Intrapyloric Botulinum Injection Increases Postoperative Esophagectomy Complications Shady M. Eldaif, MD, Richard Lee, MD, Kumari N. Adams, MD, Patrick D. Kilgo, BS, Mark A. Gruszynski, MD, Seth D. Force, MD, Allan Pickens, MD, Felix G. Fernandez, MD, Theresa D. Luu, MD, and Daniel L. Miller, MD Department of Surgery, Section of General Thoracic Surgery, Emory University School of Medicine, Atlanta, Georgia
Background. Intrapyloric botulinum toxin injection has emerged as a possible alternative to standard pyloric drainage procedures. Possible advantages include decreased operative time and less postoperative dumping and bile reflux symptoms. However, data are lacking to show its effectiveness versus standard drainage procedures. The purpose of this review is to compare the results in a prospective cohort of patients who received pyloric botulinum injection versus patients who received pyloromyotomy or pyloroplasty with esophagectomy. Methods. We performed a retrospective review of a prospective database of all patients who underwent an open esophageal resection at a single institution from 2005 through 2010. Three hundred twenty-two patients were divided into 3 groups for analysis: botulinum injection (n [ 78), pyloromyotomy (n [ 45), and pyloroplasty (n [ 199). We compared these groups with respect to duration of the procedure, presence of delayed gastric emptying on postoperative swallow studies, requirement of anastomotic dilation, requirement of pyloric dilation, use of postoperative promotility agents, and patient experience of postoperative symptoms of reflux or dumping, or both. Results. Patients receiving botulinum injections experienced similar delayed gastric emptying on postoperative radiologic evaluation as did patients undergoing
pyloromyotomy and pyloroplasty (16% versus 5% and 13%, respectively; p [ 0.14). Mean operative time was significantly shorter for the patients receiving botulinum as expected (239 minutes versus 312 minutes and 373 minutes, respectively; p < 0.001). However, more patients receiving botulinum and pyloric dilation (22% versus 4% and 2%, respectively; p < 0.001) experienced postoperative reflux symptoms (32% versus 12% and 13%, respectively; p [ 0.001) and used postoperative promotility agents (22% versus 5% and 15%, respectively; p [ 0.04). There was no statistical difference between the groups regarding postoperative dumping. Conclusions. Use of intrapyloric botulinum injection significantly decreased operative time. However, the patients receiving botulinum experienced more postoperative reflux symptoms, had increased use of promotility agents as well as a requirement for postoperative endoscopic interventions, and postoperative dumping was not reduced by the reversible procedure. Intrapyloric botulinum injection should not be used as an alternative to standard drainage procedures. Pyloromyotomy appears to be the drainage procedure of choice to accompany an esophagectomy.
E
The largest randomized controlled trial to date demonstrated significantly shorter gastric emptying times in patients undergoing pyloroplasty versus patients not undergoing drainage procedures, as well as earlier tolerance of solid diet and decreased symptoms with meals at 6 months [1]. Likewise, a literature review of 9 randomized controlled trials with meta-analysis by Urschel and colleagues [2] concluded that pyloroplasty compared with no intervention reduced the incidence of gastric outlet obstruction but had little effect on operative morbidity or mortality and late postoperative gastric conduit function. There remains no consensus because numerous other reports from high-volume centers have described no decrease in delayed gastric emptying after drainage procedures such as pyloroplasty or pyloromyotomy [3, 4]. Furthermore, detractors emphasize the increased prevalence of bile reflux, esophagitis, and dumping syndrome after these emptying procedures [5, 6].
sophagectomy with gastric conduit reconstruction is the preferred operation for esophageal cancer and certain benign diseases. However, the sequelae of gastric interposition can contribute to significant morbidity and mortality in this population. Respiratory complications associated with aspiration of gastrointestinal material and subsequent pneumonias are the most frequent complications in the early postoperative period [1]. Gastric outlet obstruction is recognized as the predominant risk factor for aspiration. Currently, the efficacy of pyloric drainage procedures in preventing delayed gastric emptying remains widely debated.
