Division of Short Gastric Vessels at Laparoscopic Nissen Fundoplication

Abstract

Objective

To determine whether division of the short gastric vessels at laparoscopic fundoplication confers long-term clinical benefit to patients.

Summary Background Data

Dividing the short gastric vessels during surgery for gastroesophageal reflux is controversial. This prospective randomized study was designed to determine whether there is a benefit in terms of patient outcome at a minimum of 5 years after primary surgery.

Methods

Between May 1994 and October 1995, 102 patients undergoing a laparoscopic Nissen fundoplication were randomized to have their short gastric vessels either divided or left intact. By September 2000, 99 (50 no division, 49 division) patients were available for follow-up, and they all underwent a detailed telephone interview by an independent and masked investigator.

Results

There were no significant differences between the groups at 5 years of follow-up in terms of the incidence of epigastric pain, regurgitation, odynophagia, early satiety, inability to belch, anorexia, nausea, vomiting, nocturnal coughing, and nocturnal wheezing. There was also no difference between the groups in the incidence of heartburn when determined by either yes/no questioning or a 0-to-10 visual analog scale. There was no difference between the groups in terms of the incidence and severity of dysphagia determined by yes/no questioning, 0-to-10 visual analog scales, or a composite dysphagia score. There was a significantly increased incidence of flatus production and epigastric bloating and a decreased incidence of ability to relieve bloating in patients who underwent division of the short gastric vessels.

Conclusions

Division of the short gastric vessels during laparoscopic Nissen fundoplication does not improve any measured clinical outcome at 5 years of follow-up and is associated with an increased incidence of “wind-related” problems.

Although the first laparoscopic antireflux procedure was performed almost a decade ago, ¹ division of the short gastric vessels during laparoscopic Nissen fundoplication remains a controversial issue. The rationale behind dividing these vessels, and thereby fully mobilizing the gastric fundus, is to facilitate the fashioning of a loose “floppy” total fundoplication, a procedure said to be associated with a lower likelihood of a tight wrap and associated postoperative dysphagia and gas bloat. Proponents of division of the short gastric vessels generally report good results when they compare current patient cohorts in whom they undertook division of the short gastric vessels with their earlier experience (or that of other surgeons) when they did not divide the vessels. However, conclusions drawn from this analysis can be subject to the influence and bias of a learning curve. ² On the other hand, other surgeons have reported good outcomes for patients undergoing total fundoplication in whom the short gastric vessels were not divided. ³

Because of this controversy, in May 1994 we commenced a prospective randomized trial of division versus no division of the short gastric vessels in patients undergoing a laparoscopic Nissen fundoplication procedure. The early outcomes from this trial have been published, ⁴ with clinical follow-up up to 6 months after surgery showing no significant difference in outcome between the two groups of patients, apart from operating time. These findings supported the proposition that division of the short gastric vessels was unnecessary for the performance of a laparoscopic Nissen fundoplication. Similar outcomes have been reported by Luostarinen and Isolauri ⁵ from a randomized trial of 50 patients undergoing open Nissen fundoplication and Blomqvist et al ⁶ from a randomized trial of 99 patients undergoing laparoscopic Nissen fundoplication. At the time of reporting the early outcomes for our trial, however, it was inappropriate to extrapolate the trial’s short-term outcomes to long-term follow-up. For this reason we are now reporting the clinical outcome of our original trial with longer-term follow-up of at least 5 years in all of the original patients.

METHODS

From May 1994 to October 1995, 102 patients undergoing a laparoscopic 360° Nissen fundoplication were entered into a randomized trial of division versus no division of the short gastric vessels. Fifty patients were randomized to undergo fundoplication without short gastric vessel division and 52 to undergo division of these vessels. Full details of the methods used in the trial have been reported previously. ⁴ Between July and October 2000, when follow-up was at least 5 years for every patient, we obtained long-term clinical follow-up by interviewing the original patients using a structured questionnaire. This questionnaire was applied using a telephone interview, with the interview conducted by a surgeon (C.O’B.) who was not involved in the previous study. Further, he was unaware of the operation that had been performed.

Patient Selection and Surgical Technique

All patients entered into this trial underwent a laparoscopic Nissen fundoplication for objectively proven gastroesophageal reflux disease. Patients were excluded if they had an esophageal motility disorder that was at the time deemed to preclude a 360° fundoplication, if they required a concurrent abdominal procedure, or if they had undergone previous antireflux surgery. Randomization was carried out after the patient was anesthetized in the operating room.

Laparoscopic Nissen fundoplication was performed using a previously described technique. ^7,8 The short gastric vessels were divided after being secured with metal clips. Division was undertaken from the lower margin of the spleen, proximally to the left pillar of the hiatus. In addition, attachments between the fundus and the undersurface of the diaphragm were routinely divided. The posterior attachments of the fundus were divided only when needed to achieve adequate fundal mobilization. Division was considered complete if the fundus of the stomach could be brought loosely around the esophagus, and the apex of the fundus could be used to construct the fundoplication. If the short gastric vessels were not divided, the anterior wall of the gastric fundus was pulled behind the esophagus, and the anterior fundus, not the apex, was used for the fundoplication.

A loose 1.5- to 2-cm wrap was fashioned with a 52F bougie sited within the abdominal esophagus for the purpose of calibration, for both techniques. If conversion to an open procedure was necessary, the randomization schedule was followed and the patients remained in the trial in the group to which they were originally allocated.

