An Analysis of Operations for Gastroesophageal Reflux Disease: Title and subTitle BreakIdentifying the Important Technical Elements FREE

A BETTER understanding of the pathogenesis of gastroesophageal reflux disease (GERD) and the development of minimally invasive techniques during the past decade have not been accompanied by a consensus on what constitutes an optimal antireflux operation. Operations are still being performed and referred to just as they were described more than 30 years ago, and discussions about technique too often seem to pivot on which one is best. The problem with this thinking is that the Nissen, Toupet, Lind, Guarner, and Hill procedures^{1 - 6} sprung from the intuition of these surgeons during a time when much less was known about the pathogenesis of GERD, and the testing and follow-up to validate their efficacy was scant. It is unlikely that any of these procedures as originally described constitutes the ideal operation for GERD. Each, however, contains important technical elements that are shared to a greater or lesser extent by the entire group. The goal of this study was to identify the elements of genuine importance.

Between October 1, 1992, and December 31, 1996, 201 patients underwent laparoscopic fundoplications for GERD. The experience can be divided into 2 periods: October 1, 1992, to June 30, 1993, when all 22 patients (group A) had 360° (Nissen-Rossetti) fundoplications without division of the short gastric vessels; and July 1, 1993, to December 31, 1996, when 82 patients (group B) had a total (360°) fundoplication and 97 patients (group C) had a partial (240°) fundoplication, the short gastric vessels being divided in each of these latter 2 groups. During the second period, a partial fundoplication was chosen for patients with abnormal esophageal peristalsis (amplitude in the distal esophagus, <40 mm Hg; >20% segmented waves; triple peaked or dropped waves or both).

The main clinical findings are given in Table 1. The preoperative workup for each patient included a barium swallow study, esophagogastroduodenoscopy, esophageal manometry, and 24-hour pH monitoring (Table 2). The rate of gastric emptying was measured in patients with symptoms suggestive of delayed emptying.

The patient was placed on the operating table in a steep reversed-Trendelenburg position, with the legs held in stirrups. The surgeon stood between the patient's legs. Five 10-mm trocars were used, as shown in Figure 1. The procedure was started by dividing the gastrohepatic ligament and separating the esophagus from the right border of the crus by blunt dissection. A Babcock clamp, placed on the stomach near the gastroesophageal junction, was used for traction. The peritoneum overlying the esophagus was divided, and the esophagus was separated from the left border of the crus by sharp and blunt dissection. With continued dissection, the esophagus and gastroesophageal junction on the one hand and the diaphragmatic crus on the other were completely separated from each other circumferentially, and both crural pillars were cleared of connective tissue down to the point where they decussated. The end result of this dissection was the creation of a large window behind the abdominal esophagus, anterior to the crural pillars.

In groups B and C, the short gastric vessels were divided starting at a point midway along the greater curvature of the stomach and continuing upward to the angle of His. Any posterior attachments of the fundus, including occasionally a posterior gastric artery, were also divided.

The reconstruction was begun by narrowing the esophageal hiatus by using interrupted sutures of 2-0 silk placed behind the esophagus and tied intracorporeally. A bougie was not in the esophagus during this step because it would interfere with exposure and suturing of the crural pillars. Instead, the hiatal reduction was gauged to preserve about a 1-cm gap between the uppermost suture and the posterior aspect of the esophagus, and the result was calibrated for tightness by the bougie used in the next step of the operation.

The fundus of the stomach was then pulled behind the esophagus, and the wrap was created over a 56F to 60F bougie. In the patients with a total wrap (groups A and B), 3 sutures of silk were used to approximate the 2 sides of the wrap. Two of these 3 sutures included generous bites of the esophageal musculature, avoiding the anterior vagus nerve.

In group A, because the short gastric vessels were not divided, the anterior wall of the gastric fundus was used for the wrap. In groups B and C, where the short gastric vessels were divided, the posterior and the anterior walls of the fundus were available to form the wrap. In group C, a 240° wrap was made by placing 3 sutures between each of the fundic flaps and the esophagus in the 10-o'clock and 2-o'clock positions in relation to the esophageal circumference.

