Congenital Duodenal Anomalies in the Adult

Background  Duodenal anomalies are defects in embryologic development and usually present as gastric outlet obstruction in infancy or early childhood. Occasionally, they remain asymptomatic until adulthood and, because they are unusual, may not be diagnosed.

Hypothesis  Based on current experience and review of the literature, recognition of diagnosis and the preferred methods of treatment of duodenal anomalies can be recommended.

Design  Retrospective study of congenital duodenal anomalies in adults.

Setting  Tertiary care university medical center.

Patients  Twenty-nine patients were observed and treated between 1983 and 1999 (19 women and 10 men; mean ± SD age, 52 ± 16 years). Twenty patients had duodenal webs, 7 had annular pancreata, and 2 had both. Nausea, vomiting, abdominal pain, and weight loss were predominant symptoms in all groups. Peptic ulceration occurred in 13 of 20 patients with webs but in none of those with annular pancreata or combined anomaly.

Main Outcome Measures  Surgical outcomes including postoperative complications, deaths, and resolution of preoperative symptoms.

Results  The treatment for patients with duodenal webs was transduodenal web excision and duodenoplasty in 19 of 22. Patients with annular pancreata were treated by transection of the annulus and duodenoplasty (n = 4) and proximal duodenal bypass (n = 3). There were no operative deaths, but 44% of patients had some complications. No pancreatic fistulas occurred in patients who had division of an annular pancreas. Outcome was considered excellent or good in 17 of 20 patients with duodenal webs, 4 of 7 with annular pancreata, and 2 of 2 with the combined anomaly.

Conclusions  Duodenal anomalies are rare in adults. Duodenal webs are best managed by transduodenal excision and duodenoplasty. Annular pancreas is generally best treated by duodenal bypass to the distal duodenum or the jejunum. Annulus division can be carried out if the annulus is extramural, without duodenal stenosis, and if access to the pancreaticobiliary sphincters is necessary.

CONGENITAL duodenal anomalies are rare lesions and originate in the early embryologic development of the foregut. Whereas the primitive foregut undergoes lengthening and rotation, the hepatobiliary and pancreatic anlagen begin as buds or diverticula at the middle of the duodenum and similarly grow and rotate. During this period, duodenal atresias, intraluminal webs, annular and ectopic pancreata, and malrotations of various types develop. The delayed presentation of these anomalies in the adult is difficult to explain, but the presence of a dilated stomach and a proximal duodenal bulb with a patulous pylorus suggests a progressive loss of compensatory peristaltic action to overcome a small duodenal aperture or narrowing of the descending duodenum. Most studies of these lesions are single case reports or small series, which do not allow a single surgeon to accumulate extensive experience; therefore, reliance on the combined experience of others in recognition and appropriate management has been the norm. This article presents a recent series of adult patients with duodenal webs, annular pancreata, or a combination of these 2 anomalies. Recognition and a rational approach to correction of these problems is based on the appropriate treatment for each patient and a knowledgeable expectation of the outcome based on the chosen therapy.

Hospital records from January 1, 1983, to December 31, 1999, were searched for patients at the Indiana University Medical Center Hospitals, Indianapolis, who were older than 18 years and had International Classification of Diseases, Ninth Revision,¹ codes for congenital small-bowel atresia (code 751.1), duodenal web (code 751.5), and annular pancreas (code 751.7). Medical charts were reviewed for patient demographics, past medical and surgical histories, signs and symptoms at presentation, location and type of duodenal anomaly, diagnostic studies, operation performed, associated anomalies, complications, mortality, and patient outcome.

Congenital duodenal anomalies were identified in 29 patients. Seven patients with duodenal webs have been reported previously.² There were 19 women and 10 men (mean ± SD age, 52 ± 16 years). Of these patients, 20 were diagnosed as having duodenal webs, 7 as having annular pancreata, and 2 as having both (Table 1).

