Diagnosis and Treatment of Abdominal Arterial Bleeding After Radical Gastrectomy: a Retrospective Analysis of 1875 Consecutive Resections for Gastric Cancer

This is a retrospective study performed in a single high-volume gastric cancer center. From January 2003 to December 2013, 1875 consecutive gastric cancer patients who underwent D2 or D2 plus gastrectomy were analyzed from our prospectively designed database. Only confirmed abdominal arterial bleeding cases after radical gastrectomy, with systematic lymphadenectomy, were included in the analysis. Patients who underwent radical gastrectomy with standard D2 or D2 plus lymph node dissection with curative intent for gastric cancer and had a confirmed diagnosis of abdominal arterial hemorrhage, either by transcatheter arterial angiography or open surgery, were included in the study. Patients who underwent laparoscopic surgery, palliative surgery, reduction surgery, bypass surgery, or biopsy alone and those who suffered transient hemorrhage, hemorrhage from anastomotic line, non-arterial hemorrhage (i.e., venous hemorrhage), diffuse gastritis, preexisting coagulation disorders, or other unclear sources of hemorrhage were excluded from the study.

Of note, in our series, anticoagulation therapy or antiplatelet therapy was not used perioperatively. Neoadjuvant chemotherapies were not performed in all patients. Stages of gastric cancer were classified using the seventh edition of UICC/AJCC TNM stage system.

Ethical subcommittee of The First Affiliated Hospital of Sun Yat-sen University approved this retrospective study (approval number: 20140429). Written consent was obtained from the patients for their information to be stored in the hospital database and used for research.

All included patients underwent total or subtotal gastrectomy with D2 or greater extended lymph node dissection, based on the recommendation from the Japanese Research Society for Gastric Carcinoma, with curative intent.22 All operations were done by experienced surgeons who had performed at least 50 D2 radical gastrectomies, which required the systematic dissection of lymph nodes in the first tier (perigastric) and the second tier (along the celiac artery and its branches). Based on preoperative staging, D2 plus or D3 lymphadenectomy was performed on selected patients. The types of gastrectomy were selected largely based on the size and location of the primary tumors. Omentobursectomy was performed for tumors located on the posterior gastric wall, especially those penetrating the serosa. Combined organ resections, such as splenectomy and distal pancreatectomy, were carried out if the primary tumor directly invaded the adjacent organs to obtain an R0 resection. Reconstruction types included Billroth type I/II and Roux-en-Y esophagojejunostomies. After anastomosis, hemostasis at the anastomotic line was checked routinely. Warm distilled water lavage and soaking were routinely performed for inactivation of deciduous gastric cancer cells and to find potential hemorrhagic spots. For all patients, one silicone drain was placed around the anastomotic site through the Winslow foramen before closure. Another drain was added at the splenic fossa for patients with total gastrectomy, especially with combined pancreatectomy or splenectomy.

Due to a lack of generally accepted definition and classification of postoperative bleeding for gastric cancer, we referenced and adopted the classification of postoperative hemorrhage for pancreatic surgery.14‐19 Based on the onset time of hemorrhage, patients who suffered from arterial bleeding within and beyond 24 h after surgery were assigned to an early arterial bleeding group and a late arterial bleeding group, respectively. For the late bleeding group, based on the primary interventional measure, we divided patients into the TAE first group or re-laparotomy first group. Sentinel bleeding was defined as any kind of minor bleeding that required no intervention and often preceded one or more major bleeding events.20,21 Re-bleeding was defined as recurrent active bleeding that required further intervention after the initial hemostatic procedure.

Abdominal arterial bleeding following gastrectomy was alerted by series of clinical manifestations including massive warm and bright red bleeding from an abdominal drain, tachycardia, hemodynamic instability, sudden abdominal pain, distension, etc. Hemodynamic instability was defined as a quick decease in the mean arterial pressure (MAP <65 mmHg) prior to fluid resuscitation or blood transfusion. Fluid resuscitation was initiated prior to and during emergency interventions. The following diagnostic procedures were considered: bedside ultrasonography, abdominocentesis, or angiography. A decision to proceed with TAE or re-laparotomy was based on hemodynamic status, extent of bleeding, risk factor, onset time of bleeding, etc.

In our hospital, a team of interventional radiologists was able to respond within 10 min, 24 h per day. All angiography and TAE were performed under local anesthesia. Celiac and superior mesenteric arterial angiographies were routinely performed to detect the bleeding in all cases. Positive angiographic findings included extravasations of contrast medium, pseudoaneurysms, or fusiform aneurysms. When the bleeding source was demonstrated at angiography, embolization was performed as selectively as possible. The choices between embolization and covered stent graft depended on the discretion of interventional radiologists and anatomic conditions. Embolization coils were placed from the distal part to the proximal part of the bleeding point to prevent retrograde perfusion through collateral vessels (Figs. 1 and 2, Supporting Files 1 and 2). A re-laparotomy was performed as a rescue treatment for a negative finding or failure of radiological intervention. Liver function was closely monitored after the procedure. Abdominal CT scans were performed to check for fluid collections after successful bleeding control by TAE.

All emergency re-laparotomies were performed under general anesthesia. In cases with less than 14 postoperative days, bleeding vessels could be generally detected and confirmed. However, in cases with severe vascular damage, extensive adhesion due to unconfined anastomotic leakage, or more than 2 weeks after gastrectomy, bleeding vessels could not always be confirmed. Surgical procedures ranged from simple ligation of bleeding arteries, over sewing the bleeding site, to vascular reconstruction, and to bleeding organ resection, like splenectomy. Gauze packing was used as a salvage measure when bleeding control failed.

Intra-abdominal infection was confirmed by a positive culture of the intra-abdominal fluid collected from patients who had a body temperature higher than 38 °C. Local sepsis and intra-abdominal abscess were defined as direct drainage of abscess from the peritoneal cavity. Pancreatic fistula was defined when drain fluid is >20 ml/24 h, after 3 days postoperatively with amylase activity exceeding three times that of the serum. Anastomotic leakage was diagnosed by digestive tract radiography. Ultrasound-guided additional drains were usually placed if intra-abdominal fluid collection was found.

All statistical analyses were performed using the Statistical Package for Social Science (SPSS) version 19.0 for Windows (IBM, Chicago, IL, USA). A Student’s t test and a Mann-Whitney U test were used to compare continuous variables. Fisher’s exact test was used to compare categorical variables. A P value of <0.05 was considered to indicate statistical significance.