ORIGINAL ARTICLE
Complications of Endoscopic Submucosal Dissection for Gastric Noninvasive Neoplasia: An Analysis of 647 Lesions Toshiyasu Ojima, MD, Katsunari Takifuji, MD, Masaki Nakamura, MD, Makoto Iwahashi, MD, Mikihito Nakamori, MD, Masahiro Katsuda, MD, Takeshi Iida, MD, Keiji Hayata, MD, and Hiroki Yamaue, MD
Purpose: This study aimed to determine risk factors for postoperative complications of gastric endoscopic submucosal dissection (ESD). Methods: This retrospective study included 647 lesions in 580 consecutive patients who underwent ESD for gastric noninvasive neoplasia from January 1, 2002 through December 31, 2011. Results: The overall perforation rate was 5.1%. Multivariate logistic regression analysis indicated that perforation was significantly associated with tumors in the greater curvature of the stomach (P < 0.0001), scars in tumor lesions (P = 0.002), long operative time (P = 0.007), and tumors in the remnant stomach (P = 0.036). The bleeding rate after gastric ESD was 3.9%. Multivariate logistic regression analysis indicated a statistically significant association between postoperative bleeding and oral anticoagulant or antiplatelet drugs (P < 0.0001), dialysis (P = 0.009), and use of antihypertensive drugs (P = 0.015). Conclusion: It is important to perform gastric ESD with particular care in patients with risk factors. Key Words: gastric cancer, noninvasive neoplasia, endoscopic submucosal dissection, complication, perforation, bleeding
(Surg Laparosc Endosc Percutan Tech 2014;24:370–374)
E
ndoscopic submucosal dissection (ESD) for gastric epithelial neoplasms is a minimally invasive endoluminal surgery.1,2 ESD has the advantage of permitting en bloc and histologically complete resection.3,4 In Japan, ESD is a standard therapy for gastric noninvasive neoplasia. However, it is a technically difficult and time-consuming procedure with potential for more complications than other procedures such as endoscopic mucosal resection. Although the safety of ESD has been substantiated, it is associated with significant complications, including perforation and post-ESD bleeding. Perforation is observed in 1% to 6% of cases,2,5–9 and postoperative bleeding occurs in approximately 5% of cases.10–12 Previous studies identified risk factors predicting perforation during gastric ESD.6–8 However, many of those studies examined small samples7,8; consequently, the risk factors they identified are not conclusive. A recent study Received for publication November 6, 2012; accepted February 1, 2013. From the Second Department of Surgery, School of Medicine, Wakayama Medical University, Wakayama, Japan. The authors declare no conflicts of interest. Reprints: Hiroki Yamaue, MD, Second Department of Surgery, School of Medicine, Wakayama Medical University, 811-1, Kimiidera, Wakayama 641-8510, Japan (e-mail: [email protected]). Copyright r 2014 by Lippincott Williams & Wilkins
370 | www.surgical-laparoscopy.com
with a large sample showed that lesions in the upper stomach and lesions >20 mm were independent risk factors for perforation.13 However, this study did not address comorbidities.13 Gastric cancer is considered a disease of the elderly. Elderly patients have an increased prevalence of comorbidities. ESD is especially useful in elderly patients with comorbidities, because they have higher rates of surgical complications and mortality. When post-ESD bleeding occurs, patients may suffer from hematemesis and hemorrhagic shock, which may require emergency endoscopy and prolonged fasting. Several studies have investigated risk factors for postoperative bleeding, but their results are inconsistent and there is no consensus with respect to these risk factors.7,11,14–16 In this study, we aimed to determine risk factors for complications of gastric ESD related to patient characteristics, tumor characteristics, and operative factors and to identify preventive measures.
