Efficacy of Laparoscopic Gastric Mobilization for Esophagectomy: Comparison with Open Thoraco-abdominal Approach

Introduction

Esophagectomy for esophageal cancer using the thoraco-abdominal approach is associated with a high rate of morbidity and mortality due to intense surgical stress, which traditionally involves opening the chest and abdomen. ¹ In gastrointestinal surgery, it has been reported that less invasive techniques such as laparoscopic or thoracoscopic surgery reduce tissue injury, resulting in less postoperative pain and earlier recovery.^2,3 Thoracoscopic surgery has been well described as a less invasive operation in the field of esophageal surgery in patients with resectable esophageal cancer. ⁴ However, few reports directly compare transthoracic esophagectomy and laparoscopic gastric mobilization (LGM) with conventional open thoraco-abdominal esophagectomy (OE) for thoracic esophageal cancer.

We have previously reported that total LGM (TLGM) and lymphadenectomy reduce operative blood loss and shorten intensive care unit (ICU) stay. ⁵ The aim of this study was to prove the clinical efficacy of LGM (TLGM or hand-assisted laparoscopic surgery [HALS]) compared with conventional OE for thoracic esophageal cancer.

Patients and Methods

Ninety-two esophageal cancer patients who underwent esophagectomy via the thoraco-abdominal Ivor Lewis approach from 1999 to 2009 at Kochi Medical School (Kochi, Japan) were reviewed. LGM has been performed since April 2005, and until 2009, 45 patients underwent the procedure. The LGM group could be subdivided as follows: TLGM, 27; HALS, 18. We performed the HALS procedure on patients who had bulky abdominal lymph nodes detected by preoperative imaging or had a history of previous major abdominal surgery. We defined bulky abdominal lymph nodes as those that were greater than 3 cm in diameter. When it was too difficult to perform HALS because of severe abdominal adhesion, we performed open laparotomy.

The TLGM technique has been described in detail in our previous report. ⁵ In brief, the patient is placed in a 20° anti-Trendelenburg position, and LGM with regional lymphadenectomy is performed using five trocars. The operation begins with division of the gastrocolic ligament and gastrosplenic ligament. The right gastroepiploic vessels are carefully preserved. The root of the left gastric vessel is divided with lymphadenectomy. After division of the phrenic vein and opening of the esophageal hiatus, lower mediastinum dissection is performed. We do not routinely perform the Kocher maneuver and pyloroplasty in order to avoid the reflux of bile. When we choose the HALS procedure, the operation begins from a midline 7-cm small incision laparotomy, and two trocars are added at the left abdomen. A feeding jejunostomy catheter is inserted for all patients.

The main operative steps in the LGM group were as follows: (1) abdominal stage, gastric mobilization with abdominal lymph node dissection and feeding jejunostomy tube construction; (2) cervical stage, cervical lymph node dissection especially around the recurrent nerve by color incision; (3) thoracic stage, muscle (latissimus dorsi and serratus anterior) preserving thoracotomy followed by thoracic esophageal mobilization with mediastinal lymph node dissection, pulling up the stomach and gastric tube formation; and (4) creation of the intrathoracic or cervical anastomosis. During the early period of this study, we created the cervical anastomosis by the hand-sewn technique. From June 2001, we changed the method to cervical anastomosis with a circular stapler. Beginning in October 2004, we changed to intrathoracic anastomosis with a circular stapler; however, we experienced a severe mediastinitis followed by intrathoracic anastomotic leak in October 2005. We therefore returned to the cervical hand-sewn method until September 2009. Since then we have routinely performed cervical anastomosis with a circular stapler.

Before surgery, patients had an epidural catheter inserted for pain control and were admitted to the ICU with a ventilator after surgery. The day after surgery, patients were removed from the ventilator and returned to the ward if their general condition was good.

The study measures included operative time, operative blood loss volume, the number of dissected lymph nodes, the duration of postoperative ICU and hospital stay, incidence of postoperative infections (POIs), and postoperative mortality within 30 days. POIs include postoperative pneumonia and surgical site infection (SSI), including wound infection after anastomotic leak. Hospital stay was calculated from the date of the operation until the date of discharge; this measure was excluded in cases of mortality. In patients who received additional in-hospital chemoradiation therapy, hospital stay was calculated from the date of the operation until the date that additional therapy was started.

The Mann–Whitney U test was used to assess correlations among the continuous variables for each group. The Pearson chi-squared test was applied to qualitative variables. P values<.05 were considered significant.

Results

Table 1 shows the characteristics for each patient group. The groups did not differ in terms of gender, age, location of tumor, histology, depth of invasion, history of preoperative chemotherapy or chemoradiotherapy, body mass index, and history of diabetes mellitus. In the OE group, 2 patients had a right colon reconstruction performed. One patient had a simultaneous gastric resection for gastric cancer, and the other had a history of gastrectomy for gastric cancer. Two surgeons performed the surgery in this study period. Before 2006, the mentoring surgeon had performed the operations. From 2006, a younger surgeon performed most of the surgery coached by an experienced mentor. In the OE group, 35 patients had cervical hand-sewn anastomoses, and 12 had intrathoracic anastomoses with a circular stapler. In the LGM group, 37 patients had cervical hand-sewn anastomoses, and 8 had intrathoracic anastomoses with a circular stapler.