The duodenal web group consisted of 14 women and 8 men, including 2 patients with both anomalies. Four patients had previously undergone cholecystectomy, 2 had undergone antrectomy, and 3 had undergone gastrojejunostomy or duodenojejunostomy for gastric outlet obstruction. Most patients with duodenal webs presented with nausea, vomiting, epigastric pain, and early satiety. The duration of symptoms ranged from 1 month to 50 years, with a mean duration of 7 years before diagnosis. Nine patients with duodenal webs manifested weight loss of an average of 6.4 kg. In 6 patients, episodic upper gastrointestinal (GI) tract bleeding occurred that was attributed to peptic ulcer disease. Subsequently, ulcers were noted in 12 of 20 patients with webs (Table 2).

Preoperative evaluation consisted of numerous radiologic and endoscopic studies, including barium contrast upper GI tract studies (Figure 1 and Figure 2), computed tomography, and percutaneous transhepatic cholangiography (1 patient with jaundice), which nicely demonstrated a saccular wind sock web. Upper GI endoscopy was performed in 20 patients, with the correct diagnosis being made in 9. Fourteen patients were diagnosed only at the time of operation.

The duodenal web was found to be preampullary in 20 of 22 patients and postampullary in 2. All but 1 web had a single aperture, and that patient had an imperforate wind sock web with a second eccentrically located aperture. There were 17 central apertures and 5 eccentric openings, with sizes ranging from 0.5 to 20.0 mm.

Surgical treatment consisted of transduodenal web excision with transverse duodenoplasty in 16 patients, excision of the web in conjunction with an extended pyloroplasty in 2, and resection of a previously undiagnosed web in 3 during antrectomy with gastrojejunostomy or revision of a Billroth II anastomosis to a Billroth I gastroduodenostomy for bile reflux (Table 3). Associated anomalies included pancreas divisum in 2 patients and intestinal malrotation, insulinoma, and paraesophageal hiatal hernia in 1 patient each, as well as the concomitant annular pancreas in 2 patients. Complications included delayed gastric emptying in 2 patients, a wound infection in 1, abdominal fluid collection in 1, and a duodenal leak in 1. The latter 2 were successfully treated without surgery.

Seven patients had annular pancreata (5 women and 2 men). Both of the combined anomaly patients were women. Their symptoms also included epigastric pain (n = 5), nausea and vomiting (n = 2), and weight loss (n = 4), but there were no patients with peptic ulcer disease (Table 2). Medical and surgical histories showed 6 patients underwent cholecystectomy previously, without mention of the annular pancreas.

Upper GI barium studies demonstrated duodenal stenosis in 3 patients (Figure 3). Abdominal computed tomographic scans confirmed the diagnosis in only 1 patient. Five patients with annular pancreata only were diagnosed before surgery by either upper GI endoscopy or endoscopic retrograde cholangiopancreatography (ERCP). Two patients were diagnosed only at the time of surgery. Of the 2 patients with combined webs and annular pancreata, 1 was diagnosed by upper GI barium study and 1 by percutaneous transhepatic cholangiogram.

Surgical interventions for 7 patients with annular pancreata included division of the annulus with transverse duodenoplasty in 3, duodenoplasty only in 1, duodenoduodenostomy in 1, and duodenojejunostomy in 2. The latter 3 had a complete annulus with a fibrotic narrowed descending duodenum that prohibited annulus division or duodenoplasty.

Overall results in patients with annular pancreata were excellent in 4 and fair in 3. The outcomes in the 2 patients with both anomalies were classified as good or excellent. There was 1 wound infection and 1 episode of pancreatitis in a patient who had duodenojejunostomy without division of the annulus. One patient had symptomatic delayed gastric emptying but has since recovered. There were no deaths in these patients, and no pancreatic fistulae occurred in the 5 patients having division of their annulus.