MATERIALS AND METHODS Patients From January 1, 2002 through December 31, 2011, ESD was performed for 647 lesions in 580 consecutive patients with gastric epithelial neoplasms at the Second Department of Surgery, Wakayama Medical University Hospital (WMUH). Diagnostic criteria for ESD were as follows: histopathologic diagnosis by biopsy of well-differentiated or moderately differentiated adenocarcinoma, or high-grade dysplasia; tumor invasion within the mucosa or minute submucosal layers (< 500 mm under the muscle membrane), defined by endoscopic ultrasonography; tumor size r30 mm if minute submucosal invasion or tumor with ulceration was suspected; and any tumor size in the absence of ulceration or submucosal invasion.17 Follow-up data were obtained from our database, including demographic information, surgical details, and tumor characteristics. Tumor invasion and lymph node status were classified according to International Union Against Cancer (UICC) criteria.18 All patients gave informed consent, in accordance with the guidelines of the Ethical Committee on Human Research at WMUH.
Surgical Procedures ESD procedures were performed while the patients were under sedation induced by intravenous diazepam and/ or pentazocine. We used a standard endoscope with a single accessory channel (GIF-H260Z, GIF-Q240, or GIF-Q260J; Olympus, Tokyo, Japan). After the tumor outline had been delineated by chromoendoscopy, marker dots were placed around the outside circumference of the tumor margin with
Surg Laparosc Endosc Percutan Tech
Volume 24, Number 4, August 2014
Surg Laparosc Endosc Percutan Tech
Volume 24, Number 4, August 2014
a flush knife (Fujinon Optical Co., Ltd.; Tokyo, Japan). A flush knife was used for circumferential mucosal incision and submucosal dissection. An insulation-tipped knife (IT knife; Olympus) was used for submucosal dissection during 2002 to 2006. An ICC-200 (Intelligent Cut and Coagulation; Erbe, Tubingen, Germany) or VIO300D (Erbe) was used as the electrosurgical unit. Saline was used for submucosal injection. Sodium hyaluronic acid (0.4%) (Mucoup; Johnson and Johnson K.K., Tokyo, Japan) was used in specific difficult situations. On postoperative day (POD) 1, patients could begin an oral diet if there was no evidence of bleeding or perforation. A second-look endoscopy was not routinely performed. An endoscopic reexamination was performed after 1 week. A proton-pump inhibitor was administered orally from POD 1 to 56.
Complications Perforation was diagnosed if free air was seen on plain x-ray or computed tomography on POD 1. When perforations were detected during resection, they were immediately closed with metal clips during the ESD, and the patient was treated with antibiotics. Postoperative bleeding was defined as clinical evidence of bleeding after ESD that required endoscopic treatment with metal clips, electrocoagulation, or both. Bleeding that occurred during the ESD and that was treated endoscopically was not included in the analysis of complications. Postoperative stricture was defined as the requirement for endoscopically guided balloon dilation against stricture of pylorus or cardia. Postoperative pneumonia was diagnosed based on computed tomography and elevated white blood cell count and serum C-reactive protein.
Statistical Analysis The StatView 5.0 software package (Abacus Concepts Inc., Berkeley, CA) was used for all statistical analyses. Quantitative results are expressed as medians and ranges. Univariate and multivariate logistic regression analyses were performed to identify risk factors for postoperative complications. Risk factors with a univariate P < 0.10 were included in the multivariate analysis. Risk factors with a multivariate P < 0.05 were defined as independent risk factors.
