative blood transfusion, reoperation, and complication in patients who had perioperative mortality compared with patients who survived operations (Table 5).

Table 3

Complications after hepatic resection

Table 4

Perioperative mortality

Table 5

Univariate analyses of factors associated with mortality

Hepatic resection is a major operation with potentially disastrous complications and fatal outcomes. Although hepatectomies are indicated for various benign and malignant conditions of the liver, they are conducted mainly in patients with primary or metastatic tumors of the liver. For resection of malignant tumors, an adequate tumor free margin is important for long-term survival. This may be very difficult to achieve. Such situations include a large tumor close to major vascular structures, hilar cholangiocarcinoma, and limited hepatic reserve from cirrhosis or a small hepatic remnant after resection. Three of 6 our patients who succumbed perioperatively had underlying cirrhosis and subsequent hepatic failure. Although the significance of liver cirrhosis and mortality could not be demonstrated statistically in this study, surgery on cirrhotic patients is dangerous and such a decision requires careful preoperative analysis. Two of our patients died from multisystem organ failure after massive blood loss during and after surgery, which was caused by unexpected difficulties and technical factors during surgery. Such fatal bleeding complications emphasized the importance of careful and faultless surgical technique. Perioperative hemorrhage is well known to be associated with hepatic resections [12-14]. However, current outcomes of hepatic resection have been greatly improved and mortality has decreased in many high-volume centers [6-9]. However, such mortality still exists when the complexity of the operations conducted is augmented by underlying cirrhosis and extensive pathology. Prolonged operative time and intraoperative blood transfusion have been shown in the current study to be significantly associated with perioperative mortality. Experience of the surgical team is unquestionably a major factor contributing to satisfactory outcomes, especially when dealing with complex liver surgery [9].

The safety of major hepatic resection has been enhanced by development of innovations, i.e. intermittent hepatic inflow occlusion (Pringle maneuver) and hepatic parenchymal transection assisting devices. In 1908, J. Hogarth Pringle from Glasgow wrote a classical article “Notes on the arrest of hepatic hemorrhage due to trauma” published in the Annals of Surgery [15]. He noted that when the portal vein and hepatic artery were compressed at the hepatoduodenal ligament with a finger and thumb in a patient with bleeding from traumatic rupture of the liver, the bleeding completely stopped and surgeons were able to clear the blood and clots and examine the wound. Subsequently, temporary control of the hepatic inflow by compression at the hepatoduodenal ligament with a finger and a thumb, a vascular clamp, or a Rummel tourniquet was generally recognized as the “Pringle maneuver”. Thus, the Pringle maneuver is widely employed in controlling hemorrhage from the injured liver during exploratory laparotomies for abdominal trauma [16, 17]. In elective hepatic resection, the Pringle maneuver is used during hepatic parenchymal transection to reduce blood loss and minimize the necessity for blood transfusion, associated with worse outcomes [18, 19]. It is now recognized that the technique of intermittent hepatic inflow occlusion by clamping the hepatoduodenal ligament for 15 minutes and releasing it for 5 minutes and repeating this until the hepatic parenchymal transection is completed can save many lives [20-23]. However, the safe time limit for total occlusion has not been clarified. In practice, prolonged ischemic injury to the liver should be avoided, and the shortest time possible must be applied. Intermittent hepatic inflow occlusion time in our patients ranged from 5 to 105 minutes (median 40, mean 43.3 ± 20.7) and we did not observe any direct detrimental effects of the Pringle maneuver in our patients postoperatively. An obvious advantage of the Pringle maneuver is a bloodless operative field during hepatic parenchymal transection. With the aid of a CUSA, precise ligations of intrahepatic vessels and bile ducts could be made without difficulty. Numerous comments have been published concerning the Pringle maneuver. Some reported adverse effects to the liver in cirrhotic patients [24]. Some authors have reported acceptable outcomes after hepatic resection without the use of the Pringle maneuver [25-27]. Man et al. reported that the Pringle maneuver was associated with less blood loss and better preservation of hepatic function postoperatively [28].

The purpose of the CUSA and the water jet employed as assisting devices during hepatic parenchymal transection is to dissect away the liver tissue by using a high-frequency ultrasonic dissector (CUSA) or high-pressure stream of water (jet) [10, 11, 29-32]. Both devices are currently employed worldwide with various modifications from the original models and decreased intraoperative blood loss and blood transfusion requirements have been reported [33, 34]. CUSA was introduced into our institution before the water jet and this explains why CUSA was employed more frequently than a water jet in the present study. We found that these two devices were both efficacious and could be used interchangeably with equal outcomes.

Despite advances in instrumentation and surgical technology, hepatic resection remains a major operation with potentially disastrous complications and relatively high mortality rates. Certain underlying pathologies carry a higher surgical risk for curative resection, such as hilar cholangiocarcinoma or a large tumor size near the inferior vena cava, the hepatic veins, or the portal pedicle. Hence, complex surgery of the liver continues to be a challenge. Our perioperative mortality of 6% was comparable to current outcomes reported from other high-volume centers [6-9]. Although this mortality rate is at the upper level of the currently acceptable range, we were satisfied with these results because approximately 90% of our patients had primary and secondary malignant tumors of the liver. Furthermore, about two-thirds of these malignant tumors in our study underwent major hepatic resection and one-third biliary-enteric reconstruction. Major complications and mortalities were thus expected. Nevertheless, we realize that with continuous improvement of preoperative, operative, and postoperative care, better outcomes may be achieved.

In conclusion, hepatic resection continues to be a challenging operation with a wide range of complexity. The procedure may be quite simple with negligible mortality or may be a formidable one with a dreadful result. Careful preoperative patient selection for major hepatic resection and meticulous operative techniques are important for good results. CUSA and the Pringle maneuver, in our opinion, are very helpful adjuncts during hepatic parenchymal transection resulting in less blood loss. The experience of the surgical team is also a major factor for acceptable outcome.