Congenital duodenal anomalies are rare in adults. In the pediatric population, the incidence is estimated to be 1 in 20 000 to 40 000 births, with incomplete obstructive lesions including duodenal webs accounting for only 2% of these defects. Ravitch³ estimated the adult incidence of duodenal stenosis from annular pancreata to be 3 in 20 000 autopsies, and Naylor and Juler⁴ described this anomaly in 2 of 20 000 laparotomies at their hospital. A review of the available published literature discovered 160 cases of annular pancreata and 76 cases of duodenal webs in adults. Most are single case reports or small series. Thus, the diagnosis has often been overlooked in the differential diagnosis of gastric outlet obstruction in the adult, and most frequently webs, at least, have been mistaken as scarring from duodenal ulcer disease.

The embryologic origin of the annular pancreas is proposed to be a defect in the normal rotation of the ventral pancreatic anlage to the right and dorsally to join the dorsal pancreas.⁵ It is thought that when this migration is imperfect, a ring of pancreatic tissue might be left around the second part of the duodenum (Figure 4). During this same process, failure of fusion of the dorsal and ventral pancreata leads to pancreas divisum. Baldwin,⁶ in 1910, reported that his studies suggested that a persistent left ventral bud was responsible for the encircling of the descending duodenum by pancreatic tissue. It was postulated that in most individuals, the left ventral anlage disappears completely and only the right ventral anlage rotates around the duodenum to join its dorsal counterpart. The annulus can completely or only partially encircle the duodenum and might be loosely applied to the serosal surface of the duodenum or, in other cases, intimately interdigitated with the muscularis mucosa of the duodenum. The pancreatic duct of the annulus has been observed to drain into either the intrapancreatic common duct or the ventral duct of Wirsung, providing further evidence that the right ventral pancreatic anlage alone is responsible for the annular pancreas.

An annular pancreas has been shown to be frequently associated with pancreas divisum, attesting to the timing of these defects during the same phase of early embryologic development. More recently, ERCP studies have demonstrated a much higher incidence of pancreas divisum in individuals with annular pancreata than in the general population.^7,8 The role of relative stenosis of the ductal orifice of the annulus is suggested to be similar to the cause of pancreatitis seen in pancreas divisum alone.⁹

The accepted cause of intraluminal duodenal anomalies such as atresias or webs arises from the seminal studies of Tandler,¹⁰ a Viennese anatomist, in 1900. He studied sections from eleven 8.5- to 20.0-mm human embryos and described a process of luminal obliteration by epithelial overgrowth. As the duodenum enlarges and lengthens, vacuoles are formed that coalesce, and eventually a lumen is reestablished¹⁰ (Figure 5). This theory has been supported by Lynn and Espinas,¹¹ Streeter,¹² Boyden et al,¹³ and others in subsequent studies. Boyden et al elegantly stratified this process and further described a stage of fusiform dilation above and below the hepatobiliary and dorsal pancreatic outpouchings. Subsequently, a wave of epithelial material clearing passes from pylorus to jejunum, leaving a patent lumen in the normal embryo by 12 weeks after gestation. Failure to clear the epithelial plug completely can result in atresias or webs anywhere in the duodenum. This process is thought to be most faulty in areas of fusion or visceral outgrowth, hence the propensity for the webs to occur in the ampullary region.

There are several varieties of webs: complete duodenal atresias or imperforate webs, intraluminal imperforate webs (wind sock webs), and perforated webs with either central or eccentric apertures. Although most webs have been observed in the preampullary or postampullary region, they have also been reported throughout the third and fourth portions of the duodenum. The juxta-ampullary position of the webs makes surgical or endoscopic extirpation a potentially dangerous procedure.

Duodenal atresias or imperforate webs always present shortly after birth; long-term survival is not possible with total duodenal obstruction, but, for the past 100 years, numerous case reports have been published describing duodenal webs and annular pancreata in the second, third, and later decades of life. The onset of symptoms in the adult seems to be the effect of a progressive decompensation of the peristaltic force of the stomach and proximal duodenum. This is attested to by marked dilation of the stomach and the first part of the duodenum as well as a patulous pylorus. Although the same mechanism might be responsible in the annular pancreas, several authors have documented acute pancreatitis in the annulus, causing high-grade duodenal obstruction, with subsequent resolution of the inflammatory process allowing relief of the obstruction.¹⁴