Beginning in mid-1994 (ie, in groups B and C), 2 coronal sutures (each incorporating the esophagus, the apex of the wrap, and the adjacent right or left crural pillar; Figure 2, A, and Figure 3, A) and 2 posterior sutures (between the posterior side of the wrap and the closed crus; Figure 2, B, and Figure 3, B) were used.

Additional procedures were performed in 4 patients in group A: 1 patient had excision of a liver cyst, 1 patient had a pyloromyotomy, and 2 patients had repair of umbilical hernias. In group B, 1 patient had a selective vagotomy, 2 patients had resection of a Zenker diverticulum and cricopharyngeal myotomies, and 1 patient had repair of an umbilical hernia. In group C, 2 patients had pyloromyotomies and 5 had laparoscopic cholecystectomies.

The Student t test and the Mann-Whitney U test were used for the statistical evaluation of data. All results are expressed as mean (±SD). Differences were considered significant at P <.05.

The operating time for groups A, B, and C was 193 (±45), 150 (±50), and 161 (±47) minutes, respectively. The hospital stay in groups A, B, and C was 52 (±21), 40 (±27), and 36 (±29) hours, respectively. The average (range) follow-up was 43 (33-52) months for group A; 23 (1-48) months for group B; and 15 (1-44) months for group C.

The heartburn score went from 3.7 (±0.5) preoperatively to 0.3 (±0.6) postoperatively (P<.05). Heartburn resolved completely in 15 patients (68%) (Table 3). In 3 patients (14%) who did not have dysphagia preoperatively, dysphagia developed postoperatively (de novo dysphagia, defined as difficulty in swallowing solids or liquids that is still present 3 months after the operation). Dysphagia resolved in 2 patients and improved in 1 patient after esophageal dilatation.

Two patients had left-sided pneumothoraces intraoperatively that required chest tube drainage.

The heartburn score went from 3.3 (±0.7) preoperatively to 0.2 (±0.6) postoperatively (P<.05). Heartburn resolved in 73 patients (89%) (Table 3). De novo dysphagia developed in 5 patients (6%), which resolved spontaneously in 2 patients and after 1 esophageal dilatation in another patient. It improved in the remaining 2 patients (1 patient had 1 dilatation).

The fundus was perforated in 1 patient, and the perforation was sutured laparoscopically without incident.

The heartburn score went from 3.4 (±0.7) preoperatively to 0.2 (±0.7) postoperatively (P<.05). Heartburn resolved completely in 88 patients (91%) (Table 3). In 2 patients (2%), de novo dysphagia developed but resolved spontaneously in 1 patient and after 2 pneumatic dilatations in the other.

One patient had an instrumental gastric perforation that was repaired laparoscopically. One patient had a pulmonary embolus.

Nine patients (4.5%) had structural failure of the wrap (Table 4) due to herniation into the chest in 7 patients (2 patients in group A, 1 patient in group B, and 4 patients in group C), disruption in 1 patient (group B), and the development of a paraesophageal hernia in 1 patient (group C).

The 7 patients whose wraps herniated into the chest had recurrent heartburn, and the problem was diagnosed by results of the barium swallow x-ray studies. Esophageal manometry demonstrated a hypotensive lower esophageal sphincter (average [range] sphincter pressure, 6.5 [2.0-10.0] mm Hg), and ambulatory pH monitoring confirmed the presence of abnormal gastroesophageal reflux (average [range] DeMeester score,⁷ 90 [20-213]; normal, <15). The videotapes of these operations were analyzed for the extent to which the esophageal hiatus had been reduced in size and the presence or absence of coronal and posterior anchoring sutures.

The esophageal hiatus was closed in all 7 of these patients, but it was left too wide in 1 patient (>2 cm between the uppermost suture and the esophagus). In 3 patients, neither coronal nor posterior sutures had been used. In 3 other patients, posterior sutures had been omitted. In 1 patient, all of these stabilizing elements had been used, and no reason could be discerned for the failure.

