| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2003;76:1694-1698
© 2003 The Society of Thoracic Surgeons
a Divisions of Cardiothoracic Surgery and Gastroenterology, Washington, DC, VAMC, and George Washington University Medical Center, Washington, DC, USA
* Address reprint requests to Dr Alexander, Cardiothoracic Surgery, 2150 Pennsylvania Ave, Suite 6B, Washington, DC 20037;
e-mail: ealexander{at}mfa.gwu.edu
Presented at the Forty-eighth Annual Meeting of the Southern Thoracic Surgical Association, San Antonio, TX, Nov 8–10, 2001.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and methodsResultsCommentDiscussionReferences |
|---|
METHODS: Experience with PEG in consecutive patients presenting with new esophageal cancer from March 1991 to March 2001 was reviewed retrospectively. PEG was planned in 119 of 179 (66%) of these patients excluding those presenting moribund and those for whom early resection was planned. The PEG was placed using an endoscopic method with wire-guided endoscopic bougienage or laser ablation or both as needed. Success of placement, requirement for dilatation and ablation, PEG-related complications, tolerance of enteral feeds, and impact on therapy were evaluated.
RESULTS: PEG placement was possible in 87% of patients (103 of 119). Dilatation or laser ablation or both was required in 46% (47 of 103). There was no procedure-related mortality. Thirty-day mortality was 13.5%. Major PEG-related complications were observed in 4% (4 of 103) and minor PEG-related complications in 12% (12 of 103). PEG removal was required in 4 patients and interruption of enteral feeds required in 33 (32%). No instances of esophageal disruption or tumor inoculation metastasis were noted. PEG takedown and site closure at the time of operation was uncomplicated and use of the stomach as an esophageal substitute was possible in all 61 resected patients. Rates of anastomotic leak, stricture, and gastric emptying delay were similar to those for patients proceeding to resection without prior PEG (leak: PEG = 8% [5 of 61] versus non-PEG = 10.5% [2 of 19]), (stricture: PEG = 37% [22 of 61] versus non-PEG = 32.5% [6 of 19]), (delay: PEG = 9.8% [6 of 61] versus non-PEG = 10.5% [2 of 19]). Analysis of variables showed PEG to be significantly related to attainment of target doses of chemoradiotherapy (p = 0.034), and survival at 12 months (p = 0.02).
CONCLUSIONS: PEG in esophageal cancer is safe and useful and does not compromise the stomach or esophagogastric anastomosis. Further study is required to define the efficacy of PEG as a means of nutritional support and its impact on survival.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and methodsResultsCommentDiscussionReferences |
|---|
Since its introduction in 1979 by Gauderer and Ponsky [2] percutaneous endoscopic gastrostomy (PEG) has been widely employed in a variety of clinical settings [3]. Estimates that between 180,000 and 200,000 PEGs are performed annually in the United States make this intervention the second most common indication for upper gastrointestinal tract endoscopy in hospitalized patients in the United States [4]. PEG has not been widely used in esophageal cancer because of concerns about the ability to routinely traverse obstructing cancer, the subsequent suitability of the stomach as an esophageal replacement, and the potential for gastric or abdominal wall tumor inoculation metastasis. PEG placement has been typically reserved for palliative nutrition where little further therapy is feasible. Reports of PEG as a supportive modality in patients with head and neck cancer generally support the safety and utility of this means of nutritional support [5–7]. Reported complications of PEG placed for a variety of indications are infrequent with a procedure-related mortality rate of 0.3% to 1% and a morbidity rate of 3% to 5.9% [3].
We initiated the routine use of PEG as an early intervention for obstructing esophageal cancer in potentially salvageable patients for whom early resection was not planned. Percutaneous endoscopic gastrostomy was intended to be a nutritional bridge during multimodal therapy including both neoadjuvant and primary chemoradiotherapy.
|
Patients and methods |
|---|
|
TopAbstractIntroductionPatients and methodsResultsCommentDiscussionReferences |
|---|
A retrospective review of this experience was conducted utilizing in and outpatient medical records including endoscopic and operative reports and tumor registry data. Recorded data included presentation, functional status, available nutritional indicators, procedural details, procedure-related complications, success of placement, tube function and dysfunction, tolerance of enteral feeds, and impact on therapy.