Operative findings and outcomes are summarized in Table 2. In the LGM group, 1 case had open conversion because of bleeding when a left gastric artery was divided by the vessel sealing system. The median operation time (P=.450) and the number of dissected lymph nodes (P=.226) in both groups were not significantly different. However, the median operative blood loss volume (430 versus 1060 mL; P<.001) and the number of patients who underwent blood transfusion (9 versus 26; P=.001) in the LGM group were significantly lower than in the OE group. The LGM patients had a shorter length of ICU stay (1 versus 3 days; P<.001) and hospital stay (35 versus 46 days; P=.006).

The mortality rate did not differ significantly between the LGM and OE groups (2.2% versus 4.3%, respectively; P=.969). In the LGM group, 1 patient died 24 days after surgery because of mediastinitis from an intrathoracic anastomotic leak. Two patients died in the OE group 22 and 24 days after the surgery because of severe pneumonia. One was a 58-year-old male patient who had a history of pulmonary tuberculosis. He developed acute respiratory distress syndrome after the surgery. Although the other patient had no previous pulmonary disease, he also developed acute respiratory distress syndrome. The incidence of POIs was significantly lower in the LGM group than in the OE group (33.3% versus 55.3%, respectively; P=.034), although there was no statistically significant difference between pneumonia or SSI individually. Few patients developed a fascial dehiscence or subsequent hernias because the wound infections mainly occurred in the neck or thorax. There was no difference in the incidence of anastomotic leak between the LGM and OE groups (19.1% versus 20.0%).

The median survival time in the LGM group was longer than in the OE group, although the difference was not statistically significant (61.7 versus 31.8 months; P=.195).

Discussion

We introduced the LGM procedure in the hope of decreasing postoperative complications. In our series of 45 consecutive LGM cases, operative blood loss and the incidence of POIs were decreased compared with OE without increasing the operation time or mortality related to esophagectomy. Significantly lower blood loss in the LGM was consistent with previous studies.^6,7 We think that LGM can be performed as a more precise operation than OE and that ligations can be made even in small vessels or for minor oozing.

Although the mortality after esophagectomy has decreased in recent years, postoperative morbidities including infectious complications still remain significantly problematic. ¹ Although previous investigators have established the feasibility of less invasive esophagectomy, several studies have not shown a reduction in postoperative pulmonary complications^8–11 (Table 3). Lawrence et al. ¹² reported that a laparoscopic procedure for gastrointestinal surgery reduces the incidence of postoperative pneumonia and SSI compared with an open procedure. In general, it is thought that laparoscopic surgery protects the gut from drying and reduces the trauma to the abdominal wall. It has been reported that laparoscopic surgery prevents desiccation of the intra-abdominal surfaces by maintaining humidity. ¹³ The use of heated humidified CO₂ for pneumoperitoneum in laparoscopic procedures has been shown to be associated with less postoperative pain and a lower risk of postoperative hypothermia. ¹⁴ It is therefore easier for the patient to breathe deeply, cough strongly to spit out phlegm, and begin to stand up and walk. This results in a reduction in POIs. Some authors have reported that esophagectomy with LGM is feasible, but the effect on decreasing complications is unclear.^15–18 In this study, the incidence of POIs in the LGM group was significantly lower than in the OE group, although there was no statistically significant difference between pneumonia and SSI individually. The incidence of pneumonia in the LGM group was less than half of that in the OE group (13.3% versus 29.8%; P=.097), and SSI in the LGM group was also reduced (28.9% vs. 42.6%; P=.172).

Anastomotic leak is the most problematic remaining issue for us. In this study, both groups had 9 patients with anastomotic leak (19.1% versus 20.0%). We need a greater reduction in incidence of leak in order to reduce SSI. Arterial blood supply, venous return, and tension for anastomosis appear to be important; therefore intrathoracic anastomosis is better than cervical anastomosis. However, if anastomotic leak occurred, there may be a serious risk of severe mediastinitis or bronchial fistula. If a leak occurred at a cervical anastomosis, we could rapidly open the wound in the neck to allow drainage. We have therefore decided to use cervical anastomosis as our routine method, and we are currently trying to evaluate arterial blood supply using indocyanine green fluorescence methods.

In this study, LGM did shorten the hospital stay compared with OE. Our studies suggest that the main reason for this is the decrease in postoperative infectious complications. The duration of hospitalization was slightly longer in our study than in other reports. The reason for this may be that our patients mostly lived in rural areas and have to travel far to our institution. They are therefore usually hospitalized until they are able to eat a reasonable amount, and this policy of postoperative management was not changed during the study period.

This study has some limitations and biases due to its retrospective nature and small size. These include the historical background and advances in surgical techniques and equipment.

Conclusions

Our results suggest that the efficacy of LGM including its precision and abdominal pain control might play an important role in the reduction of POIs and early postoperative recovery for patients with esophageal cancer. Further efforts are required to reduce the incidence of postoperative pneumonia and SSIs.