Of these 7 patients, 4 have had second laparoscopic operations to correct the abnormality, and they are now asymptomatic. One patient is scheduled for a repair in the near future. Two patients' symptoms are well controlled with omeprazole therapy, and they have chosen not to undergo another operation.

One patient had severe dysphagia within 2 days of the operation, and barium x-ray films showed a paraesophageal hernia that required an open repair. The videotape of the first operation showed that coronal sutures had not been used.

A gastric perforation developed acutely in 1 patient 7 months after a total fundoplication. At reoperation, a 4-mm perforation was found on the anterior wall of the gastric body, and the wrap had completely disrupted. The perforation was closed, and a Dor fundoplication was created. Analysis of the videotape of the first operation showed that posterior sutures had not been used, but the cause of this peculiar complication is obscure.

This study is a retrospective analysis. Even though the findings may be influenced by a learning curve, the results show that laparoscopic antireflux operations control reflux reliably and with few adverse effects, and they suggest that the following technical elements are important:

These and other technical steps of importance are listed in Table 5.

Stein et al⁸ showed that recurrent reflux after antireflux operations is largely attributable to structural failure of the reconstruction. Our findings are in accord with that observation. They identified disruption of the wrap as the principal mode of failure, with herniation of the wrap being the second major category. Herniation occurred in 7 (3.5%) of our patients and wrap disruption in 1.

The reason why a wrap occasionally herniates into the chest is not entirely clear. Our 7 cases compare favorably with those of other authors who report a 5% to 10% incidence of this complication.^{8 - 9} We could see no evidence that short esophagus was a contributing factor because the hernias in our patients who had this complication were not larger than average; they ranged from no hernia to a maximum of a 3-cm hernia. Thus, the criteria for a short esophagus were absent in these 7 patients, and none would have been considered to be a candidate for an esophageal-lengthening procedure (eg, Collis-Nissen operation). Herniation was also no more common in patients with esophageal dysmotility than in those with normal esophageal motility.

In an attempt to prevent herniations, in mid-1994 we began using anchoring stitches between the wrap and the diaphragm. The posterior anchoring stitches between the wrap and the approximated crural pillars (Figure 2, B, and Figure 3, B), analogous to the posterior sutures used in a Toupet or Hill fundoplication, counteract cephalad traction forces that can displace the wrap into the posterior mediastinum. The coronal sutures (Figure 2, A, and Figure 3, A) also counteract cephalad traction forces, and they help to prevent the wrap from unfolding (ie, deintussusception), a type of disruption often referred to as a "slipped Nissen."

The subsequent 3 occurrences of wrap herniation were confined to patients in whom this array of anchoring stitches was incomplete. Thus, this experience supports the contention that steps should be taken during these operations to secure the wrap in the abdomen. Whether our current methods are optimal will be evident with longer follow-up.

The terms Nissen, Nissen-Rossetti, Toupet, Lind, Guarner, Hill, and Dor are used to denote various antireflux operations (Table 6). Unfortunately, exactly what each stands for in the technical sense is not agreed on, and in several instances an eponym has come to be used today for an operation that is substantially different from what the author described. For example, the operation described by Toupet⁶ consisted of a partial wrap with posterior sutures between the wrap and the diaphragm. A unique aspect of this fundoplication was Toupet's belief that the hiatus should not be narrowed, or in the few cases where this was allowed, the stitches were to be placed anterior rather than posterior to the esophagus. In recent literature, however, the term "Toupet procedure" has been used loosely to refer to any partial fundoplication even when it includes closure of the hiatus. Furthermore, whether the wrap was fixed posteriorly or the hiatus narrowed has often not been mentioned in reports about this and other types of fundoplication. Consequently, it may be impossible to know exactly what was done, for the term Toupet procedure has been used for a variety of reconstructions that differ in important details.

Toupet⁶ and Hill² considered posterior gastropexy to be central elements of their repairs. The Hill repair has evolved so that it now includes a fundoplication. The Nissen⁴ and Guarner¹ fundoplications, which originally did not include a posterior fixation, have come to include this step in many cases. Consequently, the drift in surgical practice is toward a common set of technical principles, including fundoplication, hiatal reduction, and gastropexy.