Surgical treatment of duodenal webs was not attempted until 1936 by Kreig,¹⁵ who performed a gastrojejunostomy, bypassing the duodenal obstruction. Vidal¹⁶ successfully treated a patient with an annular pancreas with gastrojejunostomy in the early 20th century. Many of the earliest reports are autopsy descriptions in patients with gastric outlet obstructive symptoms or incidental observations. The infrequent occurrence of these anomalies has resulted in great variability in the surgical treatments carried out. During the past several decades, however, new technology has produced the means to accurately diagnose the condition and to make accurate plans for correction and treatment.

In the treatment of duodenal webs, the most acceptable procedure currently is longitudinal duodenotomy followed by careful excision of the web—either completely or partially—mucosal reapproximation, and subsequent transverse closure of the duodenum. In the early literature, bypass of the web by posterior gastroenterostomy, duodenoduodenostomy, or duodenojejunostomy was carried out, but it has been largely abandoned. In the face of concomitant ulcer disease, ulcer operations such as vagotomy and pyloroplasty and antrectomy with gastrojejunostomy have also been performed, but these procedures are more treatment than is needed. Relief of the duodenal obstruction should allow the stomach to regain its peristaltic function, and antral decompression should diminish gastrin production, with return of gastric acidity to normal levels and healing of ulcers. When web excision is performed, careful attention should be given to protection of the biliopancreatic sphincter mechanism. By either leaving the periampullary portion of the web or intubating the bile duct and ampulla with a probe, injury to this structure can be prevented. Several reports^17,18 of endoscopic incision or Nd:YAG laser ablation of a web appear in the literature, but subsequent scar formation has resulted in stenosis and the need for subsequent surgery.

The embryologic development of the annular pancreas also affects its clinical presentation, diagnosis, and eventual treatment. The "trail" of pancreas around the descending duodenum can be complete, partial, extramural, or intermingled with the duodenal muscular wall, causing a fibrotic stricture. There is also an increasing awareness of coexistent pancreas divisum in some of these patients, and a surgical approach to the ampulla and accessory duct might be required through the annulus. Preoperative endoscopy and ERCP will be revealing in this situation and assist in planning the operation. More recently, use of magnetic resonance cholangiopancreatography has been useful in making the diagnosis without the need for a more invasive procedure. In the early literature, transection or resection of the annulus frequently resulted in pancreatic fistula formation. Early surgeons did not have the advantage of radiologic studies to demonstrate the presence and location of the annular ductal system and did not recognize or control the transected duct. If the annulus is partial and not densely adherent to the duodenal serosa, and the duodenal segment beneath the annulus is not fibrotic and strictured, then dissection of the annulus, with or without complete removal, might be considered and transverse duodenoplasty carried out. A recent report¹⁹ of successful staple management of the transected annulus presents another surgical option in this condition. Otherwise, its seems safest to bypass the annular constriction by duodenoduodenostomy or duodenojejunostomy to the proximal duodenal bulb. Early articles in the literature frequently document gastric resection with posterior gastroenterostomy, but again this seems to be too much surgery and might predispose to postgastrectomy problems such as stomal ulceration, stricturing, dumping syndrome, bile reflux gastritis, and anemias.

During the past 100 years, to our knowledge, 160 adults with annular pancreata and 76 with duodenal webs have been described. In the 76 patients with duodenal webs, there is no sex difference, and the mean ± SD age at diagnosis is 51.5 ± 17.1 years (range, 17-81 years). Patients with annular pancreata had a slightly greater male incidence (ratio, 1.78:1). The mean age of patients with annular pancreata is 44.4 ± 15.1 years (range, 17-79 years). There seems to be a difference in duration of symptoms before presentation between the 2 groups, with patients with webs averaging 137 ± 208 months compared with 68 ± 84 months for those with annular pancreata. Most patients in both groups presented with signs of gastric outlet obstruction, nausea and vomiting, upper abdominal discomfort, and early satiety. Patients in both groups were noted to have gastric or duodenal ulcers, and a few patients with annular pancreata presented with biliary obstruction and jaundice.