RESULTS Table 1 summarizes detailed characteristics of 647 lesions in 580 patients, including 393 men and 187 women with a median age of 71 years. The median tumor size was 12 mm. The primary tumors were adenocarcinomas (499; 77.1%), adenomas (138; 21.3%), and others (10; 1.5%). Table 2 presents surgical data. The median duration of ESD was 60 minutes. En bloc resection and en bloc R0 resection were achieved in 609 (94.1%) and 575 (88.9%) lesions, respectively. The median size of resected lesions was 32 mm. The overall perforation rate was 5.1% (33/647). All of the perforations occurring during ESD were immediately closed with endoclips. All patients were treated by fasting and intravenous administration of antibiotics. No emergency surgery was required. Postoperative bleeding occurred in 25 (3.9%) lesions. All hemorrhagic episodes were successfully treated by endoscopic clipping, coagulation, or both. Five patients required a blood transfusion. Postoperative stricture was observed in 5 (0.8%) patients, all of whom were treated by endoscopically guided balloon r
2014 Lippincott Williams & Wilkins
Complications of Gastric ESD
TABLE 1. Clinicopathologic Characteristics of 647 Lesions in 580 Patients With Gastric Neoplasms Median age (range) (y) Men/women Whole stomach/remnant stomach Location Upper third Middle third Lower third Circumference Anterior wall Greater curvature Lesser curvature Posterior wall Median tumor size (range) (mm) Ulcer present/absent Histologic type Adenocarcinoma Adenoma Carcinoid Hyperplastic polyp No tumor
71 (30-94) 393/187 608/39 99 181 367 113 130 283 121 12 (0-130) 35/612 499 138 1 3 6
dilation. Postoperative pneumonia occurred in 2 (0.3%) patients. Univariate and multivariate analyses were performed to identify risk factors for perforation during ESD. Table 3 presents the results of univariate analysis of 26 variables for the 33 lesions with perforation versus the 614 lesions without perforation. Eight of 26 factors differed significantly between these groups (P < 0.10). The significant factors related to tumor characteristics were large size (P = 0.028), location in the upper third of the stomach (P = 0.082), location in the greater curvature of the stomach (P < 0.0001), location in the remnant stomach (P = 0.005), malignancy (P = 0.032), and scar in tumor lesions (P = 0.0003). The significant operative factors were long operative time (P < 0.0001) and large resected lesion (P = 0.004). The multivariate logistic regression analysis indicated that perforation during ESD was significantly associated with tumor location in the greater curvature of the stomach [P < 0.0001; odds ratio (OR), 7.0; 95% confidence interval (CI), 3.1-15.8], scar in tumor lesion (P = 0.002; OR, 4.6; 95% CI, 1.8-12.0), long operative time (P = 0.007; OR, 4.4; 95% CI, 1.5-12.9), and tumor location in the remnant stomach (P = 0.036; OR, 3.5; 95% CI, 1.111.4). Univariate and multivariate analyses were performed to identify risk factors for postoperative bleeding. Table 4 presents the results of univariate analysis of 26 variables for the 25 lesions with bleeding versus the 622 lesions without bleeding. Ten of 26 factors differed significantly between
TABLE 2. Surgical Outcomes Median operation time (range) (min) En bloc resection [n (%)] En bloc with R0* resection [n (%)] Median size of resected lesion (range) [mm] Perforation [n (%)] Postoperative bleeding [n (%)] Stricture [n (%)] Pneumonia [n (%)]
60 609 575 32 33 25 5 2
(15-420) (94.1) (88.9) (6-140) (5.1) (3.9) (0.8) (0.3)
*International Union Against Cancer TNM classification.