The lack of agreement concerning the definitions of the eponymous fundoplications has reached the point where different authors might refer to the same reconstruction as a Toupet fundoplication, a Guarner fundoplication, or a Lind fundoplication. Trying to understand from the literature what makes for a successful operation is bound to be confused by this practice. In Table 5 is a list of technical steps that seem likely to contribute to the efficacy of fundoplications. We think that reports on this subject should include data concerning each of these steps.

The esophageal hiatus was reduced to a more normal size in all patients. In general, the incidence of herniation of the wrap has decreased as closure of the hiatus has become a routine part of fundoplications.^{10 - 11} Furthermore, hiatal closure probably helps to strengthen the lower esophageal sphincter.^{12 - 13} Mittal et al^{12 - 13} have shown that the crus functions as an extrinsic sphincter, which works synergistically with the intrinsic sphincter (ie, the lower esophageal sphincter) to produce competence at the gastroesophageal junction. The extrinsic sphincter acts during inspiration to prevent reflux associated with rises in the intraabdominal pressure.^{12 - 13} Thus, there are compelling reasons to resize the hiatus.

In the first 22 operations (group A), we did not divide the short gastric vessels, and 3 patients (14%) had postoperative dysphagia requiring esophageal dilatation. After we began dividing the short gastric vessels, the incidence of dysphagia decreased to 4%, an experience similar to that of others.^{14 - 16} For example, Hunter et al¹⁵ noted an 11% incidence of postoperative dysphagia when the short gastric vessels were not divided, which dropped to 2% when they were divided. Although these findings could be influenced by a learning curve bias, dividing the short gastric vessels does facilitate creating a floppy tension-free wrap. If the short gastric vessels are left intact, the anterior wall of the fundus must be used for the fundoplication, and the short gastric vessels remain visibly under tension. Consequently, dysphagia is often more common after this procedure (ie, the Nissen-Rossetti technique) because of 1 or more of the following: the wrap is too tight; the wrap is being pulled to the left because the fundus, wrapped around the midline esophagus, is still tethered to the spleen; and asymmetry between the lateral forces on the 2 sides of the wrap produces a rotational effect on the gastroesophageal junction, which creates a spiral deformity.

About 20% of patients with GERD have severely abnormal esophageal motility, and the extent of the abnormality is proportional to the degree of esophageal mucosal injury.^{16 - 17} Neither the control of reflux nor reduction of the hiatal hernia corrects the peristaltic dysfunction.¹⁴ Some physicians think that patients with abnormal peristalsis are poor candidates for surgical treatment, fearing that they will inevitably have a high incidence of postoperative dysphagia and gas bloat syndrome.¹⁸ These are the patients with the most severe symptoms, however, and in the most need of treatment. They have a high incidence of esophageal stricture formation and Barrett esophagus.^{19 - 20}

Performing a total fundoplication in every patient may control reflux, but it runs the risk of producing postoperative dysphagia.^{9 ,21} We tailored the fundoplication (240° vs 360°) to the results of the esophageal function tests, choosing a partial fundoplication for patients with abnormal peristalsis.^{10 ,14 ,22 - 23} Following this approach, reflux was controlled in 88 patients (91%) with abnormal peristalsis (group C), and the rate of postoperative dysphagia was low. To obtain such results, however, all wraps (total or partial) should be formed over a 56F to 60F bougie.

The goal of antireflux surgical repair is to control reflux without creating troublesome adverse effects. This can be achieved if a thorough preoperative evaluation is followed by an operation that includes the elements listed above. Analysis of the failures of antireflux surgical repair has shown that the most common reasons for a poor outcome are either an incorrect diagnosis or an unstable reconstruction.⁸ It is critical to take all appropriate steps to create an effective and stable fundoplication.²⁴

Alex G. Little, MD, Las Vegas, Nev: The authors have made a significant contribution to developing what you might call the ideal operation for gastroesophageal reflux or, maybe a better way of saying it, to identifying the key or essential elements in these operations. It is interesting that after half a century or so of doing these types of operations, there still is not anything close to a consensus on how they should be done. This is an important paper in that context.