Technique
PEG placement was attempted as an isolated procedure in 74 of 119 patients (62%) and in conjunction with operative staging in 45 of 119 (38%). Percutaneous endoscopic gastrostomy was placed using an endoscopic guided "push" technique. Wire-guided endoscopic bougienage or laser ablation or both were performed before PEG placement to establish esophageal luminal patency as needed. As a primary procedure PEGs were placed using midazolam and fentanyl sedation. In conjunction with operative staging PEG was placed with general anesthesia after thoracoscopy or laparoscopy. Less operative impact on the stomach was anticipated with PEG than with laparoscopic gastrostomy. All patients received antibiotic prophylaxis.
Percutaneous endoscopic gastrostomy was used to provide or supplement enteral nutrition using standard commercially available tube feedings. Tolerance and cessation of tube feedings and instances of tube dysfunction were noted.
Subsequent course
Weight change data over the course of therapy were available in 101 patients. Tolerance and outcome of chemoradiotherapy was recorded as was survival. Records were carefully reviewed for the occurrence of tumor inoculation metastasis, a previously described phenomenon after PEG placement for head and neck and esophageal cancer patients [8, 9]. In 61 patients progressing to resection, the gastric PEG site was oversewn with a purse string suture; 38 of 61 resected patients (62%) underwent transhiatal esophagectomy and 23 (38%) had a thoracotomy. The anastomosis was hand sewn and placed in the neck in all patients. Suitability of the stomach as an esophageal substitute was noted, including rates of anastomotic leak and stricture and gastric emptying delay. These were compared with patients proceeding to resection without prior PEG. Nutritional input was assessed in patients proceeding to operation as these patients provided a definable end point for analysis of nutritional status.
Statistical methods
The 
2 test and Fisher’s exact test were used to test whether relationships existed between nominal variables. Survival was measured from the date of diagnosis to the date of death or October 1, 2001. Actuarial survival was calculated using the Kaplan-Meier method and comparisons made using the log rank test. Significant predictors from a univariable analysis of independent variables were subjected to multivariable analysis.
|
Results |
|---|
|
TopAbstractIntroductionPatients and methodsResultsCommentDiscussionReferences |
|---|
|
Tube feeding
Tube feedings were given through the PEG in 96 of the 103 patients (96%) in whom placement was possible. In the remaining 7 patients, feeds were avoided because of tolerance of oral feeds in 4 and critical illness in 3. Feedings were regularly adjusted to compensate for changes in oral intake as responses to therapy occurred including decreasing or increasing dysphagia, esophagitis, and mucositis. Diarrhea or abdominal distention was noted during the course of 42 patients (41%). Interruption of enteral feeds was required at some point in 33 patients (32%) during tube feeding.
Median duration of PEG placement was 15 weeks (range, 3 days to 10 months). There was at least one instance of tube dysfunction, typically clogging, in 23 of the 103 patients (22.3%).
Nutritional outcome
Weight loss during chemoradiation generally declined and the mean percent weight loss for the 61 patients proceeding to surgery was 3.8% (range, 0% to 9%) from the time of PEG placement. Serum albumin data were available in 51 of the 61 patients. Mean serum albumin at PEG placement was 3.3 g/dL (range, 1.6 to 4.4 g/dL). Mean serum albumin immediately before surgery increased to 3.6 g/dL (range, 2.8 to 4.8g/dL), an improvement that was statistically significant (p = 0.001).
Completion of chemoradiotherapy
Attainment of target doses of chemoradiotherapy was noted in 71 of the 103 patients (70%) in whom PEG placement was successful. This compared favorably with the 16 patients in whom PEG placement was not possible, in whom only 6 of 16 (37.5%) attained target doses. This difference was significant (p = 0.03). Analysis of multiple variables potentially associated with completion of therapy including age, weight loss at presentation, T stage, N stage, concomitant laser therapy, and PEG placement showed PEG to be significantly related to attainment of target doses of chemoradiotherapy (p = 0.034).
Inoculation metastasis
No instance of inoculation metastasis was noted at the gastrostomy site, the abdominal wall, or the stomach.
Surgical outcomes
Tube removal and gastric site closure were uncomplicated and the stomach was an adequate esophageal substitute in all 61 patients proceeding to resection. Rates of anastomotic leak, stricture, and gastric emptying delay were similar to contemporaneous patients proceeding to resection without prior PEG (leak: PEG = 8% [5 of 61] versus non-PEG = 10.5% [2 of 19]; stricture: PEG = 37% [22 of 61] versus non-PEG = 32.5% [6 of 19]; and delay: PEG = 9.8% [6 of 61] versus non-PEG = 10.5% [2 of 19]).
Survival
Actuarial survival for all patients in whom PEG was planned and in whom it was successful and unsuccessful are shown in Figure 1.