Accepted for publication Nov 11, 2013. Presented at the Fifty-ninth Annual Meeting of the Southern Thoracic Surgical Association, Naples, FL, Nov 7–10, 2012. Address correspondence to Dr Miller, Emory University Clinic, 1365 Clifton Rd NE, Atlanta, GA 30322; e-mail: daniel.miller@ emoryhealthcare.org.
Ó 2013 by The Society of Thoracic Surgeons Published by Elsevier Inc
(Ann Thorac Surg 2013;-:-–-) Ó 2013 by The Society of Thoracic Surgeons
0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2013.11.026
2
ELDAIF ET AL INTRAPYLORIC BOTULINUM INJECTION
These disparities paved the way for the introduction of intrapyloric botulinum injection as an alternative to the irreversible gastric emptying procedures and as a solution to early delayed gastric emptying and dumping and reflux secondary to a wide open pylorus. Kent and colleagues [7] in 2007 were the first to show in a pilot study of 12 patients who had undergone minimally invasive esophagectomy (MIE) that intrapyloric injection of botulinum toxin prevented delay in gastric emptying or caused aspiration pneumonia. The purpose of this review is to compare the results in a large prospective cohort of patients who underwent 1 of 3 drainage procedures: pyloric botulinum injection, pyloromyotomy, or pyloroplasty at the time of open esophageal resection.
Material and Methods Study Design and Surgical Methods With approval from our institutional review board, we performed a retrospective review of a prospective database of all patients who underwent esophagectomy at a single institution, Emory University Hospital, from 2005 through 2010, using the stomach as the reconstruction conduit. Three hundred twenty-two patients had open esophageal resection with gastric conduit reconstruction by an Ivor Lewis, McKeown, or transhiatal technique. The study population was then divided into 3 groups for analysis based on the pyloric drainage procedure: botulinum injection (n ¼ 78), pyloromyotomy (n ¼ 45), or pyloroplasty (n ¼ 199). All patients received a drainage procedure. The type of resection and drainage procedure performed was based on the preference of the surgeon and the tumor location. Pyloromyotomy and pyloroplasty procedures were performed in standard fashion: A Heineke-Mikulicz pyloroplasty was performed with a full-thickness 2- to 3-cm longitudinal incision extending on each side of the pyloric vein (vein of Mayo) and closed vertically in a single layer with either 3-0 silk or 3-0 polydiaxanone sutures, whereas the pyloromyotomy was performed as a Freder-Ramstedt procedure with a 3-cm longitudinal incision at the pyloric vein through the serosa and muscle fibers, with the final layer of muscle divided with sharp dissection to allow bulging of the mucosa above the serosa level. Extraluminal pyloric injection was performed using 100 U of botulinum toxin in 2 mL of normal saline (NS) (equally distributed into 25 U per injection into 4 separate quadrants) after gastric mobilization. The dose of botulinum toxin was the same used by Cerfolio and colleagues [8]—100 U in 4 mL NS in 4 quadrants—but less than that used by Kent and colleagues [7]—200 U in 5 mL NS, also in 4 quadrants. Botulinum toxin is an inhibitor of cholinergic neuromuscular transmission and has been used to treat spastic disorders of both striated and smooth muscles by local injections [9]. This has been reported to be successful in patients with gastroparesis secondary to diabetes and idiopathic causes. The effects of the botulinum toxin usually last 3 to 5 months [10].
Ann Thorac Surg 2013;-:-–-
Postoperative Evaluation and Follow-Up All patients were managed with nasogastric tubes postoperatively for 3 days with a cervical anastomosis and for 5 days with a thoracic anastomosis. All patients underwent a fluoroscopic water-soluble contrast agent or barium contrast agent swallow study, or both, between postoperative days 5 and 7 to access anastomotic integrity and gastric emptying in relation to the conduit and the pylorus. Patients with a cervical anastomosis or those who received preoperative radiotherapy were also evaluated further for the risk of aspiration with a modified barium swallow study performed by one of our speech pathologists. All patients were given esomeprazole (Nexium, AstraZeneca Pharmaceuticals) starting on postoperative day 1 and promotility agents, either metoclopramide (Reglan, Baxter Pharmaceuticals) or erythromycin if delayed gastric emptying was demonstrated on the radiologic swallow study. We compared the 3 groups with respect to duration of the procedure, delayed gastric emptying on postoperative radiologic studies, requirement of anastomotic dilation, requirement of pyloric dilation, use of postoperative promotility agents, and patient experience of postoperative symptoms. A standard postoperative review of systems was completed by each patient at their 6-month follow-up clinic visit. Gastrointestinal symptoms focused on for our analysis were dysphagia, odynophagia, chest or abdominal pain, reflux, and dumping.