Gastric and proximal duodenal dilation are common to both groups on upper GI contrast studies, but in patients with webs, a transverse diaphragm in the descending duodenum with an eccentric or central aperture might be seen. Hypotonic duodenography is more likely to draw the radiologist's attention to such a finding than a usual upper GI barium study. The annular pancreas, since it is an extrinsic lesion, will be seen as a smooth or tapered narrowing of the second part of the duodenum. Endoscopic examination will reveal only a circumferential narrowing, but an ERCP might delineate the ductal system of the annulus and its junction with the ventral pancreatic and biliary ducts. Computed tomography and magnetic resonance imaging have recently been shown to benefit the diagnosis and overall management of these patients.

The presentation, location,and number of webs is variable, with a preponderance of preampullary positioning in 70% of patients described. Of the remaining webs, 25% were postampullary and 5% intra-ampullary, ie, at the ampulla itself. Virtually all of these webs had a single aperture, either central or eccentric. Eighty-six percent of patients had single webs, and 14% had multiple webs occurring in preampullary and postampullary positions. Several authors suggest passing a Foley catheter distally in the duodenum to ensure that one or more distal webs are not overlooked. The evolution of therapy for duodenal webs has been toward transduodenal web excision and duodenoplasty (Table 4). Fifteen percent of reported patients had proximal duodenal diversion, and 9% had partial gastrectomy with gastroenterostomy. The resections were done in the earlier years, and this might be because of lack of the ability to make the diagnosis before surgery. In addition, many of these patients had signs and symptoms of chronic peptic ulcer disease and, therefore, the operation was intended to treat the ulcer diathesis. Complications occurred in 16% of these patients, and most deaths were in patients without surgical treatment. Complications of web excision included pancreatitis, duodenal stenosis, and leaks. Symptomatic relief was observed in 61 patients (77%), and recurrent symptoms occurred in 4%, with stomal ulceration in 1 patient after gastroenterostomy without vagotomy.

In patients with annular pancreata, surgical therapy has been more varied, without arriving at a single best solution as seen with duodenal web excision. Early attempts at resection or division of the annulus resulted in more complications of duodenal leak, pancreatitis, and pancreatic fistula. Bypass of the annulus by duodenoduodenostomy, duodenojejunostomy, and antrectomy with gastrojejunostomy have evolved as the seemingly preferred methods of therapy (Table 5). In patients who had no surgical treatment, mortality was 26.3%. After gastric resection and gastroenterostomy, patients still developed pancreatitis, pancreatic and enteric fistulae, stomal ulceration, and death. On the other hand, the outcomes were considered satisfactory in 73% of all surgical patients.

In summary, awareness of the problem and preoperative diagnosis are extremely important. Not all patients need surgery, but those with significant gastric outlet obstruction will be best served by surgery. Newer techniques such as computed tomographic scanning, magnetic resonance cholangiopancreatography, and endoscopic evaluation by experienced GI endoscopists will make a difference in preoperative planning. Patients with associated anomalies such as malrotation, pancreas divisum, and combined anomalies will present challenges to the surgeon, but knowledge of the anatomy and awareness of potential complications should lead to more satisfactory outcomes.

As a result of this extensive review of a single institutional experience and an exhaustive literature search, the following recommendations are made:

• In patients with duodenal webs, a transduodenal web excision should be carried out through a longitudinal incision that is then closed transversely in Heineke-Mikulicz fashion. The ampullary structures should be protected by either avoiding resection of the juxta-ampullary portion of the web or intubating the common duct to protect the ducts. The duodenum should be inspected for distal webs by passing a Foley catheter distally and withdrawing it with the balloon inflated. Duodenal bypass is usually not indicated, and neither is major gastric or duodenal resection.