Comparison of successful and unsuccessful placement shows a trend toward an advantage for PEG placement (log rank p = 0.086).
|
|
Comment |
|---|
|
TopAbstractIntroductionPatients and methodsResultsCommentDiscussionReferences |
|---|
The stomach was not compromised as an alimentary conduit and with the technique of PEG employed there was no evidence of tumor implantation or inoculation metastasis in the study cohort.
Is "nutritional bridge" a meaningful concept in the management of esophageal cancer? Prior studies of parenteral nutrition in esophageal cancer suggest that it may be [13, 14]. A recent American Gastroenterological Association Medical position statement endorses the use of parenteral nutrition for this indication [15]. Enteral nutrition is thought to have numerous advantages over parenteral nutrition, though this is not universally accepted [16, 17]. A multitude of possible selection biases and confounding variables in a noncontrolled retrospective study such as this make the survival data difficult to interpret. Chief among these is that PEG placement may be systematically limited in more advanced lesions.
Given the limitations of a retrospective review, the favorable impact of PEG on completion of chemoradiotherapy and survival in this cohort of patients should be further investigated.
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and methodsResultsCommentDiscussionReferences |
|---|
DR MARGOLIS: We looked at the albumin patients at the time of diagnosis, at the time of inserting the PEG, and that was a mean of 3.3, and we rechecked that again at the time of surgery and that had increased to a mean of 3.6.
DR DANIEL L. MILLER (Rochester, MN): I have several questions especially in regard to your conclusions. You stated in your conclusion slide that a PEG had no effect on the stomach as a conduit but you failed to present any data in your talk to support this conclusion. I would appreciate your comment please. A significant percentage of our patients are undergoing neoadjuvant treatment prior to a planned esophagectomy. We usually place a feeding jejunostomy prior to initiation of the neoadjuvant treatment in those patients. You can do a minilaparotomy at the time of the jejunostomy placement and do an abdominal inspection to see if the patient has advanced disease or carcinomatosis. I have gotten burned a couple times when a PEG was placed elsewhere and when the patient came back for resection they had destruction of the right gastroepipolic artery, which prevented use of the stomach, and 1 patient died from sepsis secondary to an infection from the PEG skin site.
I would like to know if you could comment on the follow-up of your patients who went on to resection. Were there any complications related to the PEG such as wound infection or that you could not use the stomach and had to use the colon instead?
DR MARGOLIS: We were able to use the stomach as a conduit in all these 61 patients that had proceeded to surgery. I did present a slide where we assessed the suitability of the stomach as a conduit, looking at three factors. We looked at any anastomotic strictures, any anastomotic leaks, and any increased evidence of gastric emptying delay, comparing those patients who had a PEG placed with those patients who did not have a PEG placed. We showed that there was no difference between those two groups of patients. We had no problems taking down the PEG, we had no problems in using the stomach as a conduit. There was no increased incidence of wound infections either between the two groups.
DR JOE B. PUTNAM, JR (Houston, TX): I enjoyed your presentation and your enthusiasm for PEGs, but I believe that your two populations were two difficult groups to compare: both groups were at the "extremes" of care. One group was very poor prognosis and the PEG was placed for palliation to enhance end-of-life care and comfort. The other group was for preoperative nutritional preparation. At The University of Texas M. D. Anderson Cancer Center, the thoracic surgical faculty have not been very enthusiastic about PEG placement prior to esophagectomy simply because the need is not consistent—even for patients receiving preoperative chemotherapy and radiation therapy. As well, the location cannot always be predicted and involvement of the midline fascia has yielded significant wound complications with both skin and fascial dehiscence, and even evisceration on one occasion. Injury to the vascular supply of the gastric conduit occurred in 1 of our patients as a result of PEG placement, precluding the use of the stomach as a conduit.
Your institution may be unique in that your gastroenterologists and surgeons have jointly created a detailed and methodical process to prepare these preoperative patients with PEG placement. In many general communities, however, this process is not followed. Perhaps the process for routine PEG is just blow some air in the stomach and "slam a tube in there." I believe that the gastroenterologist rarely if ever confers with the surgeon prior to PEG placement. This casual placement of a PEG (often not even warranted) may provide more morbidity than benefit.
DR MARGOLIS: Thank you. What you say is a hundred percent correct. We have a unique setup at the VA Hospital. We have a nutritional unit with an endoscopist who is particularly devoted to doing this in our patients with esophageal cancer. So these PEGs are placed with a lot of care, realizing that we do plan to absolutely use the stomach as a conduit.
|
References |
|---|
|
TopAbstractIntroductionPatients and methodsResultsCommentDiscussionReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