Statistical Evaluation Unadjusted pyloric group comparisons were made using c2 tests and analysis of variance methods for categorical and continuous variables, respectively. Logistic regression was used to determine the association between each binary clinical end point and pyloric drainage group, adjusted for age, sex, preoperative use of chemotherapy and radiation treatment, and surgical technique. Adjusted odds ratio (AOR) with 95% confidence interval (CI) was used as the measure of association. The botulinum group was the reference group. Using the same adjustment methodology, general linear modeling was used to model the lone numerical end point—length of stay. All analyses were performed using SAS, version 9.3 (SAS Institute, Cary, NC) and all tests were evaluated at the 0.05 alpha level. Missing data among predictors were present in less than 2% of cases. Missing data were more prevalent in outcomes; on average, outcome variables had about 8% missing data (range, 1.2%-14.3%). A multiple imputation procedure was implemented for cases with missing predictors so that the maximum number of cases could be used. However, outcomes were not imputed when missing, and thus a complete case analysis was performed for these cases.
Results Unadjusted Analyses Preoperative characteristics are included in Table 1. The majority of procedures (86%) were performed for
Ann Thorac Surg 2013;-:-–-
3
ELDAIF ET AL INTRAPYLORIC BOTULINUM INJECTION
Table 1. Preoperative Characteristics Botulinum n ¼ 78
Variable a
62.2 27 34 35
Patient age (SD) Female sex (n, %) Preoperative radiationa (n, %) Preoperative chemotherapya (n, %) a
(12.0) (34.6) (44.2) (45.5)
Pyloromyotomy n ¼ 45 60.5 13 21 22
(10.0) (28.9) (46.7) (48.9)
Pyloroplasty n ¼ 199 62.2 39 56 80
(11.0) (19.6) (28.3) (40.4)
p Value 0.64 0.025 0.009 0.50
There are significant mean differences by group (p < 0.001) and by operative type (p < 0.001).
SD ¼ standard deviation.
esophageal cancer; 11% were performed for benign disease, including achalasia, stricture, and perforation; and 3% were performed for Barrett’s high-grade dysplasia. The number of patients with benign disease was not statistically different among the groups. Approximately 42% of the patients with cancer underwent neoadjuvant treatment, with significantly fewer patients receiving preoperative radiation in the pyloroplasty group. Intraoperative data are included in Table 2. Mean operative time was significantly shorter for the botulinum group in comparison with the pyloromyotomy and pyloroplasty groups (239 minutes versus 312 minutes and 373 minutes, respectively; p < 0.001). Mean operative time was further subclassified (Table 2) to demonstrate if the length of operation was different for the 3 esophagectomy techniques. Gastric conduit width and anastomosis location were similar between the groups. The average width of the gastric conduits was 5 to 8 cm for both the intrathoracic and cervical reconstructions. The location of the intrathoracic anastomosis was above the azygous vein, with distance from the incisors of 21 to 23 cm, and the cervical anastomosis was located at 17 to 19 cm from the incisors. A delay on the postoperative swallow study was defined according to the method described by Cerfolio and colleagues [8] to unify the definition in the literature. Patients receiving botulinum injections experienced similar delayed gastric emptying, but the trend was toward significantly slow emptying on postoperative swallow studies compared with patients who underwent pyloromyotomy and pyloroplasty (18% versus 5% and 13%, respectively; p ¼ 0.08). The use of promotility agents starting in the hospital and continuing after discharge was more significant in the botulinum group than in the pyloromyotomy and pyloroplasty patients (22% versus 5% and 15%, respectively; p ¼ 0.04). Compared with the Table 2. Median Operative Time by Group and Type of Operation