www.surgical-laparoscopy.com |
371
Surg Laparosc Endosc Percutan Tech
Ojima et al
Volume 24, Number 4, August 2014
TABLE 3. Univariate and Multivariate Analyses of Risk Factors for Perforation During ESD
Univariate Risk Factors
Categories
Patient characteristics Sex Age Dialysis Cirrhosis Antihypertensive drugs Anticoagulants and/or antiplatelets Home oxygen therapy Tumor characteristics Tumor size Location Circumference Remnant stomach No. resected lesions History of ESD for gastric neoplasm Histologic type Scar in tumor lesions Including IIc type Ulcer in tumor lesions Tumor invasive depth Lateral margin Vertical margin Operative factors Operation time En bloc resection Size of resected lesions Experience of endoscopist Snare use Device
Male; female Z70;
Complications of Endoscopic Submucosal Dissection for Gastric Noninvasive Neoplasia: An Analysis of 647 Lesions Toshiyasu Ojima, MD, Katsunari Takifuji, MD, Masaki Nakamura, MD, Makoto Iwahashi, MD, Mikihito Nakamori, MD, Masahiro Katsuda, MD, Takeshi Iida, MD, Keiji Hayata, MD, and Hiroki Yamaue, MD
Purpose: This study aimed to determine risk factors for postoperative complications of gastric endoscopic submucosal dissection (ESD). Methods: This retrospective study included 647 lesions in 580 consecutive patients who underwent ESD for gastric noninvasive neoplasia from January 1, 2002 through December 31, 2011. Results: The overall perforation rate was 5.1%. Multivariate logistic regression analysis indicated that perforation was significantly associated with tumors in the greater curvature of the stomach (P < 0.0001), scars in tumor lesions (P = 0.002), long operative time (P = 0.007), and tumors in the remnant stomach (P = 0.036). The bleeding rate after gastric ESD was 3.9%. Multivariate logistic regression analysis indicated a statistically significant association between postoperative bleeding and oral anticoagulant or antiplatelet drugs (P < 0.0001), dialysis (P = 0.009), and use of antihypertensive drugs (P = 0.015). Conclusion: It is important to perform gastric ESD with particular care in patients with risk factors. Key Words: gastric cancer, noninvasive neoplasia, endoscopic submucosal dissection, complication, perforation, bleeding
(Surg Laparosc Endosc Percutan Tech 2014;24:370–374)
E
ndoscopic submucosal dissection (ESD) for gastric epithelial neoplasms is a minimally invasive endoluminal surgery.1,2 ESD has the advantage of permitting en bloc and histologically complete resection.3,4 In Japan, ESD is a standard therapy for gastric noninvasive neoplasia. However, it is a technically difficult and time-consuming procedure with potential for more complications than other procedures such as endoscopic mucosal resection. Although the safety of ESD has been substantiated, it is associated with significant complications, including perforation and post-ESD bleeding. Perforation is observed in 1% to 6% of cases,2,5–9 and postoperative bleeding occurs in approximately 5% of cases.10–12 Previous studies identified risk factors predicting perforation during gastric ESD.6–8 However, many of those studies examined small samples7,8; consequently, the risk factors they identified are not conclusive. A recent study Received for publication November 6, 2012; accepted February 1, 2013. From the Second Department of Surgery, School of Medicine, Wakayama Medical University, Wakayama, Japan. The authors declare no conflicts of interest. Reprints: Hiroki Yamaue, MD, Second Department of Surgery, School of Medicine, Wakayama Medical University, 811-1, Kimiidera, Wakayama 641-8510, Japan (e-mail: [email protected]). Copyright r 2014 by Lippincott Williams & Wilkins
370 | www.surgical-laparoscopy.com
with a large sample showed that lesions in the upper stomach and lesions >20 mm were independent risk factors for perforation.13 However, this study did not address comorbidities.13 Gastric cancer is considered a disease of the elderly. Elderly patients have an increased prevalence of comorbidities. ESD is especially useful in elderly patients with comorbidities, because they have higher rates of surgical complications and mortality. When post-ESD bleeding occurs, patients may suffer from hematemesis and hemorrhagic shock, which may require emergency endoscopy and prolonged fasting. Several studies have investigated risk factors for postoperative bleeding, but their results are inconsistent and there is no consensus with respect to these risk factors.7,11,14–16 In this study, we aimed to determine risk factors for complications of gastric ESD related to patient characteristics, tumor characteristics, and operative factors and to identify preventive measures.
