Discussion

Article Outline

• Copyright

Dr Henning A. Gaissert (Boston, Mass). I have no disclosures.

To summarize, Dr Freeman and associates report 21 patients in whom acute postoperative esophageal leaks of fistulae were treated with a self-expanding Polyflex stent after failure of at least 1 operative repair. A seal was achieved in 20 of these 21 patients, and stenting lasted for a mean of 51 days. There were 3 complications. One patient died from disruption of an aortic repair, 1 patient had disruption of an esophageal repair, and 1 patient had a stricture.

I have 2 comments and several questions. The first point is that successful sealing of the leak is gratifying and important, but it is not the complete management of an esophageal leak. Other important components are the drainage of extraluminal fluid collections, decortication for pleural sepsis, and separation of the esophagus from adjacent structures, such as the aorta or airway, with vascularized tissue.

The second point is that self-expanding stents have important disadvantages related to the radial force necessary to maintain their position. If this radial force is too high, the stent erodes the mucosa and causes ulceration and stricture, and if it is too low, the stent slips and migrates, which occurred in 24% of patients. Placed across the esophagogastric junction, the stent might cause debilitating reflux and a stricture above. This treatment is therefore not without complications.

Now to my questions. There were 29 patients in the 2½ years with postoperative esophageal leaks and fistulae. I realize many of these were referred to you, but is there anything that you have learned in this period from the primary repair that would lead you to improve that management to decrease the need for stent intervention?

Second, do you trust the scar that results from stenting alone? I note that one of your patients had an erosion after an aortic repair through the aortic suture line, and there were 4 patients with airway fistulae, and I wonder whether in each of these patients you simply trusted the scar caused or occasioned after stent placement.

Third, there was an average of 6 days that passed before you knew whether the stent sealed the hole. That is a long time not knowing whether your management succeeded. Would you consider obtaining contrast studies at the time of stent placement?

This was a very nice article and very well presented, and I thank the Association for the opportunity to discuss it.

Dr Freeman. Thank you, Dr Gaissert.

You are correct in that only 3 of these patients came from our practice. The remaining patients either came from a gastroenterologist at our institution or were transferred in 16 cases. I think that looking at these patients brings to light the things that we are taught, that repairs should try and be without tension and to have some kind of vascularized tissue applied to them. We usually use muscle. In some of these patients, this had not been done initially and might have contributed to the failure.

As far as trusting the repair, we have had no problems with recurrent fistulae. That being said, 51 days was our mean time until stent removal, and I think we were very leery of taking these out too soon. In fact, we wanted to see excellent nutritional parameters in the patient and resolution of any infectious or septic parameter, including resolution of associated organ failure. Therefore I think we were very conservative when we removed these, and hopefully that has contributed to the lack of recurrence.

As far as knowing that the stent had sealed, to be honest with you, I think we have a pretty good idea of that at the time of stent placement. Because we were doing a study, we wanted a contrast esophagram for every patient. However, after stent placement, we obviously would perform another flexible esophagoscopy and would insufflate, and you get a pretty good idea that the leak has sealed if you have a chest tube in place, or if you have an airway fistula, you are going to do a bronchoscopy at the same time. Therefore I think some of the esophagrams were delayed because of the patients’ other comorbidities, but in general, we usually knew that we had a good seal on leaving the operating room.

Dr Jonathan C. Nesbitt (Nashville, Tenn). Richard, I enjoyed your presentation. This is a difficult problem in a select cohort of patients, and I congratulate you on your results.

Dr Gaissert mentioned the problem with migration, and certainly I think with this particular stent it is one of the biggest issues, and, as you have shown, 24%, that is a relatively high rate.

How do you size the stent, not only the diameter but also the length? Also, with regard to the actual perforation or the fistula, what do you believe is the optimal position, either high or low, in the esophagus for this particular stent? That is my first question.

Dr Freeman. We tend to oversize the diameter of the stent slightly and the length significantly, and we do that at the time of endoscopy by insufflating and also by using fluoroscopy. Therefore it is more of a feel. It also depends on where the fistula is located. Obviously you cannot do that if it is extremely proximal or extremely distal.

Dr Nesbitt. Certainly you cannot because the sizes vary significantly, and there is quite a difference between the diameters of each stent. The stents might slide and migrate, and sealing of the fistula or perforation is directly related to the stability of the stent position.

Dr Freeman. This article does not address acute perforations, but in these patients who have all had operations and some sort of repair, they do have more scarring, and I think you have a little bit of an advantage in that respect.

Dr Nesbitt. My next question pertains to the disparity between removal times of the stents. What do you believe is the optimal time for removal?

Dr Freeman. Again, that is a very individualized decision. We were very conservative and still are. We like to see positive nitrogen balance, a good prealbumin value, resolution of any infectious or septic problems, and obviously no other signs of leak. We have had several persons in this series in whom we could actually monitor the healing either because it was an airway fistula or in one case a cervical esophagocutaneous fistula, and that made it a little easier, but it is a very individualized decision and it is based mainly on the patient’s global status.

Dr Nesbitt. Finally, did you perform follow-up studies once you removed the stent, and if so, did you have any leaks, or did you have to replace the stent?

Dr Freeman. We performed esophagoscopy in the operating room before and after the removal of each stent, and a minimum of 24 to 48 hours after stent removal, the patients had another contrast esophagram. No patient had a residual leak or fistula in this series.

Dr Bryan F. Meyers (St Louis, Mo). Congratulations on your article.

You mentioned oversizing, and with these stents, if you oversize them too much, then you get an infolding at the top of the stent that is difficult to pop out, and then you are forced to stick a dilator or a balloon down there and blow up a balloon in a perforated or leaking esophagus. I just wondered whether you have encountered that problem. And just following up on the last question, how specifically would you pick the size to make it not migrate but make it not so large that you end up with an infolding and more of a tendency to leak around it?

Dr Freeman. In general, getting back to size, you really get a feel when you insufflate and distend the esophagus, and we generally go very large to make sure we have enough radial force to seal this, and I do not think we have used anything smaller than a size 25 in these patients. As far as the folding, that is a very frustrating problem. In early patients we would try and use a balloon dilator. The last few times that has happened, I basically left it alone and then did another examination in about 24 to 48 hours, and the fold comes out and opens up over time as the stent warms up.

Dr Rafael S. Andrade (Minneapolis, Minn). I commend you for trying to shift the therapeutic paradigm to this problem.

I have a question for you in terms of the esophagram. We have seen, particularly after an anastomotic leak, that the esophagram is negative in the upright position, but in the supine position contrast trickles distally around the stent and out of the fistula. Now, that is not necessarily a failure because you might still be slowing down the leak and eventually the patient will do well. I want to have your opinion on how you are doing your esophagrams after stent placement.

Dr Freeman. First of all, I think that hopefully has been a minimal problem because we do oversize significantly in diameter of the stent. We do wait generally a minimum of 48 hours before we do an esophagram, and they do a standard esophagram, which is sitting and lying down, and it is a video esophagram.

Dr Stephen G. Swisher (Houston, Tex). As you know, there is a variation in severity of esophageal leaks, and I was just wondering whether you could comment on how many of these leaks were contained and how many were free flowing, and how many of these patients were septic or were looking pretty good?

Dr Freeman. One indication of how sick they were is that most of them were transferred from other places. They were critically ill. Their leaks were, for the most part, drained, at least into the pleural or peritoneal space. These were not small, contained postoperative leaks that just needed a few more days to heal. These were fairly significant leaks, and I think you can tell that also by the associated procedures that we had to do to try and remove areas of infection.