Procedure Ivor Lewis McKeon Transhiatal esophagectomy
Botulinum Pyloromyotomy Pyloroplasty (median (median (median time, minutes) time, minutes) time, minutes) 233 211 195
264 229 317
362 421 287
patients who underwent pyloromyotomy and pyloroplasty, by the 6-month follow-up more patients who received botulinum injections required endoscopic anastomotic dilation (41% versus 22% and 14%, respectively; p < 0.001) or endoscopic pyloric dilation (22% versus 4% and 2%, respectively; p < 0.001) and experienced more postoperative reflux symptoms (32% versus 12% and 13%, respectively; p ¼ 0.001). In the botulinum injection group, 16 patients underwent dilation (59%) with Savory dilators (60F) and 11 underwent dilation with a balloon dilating system (20-mm balloon at 2-3 atm of pressure). In the 16 patients who underwent dilation with the Savory dilators, both the anastomosis and pylorus were dilated, but the main focus was on the pylorus. Savory dilators were used in 97% of the patients who underwent pyloromyotomy, and 44% of the patients who underwent pyloroplasty that required dilation. There was no statistical difference between the groups regarding postoperative dumping (Table 3) at the 6-month followup.
Adjusted Comparisons After multivariable adjustment, the differences between the botulinum injection group and the other 2 groups persisted. Compared with the botulinum group, the pyloromyotomy and pyloroplasty groups had statistically smaller incidences of endoscopic anastomotic dilation (AOR, 0.39 and 0.23, respectively), endoscopic pyloric dilation (AOR, 0.17 and 0.04, respectively), and follow-up reflux (AOR, 0.28 and 0.29, respectively), but more had undergone preoperative radiation (AOR, 4.87 and 3.52, respectively). Additionally, pyloroplasty was associated with statistically fewer deaths (AOR, 0.27). In the botulinum injection group, 4 of the 9 30-day mortalities (44%) were related to respiratory issues, ie, aspiration pneumonia. Three of the 4 patients had undergone neoadjuvant chemotherapy and radiation treatment. Also, pyloromyotomy was statistically associated with less delayed gastric emptying (AOR, 0.20) and less use of promotility agents (AOR, 0.18). Adjusted length of stay was not significantly different between the groups (Table 4).
Comment Patients undergoing esophagectomy with a gastric conduit experience significant abnormal gastrointestinal
4
ELDAIF ET AL INTRAPYLORIC BOTULINUM INJECTION
Ann Thorac Surg 2013;-:-–-
Table 3. Postoperative Outcomes by Group Botulinum n ¼ 78
Outcome a
Pneumonia 30-day mortalitya (n, %) Leakagea (n, %) Chemotherapya (n, %) Radiationa (n, %) Delayed gastric emptying (n, %) Dilation of esophagusa (n, %) Dilation of pylorusa (n, %) Dumpinga (n, %) Refluxa (n, %) Promotility agentsa (n, %) Operative time (SD) Length of stay (d) Type of procedure (n, %) Ivor Lewis McKeown Transhiatal esophagectomy a
17 9 10 11 6 13 27 16 7 20 15 239 17.3
(21.8) (11.8) (12.8) (16.4) (8.8) (18.8) (40.9) (24.2) (11.7) (32.8) (22.7) (71) (18.0)
51 (65.4) 4 (5.1) 23 (29.5)
Pyloromyotomy n ¼ 45 8 3 8 12 11 2 9 2 1 5 2 312 13.9
(17.8) (6.7) (18.2) (30) (27.5) (4.6) (22.0) (4.9) (2.4) (12.2) (4.7) (97) (10.0)
18 (40.0) 4 (8.9) 23 (51.1)
Pyloroplasty n ¼ 199 21 5 20 46 33 25 26 2 15 23 28 373 17.7
p Value
(10.8) (2.6) (10.5) (24.6) (18.0) (12.9) (14.4) (1.1) (8.6) (13.1) (15.7) (76) (15.2)
0.051 0.009 0.36 0.23 0.04 0.08