MATERIALS AND METHODS Patients From January 1, 2002 through December 31, 2011, ESD was performed for 647 lesions in 580 consecutive patients with gastric epithelial neoplasms at the Second Department of Surgery, Wakayama Medical University Hospital (WMUH). Diagnostic criteria for ESD were as follows: histopathologic diagnosis by biopsy of well-differentiated or moderately differentiated adenocarcinoma, or high-grade dysplasia; tumor invasion within the mucosa or minute submucosal layers (< 500 mm under the muscle membrane), defined by endoscopic ultrasonography; tumor size r30 mm if minute submucosal invasion or tumor with ulceration was suspected; and any tumor size in the absence of ulceration or submucosal invasion.17 Follow-up data were obtained from our database, including demographic information, surgical details, and tumor characteristics. Tumor invasion and lymph node status were classified according to International Union Against Cancer (UICC) criteria.18 All patients gave informed consent, in accordance with the guidelines of the Ethical Committee on Human Research at WMUH.
Surgical Procedures ESD procedures were performed while the patients were under sedation induced by intravenous diazepam and/ or pentazocine. We used a standard endoscope with a single accessory channel (GIF-H260Z, GIF-Q240, or GIF-Q260J; Olympus, Tokyo, Japan). After the tumor outline had been delineated by chromoendoscopy, marker dots were placed around the outside circumference of the tumor margin with
Surg Laparosc Endosc Percutan Tech
Volume 24, Number 4, August 2014
Surg Laparosc Endosc Percutan Tech
Volume 24, Number 4, August 2014
a flush knife (Fujinon Optical Co., Ltd.; Tokyo, Japan). A flush knife was used for circumferential mucosal incision and submucosal dissection. An insulation-tipped knife (IT knife; Olympus) was used for submucosal dissection during 2002 to 2006. An ICC-200 (Intelligent Cut and Coagulation; Erbe, Tubingen, Germany) or VIO300D (Erbe) was used as the electrosurgical unit. Saline was used for submucosal injection. Sodium hyaluronic acid (0.4%) (Mucoup; Johnson and Johnson K.K., Tokyo, Japan) was used in specific difficult situations. On postoperative day (POD) 1, patients could begin an oral diet if there was no evidence of bleeding or perforation. A second-look endoscopy was not routinely performed. An endoscopic reexamination was performed after 1 week. A proton-pump inhibitor was administered orally from POD 1 to 56.
Complications Perforation was diagnosed if free air was seen on plain x-ray or computed tomography on POD 1. When perforations were detected during resection, they were immediately closed with metal clips during the ESD, and the patient was treated with antibiotics. Postoperative bleeding was defined as clinical evidence of bleeding after ESD that required endoscopic treatment with metal clips, electrocoagulation, or both. Bleeding that occurred during the ESD and that was treated endoscopically was not included in the analysis of complications. Postoperative stricture was defined as the requirement for endoscopically guided balloon dilation against stricture of pylorus or cardia. Postoperative pneumonia was diagnosed based on computed tomography and elevated white blood cell count and serum C-reactive protein.
Statistical Analysis The StatView 5.0 software package (Abacus Concepts Inc., Berkeley, CA) was used for all statistical analyses. Quantitative results are expressed as medians and ranges. Univariate and multivariate logistic regression analyses were performed to identify risk factors for postoperative complications. Risk factors with a univariate P < 0.10 were included in the multivariate analysis. Risk factors with a multivariate P < 0.05 were defined as independent risk factors.