• In patients with annular pancreata, one must determine by inspection if the annulus is complete, partial, intramural, or extramural. If the annulus is partial and the duodenum is not densely strictured, incision or partial excision of the annulus and a Heineke-Mikulicz–type duodenoplasty can be carried out, especially if access to the biliopancreatic sphincter is deemed surgically necessary. Careful search for and control of the transected pancreatic ducts of the annulus is mandatory in that procedure. Otherwise, duodenoduodenostomy and duodenojejunostomy are the safest procedures to bypass the obstructing annulus and prevent long-term postgastrectomy problems, anastomotic leaks, and pancreatic fistulae. Care must be taken, however, to use an incision high enough on the proximal dilated duodenum to achieve a widely patent anastomosis that will ensure adequate duodenal drainage.

A useful clinical observation at the time of surgery for presumed duodenal stenosis from ulcer scarring is that external transverse duodenal dimpling or obstruction to the finger in the descending duodenum is probably not from an ulcer but might be an intraluminal web. An annular pancreas, on the other hand, should be obvious.

Presented at the 108th Scientific Session of the Western Surgical Association, Dana Point, Calif, November 15, 2000.

Corresponding author and reprints: James A. Madura, MD, Department of Surgery, Indiana University Medical Center, 545 Barnhill Dr, EM 244, Indianapolis, IN 46202-5125 (e-mail: jmadura@iupui.edu).

Jack Pickleman, MD, Maywood, Ill: I commend the Western Surgical Association in seeking out experts in their field to discuss each of these excellent papers. It is important to define an expert. For the purposes of the present discussion, it will be as follows. The expert should have heard of the condition, the expert should have treated 3 patients with the condition, and the expert should have a beating heart. As such, I stand before you as an expert.

The GI surgery group in Indiana brings to our attention the largest reported series of congenital duodenal abnormalities in adults, all presenting with symptoms of chronic gastric outlet obstruction. My first comments have to do with the diagnosis of these conditions. With the hairpin triggers possessed by most gastroenterologists for inserting an endoscope into both northern and southern orifices at the onset of any GI symptom, it is not surprising that 20 of 29 of these patients underwent this procedure. However, only a few had a correct endoscopic diagnosis. The authors state that the upper GI studies were done and 3 were diagnostic.

My first question is how many upper GI studies were done in total, as in my experience this is the most accurate means of securing the diagnosis. I continue to believe that in any upper GI dysmotility syndrome, both endoscopy and contrast studies are needed as they provide different information. And I suspect that the 11 endoscopy patients who were not diagnosed by these means would have been identified by an upper GI. Regarding the treatment of webs, the authors' wisely urge caution as many of these are juxta-ampullary, and if you merely evert and oversew the wind sock, you can compromise the ampulla. Their advice to only perform a lateral web excision or to protect the ampulla by selectively cannulating it from above or below should be considered by anybody treating this condition. My only experience with endoscopic excision of a duodenal web came in the form of a patient transferred to me with a near fatal pancreatitis secondary to this procedure, and so I am not particularly sanguine about this form of treatment as a routine approach.

My last comment has to do with the treatment of the annular pancreas. Although the authors appear to advocate excision of the annulus with a transverse duodenoplasty, I would prefer a duodenoduodenostomy after complete kocherization of the duodenum. Both procedures necessitate a duodenal suture line, but my approach eliminates the possibility of postoperative pancreatitis or a pancreatic fistula. Perhaps the authors could comment on my bias.

In closing, these conditions are rare, and each of you will therefore see few of them in your careers. However, the potential for diagnostic and therapeutic mishaps are real. I feel this paper belongs in the permanent reprint file of all practicing general surgeons.

Thomas A. Stellato, MD, Cleveland, Ohio: The authors suggest that excision can be performed for a loose annulus. Yet, the only fair to poor result in the study seemed to be in those patients who had an excision of the annular pancreas. All of the other results seem to be excellent, and I wonder once again whether the authors could reevaluate that recommendation.