Background. Intrapyloric botulinum toxin injection has emerged as a possible alternative to standard pyloric drainage procedures. Possible advantages include decreased operative time and less postoperative dumping and bile reflux symptoms. However, data are lacking to show its effectiveness versus standard drainage procedures. The purpose of this review is to compare the results in a prospective cohort of patients who received pyloric botulinum injection versus patients who received pyloromyotomy or pyloroplasty with esophagectomy. Methods. We performed a retrospective review of a prospective database of all patients who underwent an open esophageal resection at a single institution from 2005 through 2010. Three hundred twenty-two patients were divided into 3 groups for analysis: botulinum injection (n [ 78), pyloromyotomy (n [ 45), and pyloroplasty (n [ 199). We compared these groups with respect to duration of the procedure, presence of delayed gastric emptying on postoperative swallow studies, requirement of anastomotic dilation, requirement of pyloric dilation, use of postoperative promotility agents, and patient experience of postoperative symptoms of reflux or dumping, or both. Results. Patients receiving botulinum injections experienced similar delayed gastric emptying on postoperative radiologic evaluation as did patients undergoing
pyloromyotomy and pyloroplasty (16% versus 5% and 13%, respectively; p [ 0.14). Mean operative time was significantly shorter for the patients receiving botulinum as expected (239 minutes versus 312 minutes and 373 minutes, respectively; p < 0.001). However, more patients receiving botulinum and pyloric dilation (22% versus 4% and 2%, respectively; p < 0.001) experienced postoperative reflux symptoms (32% versus 12% and 13%, respectively; p [ 0.001) and used postoperative promotility agents (22% versus 5% and 15%, respectively; p [ 0.04). There was no statistical difference between the groups regarding postoperative dumping. Conclusions. Use of intrapyloric botulinum injection significantly decreased operative time. However, the patients receiving botulinum experienced more postoperative reflux symptoms, had increased use of promotility agents as well as a requirement for postoperative endoscopic interventions, and postoperative dumping was not reduced by the reversible procedure. Intrapyloric botulinum injection should not be used as an alternative to standard drainage procedures. Pyloromyotomy appears to be the drainage procedure of choice to accompany an esophagectomy.
E
The largest randomized controlled trial to date demonstrated significantly shorter gastric emptying times in patients undergoing pyloroplasty versus patients not undergoing drainage procedures, as well as earlier tolerance of solid diet and decreased symptoms with meals at 6 months [1]. Likewise, a literature review of 9 randomized controlled trials with meta-analysis by Urschel and colleagues [2] concluded that pyloroplasty compared with no intervention reduced the incidence of gastric outlet obstruction but had little effect on operative morbidity or mortality and late postoperative gastric conduit function. There remains no consensus because numerous other reports from high-volume centers have described no decrease in delayed gastric emptying after drainage procedures such as pyloroplasty or pyloromyotomy [3, 4]. Furthermore, detractors emphasize the increased prevalence of bile reflux, esophagitis, and dumping syndrome after these emptying procedures [5, 6].
sophagectomy with gastric conduit reconstruction is the preferred operation for esophageal cancer and certain benign diseases. However, the sequelae of gastric interposition can contribute to significant morbidity and mortality in this population. Respiratory complications associated with aspiration of gastrointestinal material and subsequent pneumonias are the most frequent complications in the early postoperative period [1]. Gastric outlet obstruction is recognized as the predominant risk factor for aspiration. Currently, the efficacy of pyloric drainage procedures in preventing delayed gastric emptying remains widely debated.
Accepted for publication Nov 11, 2013. Presented at the Fifty-ninth Annual Meeting of the Southern Thoracic Surgical Association, Naples, FL, Nov 7–10, 2012. Address correspondence to Dr Miller, Emory University Clinic, 1365 Clifton Rd NE, Atlanta, GA 30322; e-mail: daniel.miller@ emoryhealthcare.org.