RESULTS Table 1 summarizes detailed characteristics of 647 lesions in 580 patients, including 393 men and 187 women with a median age of 71 years. The median tumor size was 12 mm. The primary tumors were adenocarcinomas (499; 77.1%), adenomas (138; 21.3%), and others (10; 1.5%). Table 2 presents surgical data. The median duration of ESD was 60 minutes. En bloc resection and en bloc R0 resection were achieved in 609 (94.1%) and 575 (88.9%) lesions, respectively. The median size of resected lesions was 32 mm. The overall perforation rate was 5.1% (33/647). All of the perforations occurring during ESD were immediately closed with endoclips. All patients were treated by fasting and intravenous administration of antibiotics. No emergency surgery was required. Postoperative bleeding occurred in 25 (3.9%) lesions. All hemorrhagic episodes were successfully treated by endoscopic clipping, coagulation, or both. Five patients required a blood transfusion. Postoperative stricture was observed in 5 (0.8%) patients, all of whom were treated by endoscopically guided balloon r
2014 Lippincott Williams & Wilkins
Complications of Gastric ESD
TABLE 1. Clinicopathologic Characteristics of 647 Lesions in 580 Patients With Gastric Neoplasms Median age (range) (y) Men/women Whole stomach/remnant stomach Location Upper third Middle third Lower third Circumference Anterior wall Greater curvature Lesser curvature Posterior wall Median tumor size (range) (mm) Ulcer present/absent Histologic type Adenocarcinoma Adenoma Carcinoid Hyperplastic polyp No tumor
71 (30-94) 393/187 608/39 99 181 367 113 130 283 121 12 (0-130) 35/612 499 138 1 3 6
dilation. Postoperative pneumonia occurred in 2 (0.3%) patients. Univariate and multivariate analyses were performed to identify risk factors for perforation during ESD. Table 3 presents the results of univariate analysis of 26 variables for the 33 lesions with perforation versus the 614 lesions without perforation. Eight of 26 factors differed significantly between these groups (P < 0.10). The significant factors related to tumor characteristics were large size (P = 0.028), location in the upper third of the stomach (P = 0.082), location in the greater curvature of the stomach (P < 0.0001), location in the remnant stomach (P = 0.005), malignancy (P = 0.032), and scar in tumor lesions (P = 0.0003). The significant operative factors were long operative time (P < 0.0001) and large resected lesion (P = 0.004). The multivariate logistic regression analysis indicated that perforation during ESD was significantly associated with tumor location in the greater curvature of the stomach [P < 0.0001; odds ratio (OR), 7.0; 95% confidence interval (CI), 3.1-15.8], scar in tumor lesion (P = 0.002; OR, 4.6; 95% CI, 1.8-12.0), long operative time (P = 0.007; OR, 4.4; 95% CI, 1.5-12.9), and tumor location in the remnant stomach (P = 0.036; OR, 3.5; 95% CI, 1.111.4). Univariate and multivariate analyses were performed to identify risk factors for postoperative bleeding. Table 4 presents the results of univariate analysis of 26 variables for the 25 lesions with bleeding versus the 622 lesions without bleeding. Ten of 26 factors differed significantly between
TABLE 2. Surgical Outcomes Median operation time (range) (min) En bloc resection [n (%)] En bloc with R0* resection [n (%)] Median size of resected lesion (range) [mm] Perforation [n (%)] Postoperative bleeding [n (%)] Stricture [n (%)] Pneumonia [n (%)]
60 609 575 32 33 25 5 2
(15-420) (94.1) (88.9) (6-140) (5.1) (3.9) (0.8) (0.3)
*International Union Against Cancer TNM classification.
www.surgical-laparoscopy.com |
371
Surg Laparosc Endosc Percutan Tech
Ojima et al
Volume 24, Number 4, August 2014
TABLE 3. Univariate and Multivariate Analyses of Risk Factors for Perforation During ESD
Univariate Risk Factors
Categories
Patient characteristics Sex Age Dialysis Cirrhosis Antihypertensive drugs Anticoagulants and/or antiplatelets Home oxygen therapy Tumor characteristics Tumor size Location Circumference Remnant stomach No. resected lesions History of ESD for gastric neoplasm Histologic type Scar in tumor lesions Including IIc type Ulcer in tumor lesions Tumor invasive depth Lateral margin Vertical margin Operative factors Operation time En bloc resection Size of resected lesions Experience of endoscopist Snare use Device
Male; female Z70;