I agree with the authors that technique, whether it's open or laparoscopic, is essential to a good outcome, but don't think that we should skip emphasizing the fact that patient selection is crucial. I would like to have the authors comment on their criteria for selection. This is particularly important as the curve seems to be shifting, at least in my practice, toward younger patients attracted by a less morbid operation. Selection is at least as important, if not more important, than it has been in the past. One specific question would be, did the authors include any patients with Barrett esophagus who did not have significant GERD symptoms?

In terms of technique, let's first focus on the question of the migration, if that is the right word, of the wrap up through the hiatus. Looking at my own experience and looking at patients with 5-year follow-up—obviously these were all patients done with an open procedure—but nonetheless in the same operation, where the hiatus is closed, the esophagus and fundus are both generously mobilized, and a loose wrap is performed, I couldn't find more than a 2% to 3% incidence of hiatal migration. I am, therefore, not really convinced that the anchoring stitch is necessary. On the other hand, I don't see any harm in doing it, but I am not convinced it is essential. But even closing the hiatus, which would seem so obviously a necessary maneuver to many of us, is still far from universally applied. I do definitely agree with the authors that the hiatus must be closed.

In terms of postoperative dysphagia, the experience has been that when the operation is done, as you did, with division of the short gastric vessels and a loose or floppy type of fundoplication, the incidence of dysphagia is low. Therefore, I am not really convinced, since you didn't do the Nissen procedure in the patients with those that you define as having poor motility, that an incomplete or partial wrap is essential. On the other hand, to ask it another way—and perhaps reminding you that you have a short follow-up in group C, and I don't think we should leap to too many conclusions when there are patients with only a 1-month follow-up—nonetheless, assuming your results hold up in the partial or incomplete fundoplication group, why not simply do that operation in all patients? Why would any patient need a total fundoplication?

Philip E. Donahue, MD, Chicago, Ill: Dr Patti and Dr Way have led the way in trying to analyze the elements of effective operations. Their incidence of dysphagia is absolutely the lowest that you will see published for these operations in the world today. While they deserve recognition for that, the incidence of heartburn is one of the highest that you will ever see after these operations; I am not sure why the incidence of heartburn is as high as it is, unless they have been so rigorous in eliciting even mild degrees of discomfort that they have included patients who are not included in other reports.

My questions are 2: Did you have any perforations of the esophagus with the intraluminal bougie? That is 1 part of the operation that can really be eliminated. If you are going to be having an adequate dissection, a large retroesophageal window, a nice floppy—that's "f l o p p y"—wrap that fits behind the esophagus, then you really don't need an intraluminal tube. On the other hand, if you put that tube into the esophagus and then stretch the fundus around so it stays there like a taut rubber band, you won't achieve the end result that you are looking for. So that's a question. Did you have any perforations? Can't those tubes with their inherent risk of perforation be eliminated?

Finally, there is increased interest in doing a less-than-complete wrap in all patients, especially for young women who might have a subsequent pregnancy. There is nothing that destroys a fundoplication like repeated episodes of emesis, whether it's gravidarum or otherwise. I have found partial wraps effective for "all comers" and think that 10 years from now, we'll all be doing partial wraps. Have you considered that approach?

Jeffrey H. Peters, MD, Los Angeles, Calif: I wonder if you would focus on 2 concepts for me. The first is the issue of partial wraps and when you use them. I, in contrast to the last 2 discussants, wonder if we aren't using them too much and think that, particularly in the young population with Barrett esophagus, the issue of recurrent reflux through a partial wrap, which some have suggested may be as high as even 50% in 3, 4, or 5 years, will be its Achilles heel. We mustn't forget that we need to provide a long-term cure in these patients. What were your indications for a partial wrap? You certainly did use a large number of them, and you also seem to have a high incidence of recurrent heartburn.