Ó 2013 by The Society of Thoracic Surgeons Published by Elsevier Inc
(Ann Thorac Surg 2013;-:-–-) Ó 2013 by The Society of Thoracic Surgeons
0003-4975/$36.00 http://dx.doi.org/10.1016/j.athoracsur.2013.11.026
2
ELDAIF ET AL INTRAPYLORIC BOTULINUM INJECTION
These disparities paved the way for the introduction of intrapyloric botulinum injection as an alternative to the irreversible gastric emptying procedures and as a solution to early delayed gastric emptying and dumping and reflux secondary to a wide open pylorus. Kent and colleagues [7] in 2007 were the first to show in a pilot study of 12 patients who had undergone minimally invasive esophagectomy (MIE) that intrapyloric injection of botulinum toxin prevented delay in gastric emptying or caused aspiration pneumonia. The purpose of this review is to compare the results in a large prospective cohort of patients who underwent 1 of 3 drainage procedures: pyloric botulinum injection, pyloromyotomy, or pyloroplasty at the time of open esophageal resection.
Material and Methods Study Design and Surgical Methods With approval from our institutional review board, we performed a retrospective review of a prospective database of all patients who underwent esophagectomy at a single institution, Emory University Hospital, from 2005 through 2010, using the stomach as the reconstruction conduit. Three hundred twenty-two patients had open esophageal resection with gastric conduit reconstruction by an Ivor Lewis, McKeown, or transhiatal technique. The study population was then divided into 3 groups for analysis based on the pyloric drainage procedure: botulinum injection (n ¼ 78), pyloromyotomy (n ¼ 45), or pyloroplasty (n ¼ 199). All patients received a drainage procedure. The type of resection and drainage procedure performed was based on the preference of the surgeon and the tumor location. Pyloromyotomy and pyloroplasty procedures were performed in standard fashion: A Heineke-Mikulicz pyloroplasty was performed with a full-thickness 2- to 3-cm longitudinal incision extending on each side of the pyloric vein (vein of Mayo) and closed vertically in a single layer with either 3-0 silk or 3-0 polydiaxanone sutures, whereas the pyloromyotomy was performed as a Freder-Ramstedt procedure with a 3-cm longitudinal incision at the pyloric vein through the serosa and muscle fibers, with the final layer of muscle divided with sharp dissection to allow bulging of the mucosa above the serosa level. Extraluminal pyloric injection was performed using 100 U of botulinum toxin in 2 mL of normal saline (NS) (equally distributed into 25 U per injection into 4 separate quadrants) after gastric mobilization. The dose of botulinum toxin was the same used by Cerfolio and colleagues [8]—100 U in 4 mL NS in 4 quadrants—but less than that used by Kent and colleagues [7]—200 U in 5 mL NS, also in 4 quadrants. Botulinum toxin is an inhibitor of cholinergic neuromuscular transmission and has been used to treat spastic disorders of both striated and smooth muscles by local injections [9]. This has been reported to be successful in patients with gastroparesis secondary to diabetes and idiopathic causes. The effects of the botulinum toxin usually last 3 to 5 months [10].
Ann Thorac Surg 2013;-:-–-
Postoperative Evaluation and Follow-Up All patients were managed with nasogastric tubes postoperatively for 3 days with a cervical anastomosis and for 5 days with a thoracic anastomosis. All patients underwent a fluoroscopic water-soluble contrast agent or barium contrast agent swallow study, or both, between postoperative days 5 and 7 to access anastomotic integrity and gastric emptying in relation to the conduit and the pylorus. Patients with a cervical anastomosis or those who received preoperative radiotherapy were also evaluated further for the risk of aspiration with a modified barium swallow study performed by one of our speech pathologists. All patients were given esomeprazole (Nexium, AstraZeneca Pharmaceuticals) starting on postoperative day 1 and promotility agents, either metoclopramide (Reglan, Baxter Pharmaceuticals) or erythromycin if delayed gastric emptying was demonstrated on the radiologic swallow study. We compared the 3 groups with respect to duration of the procedure, delayed gastric emptying on postoperative radiologic studies, requirement of anastomotic dilation, requirement of pyloric dilation, use of postoperative promotility agents, and patient experience of postoperative symptoms. A standard postoperative review of systems was completed by each patient at their 6-month follow-up clinic visit. Gastrointestinal symptoms focused on for our analysis were dysphagia, odynophagia, chest or abdominal pain, reflux, and dumping.