Theodore X. O'Connell, MD, Los Angeles, Calif: I also have a question about the annular pancreas because maybe in my simple-minded approach they seem to be taken care of fairly easily with a duodenoduodenostomy or a duodenojejunostomy where you would predict a good outcome. Yet, 40% of their 5 patients, 2 of 5 patients, had a fair to poor outcome. I wish you would explain that a little bit more. Exactly what was the poor outcome due to and was it due to the operative intervention?

Philipe E. Donahue, MD, Chicago, Ill: I, too, congratulate the authors on an important study showing the importance of complete evaluation in foregut symptoms, especially the upper GI series. My one question: "What was the time interval from first symptoms until final diagnosis?" These patients often creep along the diagnostic algorithm of "diagnosis of symptom," "proton pump inhibitor," "watchful waiting," "endoscopy," "repeat endoscopy," . . . a process which can consume years. It is very expensive and often wasteful of time and resources. Tell us something about that if you can.

Dr Madura: Dr Ladd and I would again like to express our thanks to the Program Committee and to the Western Surgical Association for allowing us to present this unusual report. Not only did we have a primary discussant with a beating heart, but it was a gentler, kinder beating heart, and we appreciate Dr Pickleman's remarks.

Of these conditions, the first web I ever saw, I missed at surgery as well. We had a patient with recurrent duodenal ulcer disease who did not get better, and we were actually doing a Jaboulay pyloroplasty when we discovered the web. I remembered back to my training with Dr Zollinger, who, in the heyday of gastric surgery, always told me that when doing a pyloroplasty or gastric resection you should always put your finger down into the duodenum because they could have a second ulcer stricture, which I never understood very well and which I never saw until my first patient with the web. These ulcer strictures that he described were probably not due to ulcers but probably were webs. There are undoubtedly a lot more individuals out there with webs than we would suspect.

Now, why were they not preoperatively diagnosed? Well, with a large, big, dilated pylorus and duodenal bulb, the endoscopist many times mistakes that web aperture for the pylorus. They really mistake the duodenal bulb dilation as antrum, and unless they biopsy it (which they don't), they will miss the diagnosis completely. So they reported a strictured pylorus in many of our cases. Upper GI series is similar, and many of these were not initially diagnosed by an upper GI series but, in retrospect, you can clearly see the web like you did in one of our photos in a patient who came from another hospital. We, too, agree with you, Dr Pickleman, that endoscopic excision should not be undertaken lightly. There are 3 reports in the literature, 2 laser ablations and 1 simple incision, with a papillotome through the endoscope, and although all of the endoscopists were cautious about how they did it, they got duodenal strictures, and several of them had to be reoperated on. So there is not a huge experience in the endoscopy literature for ablating these endoscopically. Now, the patients in whom we excised or divided an annulus were not patients who had the tight fibrotic strictures of the duodenum. These were the patients who either had an incomplete annulus or in whom the annulus was loosely applied to the duodenal serosa, and we could dissect it off and carefully control the ductal system. Why didn't we just do a bypass? Well, these patients also had concomitant stenosis of the sphincter of Oddi, which had been diagnosed by pancreatic duct manometry and, because of the annulus, it was difficult for our ERCP service to get there and divide the pancreatic sphincter. So they referred them to us. These were combined procedures in these patients with excision of the web, duodenoplasty, and transduodenal sphincteroplasty. The reason they did less well in some of these cases was they had continued problems with their pancreas and some had delayed gastric emptying, and, therefore, we concluded that those patients were not excellent outcomes. That answers Dr Stellato's question as well as Dr O'Connell's question. The fair to poor outcome was due to pancreatitis and delayed gastric emptying.

One other point that we want to make is that these patients, a third of them, will present with ulcer disease and they don't need an ulcer operation. Once you eliminate the web and open up the duodenum, their gastric dilation and antral distention goes away. Their resting gastrin levels decrease, and their ulcers can be easily managed with either no medication or short-term antacid medications.

The timing interval, Dr Donahue, from onset of symptoms to diagnosis was up to 10 to 15 years. One of our patients who was in her 70s said that she had had problems with digestion ever since she was a child but would eat more frequently and would not eat solid and high-residue foodstuffs but ate more of a liquid diet.