The second comment is the importance of the geometry of the fundoplication. My own bias is that it makes a big difference and may be responsible for some of the dysphagia that we see. In the group of Rossetti's that you did, why do you think they get dysphagia? When we see some of these patients back in our clinic now following fundoplication with recurrent dysphagia, what should we do with them?

Raymond J. Joehl, MD, Chicago: Have you had an opportunity to assess manometry postoperatively in some of these patients, not only in those who have failed or who have recurrent symptoms but also in those patients who have had a successful procedure, particularly to dissect the contributions to gastroesophageal competence or the new high-pressure zone, dissecting the component due to the fundoplication and that due to the tightened crus?

For the patients who had recurrent symptoms and some degree of failure, how many were you able to reoperate on laparoscopically and successfully redo the procedure?

Stephen E. Reid, MD, Chicago: Would the authors tell us what the specific criteria are that would make them do a partial wrap? Does everyone get manometry, and what are the exact criteria used to make that decision? There were a lot of partial wraps, and it is not clear to me why those patients had them.

Edward H. Phillips, MD, Los Angeles: Did you have the opportunity to randomly pick 20 patients who underwent 360° and/or 240° wraps and were asymptomatic postoperatively without dysphagia or recurrent reflux and compare the incidence of the failure to perform posterior, crural, or hiatal closure sutures?

Jorge L. Rodriguez, MD, Minneapolis, Minn: This operation is being undertaken in young people to avoid long-term morbidity with medication. Since you have such a high incidence of reflux, how many of these people were treated with medication, and what was their final outcome? Did medication resolve the issue?

Dr Way: First, I wish to stress that our principal point is that the details of operations to combat reflux should be given more attention because there is increasing evidence that the details are of genuine importance. We do not expect that this article necessarily contains definitive answers, but we feel that the more precise recording of wrap construction, the use of anchoring stitches, and other technical steps is called for to increase further the effectiveness of an already excellent operation.

Dr Little asked for our indications for an antireflux operation. The patient should have severe symptoms requiring continual treatment. The diagnosis of GERD should be well established as the cause of the symptoms, which with few exceptions should include pH monitoring and motility testing. If heartburn is the major symptom, one would expect a good response to proton pump inhibitors; symptoms that do not respond to these drugs may not result from acid reflux. In other words, genuine reflux is now rarely intractable to drug therapy. The more atypical the symptoms (eg, chest pain, nausea, diarrhea, abdominal pain), the less likely they will respond to an operation that corrects reflux.

Hubert Stein reported that the causes of failure of antireflux operations on late follow-up could be divided into 2 categories: patients who were poorly selected for the operation (eg, reflux was not the cause of their symptoms), and those whose fundoplications were faulty or unstable. We must do our best to avoid each. The second of these categories is the subject of this article.

Dr Little was skeptical about the importance of anchoring stitches. Time will tell, especially if we record their use systematically and take them into account when analyzing results. Our data suggest that the skeptics may have to change their opinions.

Why not do a partial fundoplication in all patients? And what are our indications for a partial wrap? We have used partial fundoplications for patients whose esophageal peristaltic pressure does not exceed 50 mm Hg or if the patient has dysphagia preoperatively in the absence of mechanical obstruction (eg, stricture formation). We believe partial wraps have not been as thoroughly validated as total wraps have, so we are reluctant to expand their use at this time. Nevertheless, the total and partial wraps have given equivalent short-term results so far, with heartburn relieved in more than 90% of patients in each group.

Dr Joehl, about 65% of patients who have had 1 unsuccessful laparoscopic fundoplication can be treated successfully by a highly experienced laparoscopic surgeon with a second laparoscopic operation. If the initial unsuccessful operation was done as an open procedure, it is unlikely that the problem can be corrected laparoscopically at the second operation.

Dr Phillips, yours is an excellent question. We have not done what you suggested.

Presented at the 105th Scientific Session of the Western Surgical Association, Colorado Springs, Colo, November 18, 1997.

Corresponding author: Marco G. Patti, MD, Department of Surgery, University of California, San Francisco, School of Medicine, 533 Parnassus Ave (Room U-122), San Francisco, CA 94143-0788.