Statistical Evaluation Unadjusted pyloric group comparisons were made using c2 tests and analysis of variance methods for categorical and continuous variables, respectively. Logistic regression was used to determine the association between each binary clinical end point and pyloric drainage group, adjusted for age, sex, preoperative use of chemotherapy and radiation treatment, and surgical technique. Adjusted odds ratio (AOR) with 95% confidence interval (CI) was used as the measure of association. The botulinum group was the reference group. Using the same adjustment methodology, general linear modeling was used to model the lone numerical end point—length of stay. All analyses were performed using SAS, version 9.3 (SAS Institute, Cary, NC) and all tests were evaluated at the 0.05 alpha level. Missing data among predictors were present in less than 2% of cases. Missing data were more prevalent in outcomes; on average, outcome variables had about 8% missing data (range, 1.2%-14.3%). A multiple imputation procedure was implemented for cases with missing predictors so that the maximum number of cases could be used. However, outcomes were not imputed when missing, and thus a complete case analysis was performed for these cases.
Results Unadjusted Analyses Preoperative characteristics are included in Table 1. The majority of procedures (86%) were performed for
Ann Thorac Surg 2013;-:-–-
3
ELDAIF ET AL INTRAPYLORIC BOTULINUM INJECTION
Table 1. Preoperative Characteristics Botulinum n ¼ 78
Variable a
62.2 27 34 35
Patient age (SD) Female sex (n, %) Preoperative radiationa (n, %) Preoperative chemotherapya (n, %) a
(12.0) (34.6) (44.2) (45.5)
Pyloromyotomy n ¼ 45 60.5 13 21 22
(10.0) (28.9) (46.7) (48.9)
Pyloroplasty n ¼ 199 62.2 39 56 80
(11.0) (19.6) (28.3) (40.4)
p Value 0.64 0.025 0.009 0.50
There are significant mean differences by group (p < 0.001) and by operative type (p < 0.001).
SD ¼ standard deviation.
esophageal cancer; 11% were performed for benign disease, including achalasia, stricture, and perforation; and 3% were performed for Barrett’s high-grade dysplasia. The number of patients with benign disease was not statistically different among the groups. Approximately 42% of the patients with cancer underwent neoadjuvant treatment, with significantly fewer patients receiving preoperative radiation in the pyloroplasty group. Intraoperative data are included in Table 2. Mean operative time was significantly shorter for the botulinum group in comparison with the pyloromyotomy and pyloroplasty groups (239 minutes versus 312 minutes and 373 minutes, respectively; p < 0.001). Mean operative time was further subclassified (Table 2) to demonstrate if the length of operation was different for the 3 esophagectomy techniques. Gastric conduit width and anastomosis location were similar between the groups. The average width of the gastric conduits was 5 to 8 cm for both the intrathoracic and cervical reconstructions. The location of the intrathoracic anastomosis was above the azygous vein, with distance from the incisors of 21 to 23 cm, and the cervical anastomosis was located at 17 to 19 cm from the incisors. A delay on the postoperative swallow study was defined according to the method described by Cerfolio and colleagues [8] to unify the definition in the literature. Patients receiving botulinum injections experienced similar delayed gastric emptying, but the trend was toward significantly slow emptying on postoperative swallow studies compared with patients who underwent pyloromyotomy and pyloroplasty (18% versus 5% and 13%, respectively; p ¼ 0.08). The use of promotility agents starting in the hospital and continuing after discharge was more significant in the botulinum group than in the pyloromyotomy and pyloroplasty patients (22% versus 5% and 15%, respectively; p ¼ 0.04). Compared with the Table 2. Median Operative Time by Group and Type of Operation
Procedure Ivor Lewis McKeon Transhiatal esophagectomy
Botulinum Pyloromyotomy Pyloroplasty (median (median (median time, minutes) time, minutes) time, minutes) 233 211 195
264 229 317
362 421 287
patients who underwent pyloromyotomy and pyloroplasty, by the 6-month follow-up more patients who received botulinum injections required endoscopic anastomotic dilation (41% versus 22% and 14%, respectively; p < 0.001) or endoscopic pyloric dilation (22% versus 4% and 2%, respectively; p < 0.001) and experienced more postoperative reflux symptoms (32% versus 12% and 13%, respectively; p ¼ 0.001). In the botulinum injection group, 16 patients underwent dilation (59%) with Savory dilators (60F) and 11 underwent dilation with a balloon dilating system (20-mm balloon at 2-3 atm of pressure). In the 16 patients who underwent dilation with the Savory dilators, both the anastomosis and pylorus were dilated, but the main focus was on the pylorus. Savory dilators were used in 97% of the patients who underwent pyloromyotomy, and 44% of the patients who underwent pyloroplasty that required dilation. There was no statistical difference between the groups regarding postoperative dumping (Table 3) at the 6-month followup.
Adjusted Comparisons After multivariable adjustment, the differences between the botulinum injection group and the other 2 groups persisted. Compared with the botulinum group, the pyloromyotomy and pyloroplasty groups had statistically smaller incidences of endoscopic anastomotic dilation (AOR, 0.39 and 0.23, respectively), endoscopic pyloric dilation (AOR, 0.17 and 0.04, respectively), and follow-up reflux (AOR, 0.28 and 0.29, respectively), but more had undergone preoperative radiation (AOR, 4.87 and 3.52, respectively). Additionally, pyloroplasty was associated with statistically fewer deaths (AOR, 0.27). In the botulinum injection group, 4 of the 9 30-day mortalities (44%) were related to respiratory issues, ie, aspiration pneumonia. Three of the 4 patients had undergone neoadjuvant chemotherapy and radiation treatment. Also, pyloromyotomy was statistically associated with less delayed gastric emptying (AOR, 0.20) and less use of promotility agents (AOR, 0.18). Adjusted length of stay was not significantly different between the groups (Table 4).
Comment Patients undergoing esophagectomy with a gastric conduit experience significant abnormal gastrointestinal
4
ELDAIF ET AL INTRAPYLORIC BOTULINUM INJECTION
Ann Thorac Surg 2013;-:-–-
Table 3. Postoperative Outcomes by Group Botulinum n ¼ 78
Outcome a
Pneumonia 30-day mortalitya (n, %) Leakagea (n, %) Chemotherapya (n, %) Radiationa (n, %) Delayed gastric emptying (n, %) Dilation of esophagusa (n, %) Dilation of pylorusa (n, %) Dumpinga (n, %) Refluxa (n, %) Promotility agentsa (n, %) Operative time (SD) Length of stay (d) Type of procedure (n, %) Ivor Lewis McKeown Transhiatal esophagectomy a
17 9 10 11 6 13 27 16 7 20 15 239 17.3
(21.8) (11.8) (12.8) (16.4) (8.8) (18.8) (40.9) (24.2) (11.7) (32.8) (22.7) (71) (18.0)
51 (65.4) 4 (5.1) 23 (29.5)
Pyloromyotomy n ¼ 45 8 3 8 12 11 2 9 2 1 5 2 312 13.9
(17.8) (6.7) (18.2) (30) (27.5) (4.6) (22.0) (4.9) (2.4) (12.2) (4.7) (97) (10.0)
18 (40.0) 4 (8.9) 23 (51.1)
Pyloroplasty n ¼ 199 21 5 20 46 33 25 26 2 15 23 28 373 17.7
p Value
(10.8) (2.6) (10.5) (24.6) (18.0) (12.9) (14.4) (1.1) (8.6) (13.1) (15.7) (76) (15.2)
0.051 0.009 0.36 0.23 0.04 0.08