Follow-Up

All patients were unaware of the status of their short gastric vessels after surgery, and follow-up was performed by a surgeon investigator who was also unaware of the status of the vessels. Clinical follow-up was performed at least 5 years after the original surgical procedure. The clinical questionnaire applied was the same as that described in our previous report. ⁴ The presence of each of the following symptoms was sought: heartburn; epigastric pain; regurgitation; dysphagia for lumpy solids, soft solids, and liquids; odynophagia; early satiety; inability to belch; epigastric bloating; anorexia; nausea; vomiting; nocturnal coughing; and nocturnal wheezing. Patients were also asked whether their heartburn was controlled, whether they could relieve bloating after eating, and whether their diet was normal. An additional question inquiring whether patients passed excessive flatus was also asked during the late follow-up interview.

Heartburn, dysphagia for solids, and dysphagia for liquids were also scored using separate visual analog scales (0 = no symptoms, 10 = severe symptoms). Dysphagia was scored using a previously validated dysphagia score ⁹ (0 = no dysphagia, 45 = severe dysphagia). Patients were asked to rank the outcome of their surgery using a modified Visick grading (range 1–5, good outcome = 1 or 2) and whether they regarded the outcome of their surgery to be excellent, good, fair, or poor. Overall satisfaction with the surgical outcome was also assessed using a visual analog scale (0 = dissatisfied, 10 = satisfied). Formal evaluation using pH monitoring or esophageal manometry was not undertaken at 5 years of follow-up. The earlier outcomes of these studies were reported previously. ⁴

Statistical Analysis

All analyses were performed on an intent-to-treat basis, with all patients remaining in their initial allocated group for this analysis. Data were entered into a dedicated database (FileMaker Pro version 4; Claris Corp., Santa Clara, CA) and analyzed using statistical software (Instat version 2.01; Graphpad Software, San Diego, CA). The chi-square test was used to determine the significance of 2 × 2 contingency tables, and the Fisher exact test was applied as appropriate where numbers were small. A two-tailed Mann-Whitney test was used to assess differences between sets of nonparametric data.

RESULTS

Between May 1994 and October 1995, 102 patients who underwent a laparoscopic Nissen fundoplication at the Royal Adelaide Hospital were entered into this trial. Long-term follow-up was obtained for all patients. Three patients were not able to provide detailed clinical follow-up. Two had died of causes unrelated to their original surgery: a 37-year-old woman with diabetes died after an insulin overdose at 12 months, and a 63-year-old man died after esophagectomy performed for severe dysplasia in Barrett’s epithelium at 13 months after surgery. A 69-year-old woman suffered a cerebrovascular accident 4 years after her fundoplication and could not communicate sufficiently to answer the follow-up questionnaire. Hence, 99 patients (vessels divided in 50, not divided in 49) completed the detailed clinical questionnaire 60 to 76 months after the original surgery (mean follow-up 68 months).

Both study groups were well matched for all preoperative factors previously reported: physical, demographic, preoperative clinical symptoms, grade of endoscopic esophagitis, presence of a hiatus hernia, and esophageal manometry and pH study outcomes. ⁴ During the initial 6 months of follow-up, two patients in the nondivision group required laparoscopic repair of a postoperative paraesophageal hernia, whereas three patients in the division group also underwent reoperation; one required laparotomy 6 hours after surgery for bleeding from a short gastric vessel, one patient required laparoscopic correction of a tight hiatal repair on postoperative day 5, and a further patient required open surgical revision for tight scarring at the esophageal hiatus 12 weeks after surgery.

Late reoperations (beyond 6 months) were performed in three patients because of failure of the original procedure. In the nondivision group a patient required open revision at 11 months for recurrent reflux resulting from a “slipped” Nissen fundoplication. In the division group dysphagia from a tight hiatus (hiatal stenosis) ⁸ led to open revision with widening of the esophageal hiatus at 9 months in one patient, and a further patient with recurrent reflux underwent successful laparoscopic correction of a “slipped” Nissen fundoplication at 6 years. In addition to these procedures, a further three patients underwent repair of a port-site hernia at 2, 4, and 6 years after surgery (two in the nondivision group, one in the division group). Eight patients (three in the nondivision group, five in the division group) also underwent endoscopy with dilatation for dysphagia. Six of these patients required a single dilatation only, whereas two patients required subsequent dilatations (both patients in the nondivision group).

At late follow-up there were no significant differences between the groups for the symptoms of epigastric pain, regurgitation, odynophagia, early satiety, inability to belch, anorexia, nausea, vomiting, nocturnal coughing, and nocturnal wheezing (Table 1). However, there was a significant difference in the incidence of flatus production, the occurrence of epigastric bloating, and a decreased ability to relieve bloating by belching, with a less satisfactory outcome in the group who underwent division of the short gastric vessels.

Table 1. SYMPTOMS

No significant differences between trial groups (P > .05 at all follow-up intervals except *P = .02,

†P = .04,

‡P = .03 at late follow-up, Fisher exact test.

There was no difference in the incidence of heartburn when estimated empirically by a yes/no question or by the visual analog scale (Table 2). Although 18% (not divided) and 12% (divided) of patients reported the presence of at least occasional heartburn when asked the yes/no question, only five (three in the nondivision group, two in the division group) scored this symptom as greater than 5 of 10 on the visual analog scale. Most patients scored more than 5 of 10 before surgery. Nine patients (9%) were taking acid suppressive medication at the time of assessment (six in the nondivision group, three in the division group). Five of these patients (two in the nondivision group, three in the division group) had no recurrence of heartburn symptoms, and one reported minimal heartburn only (score 1 of 10 on visual analog scale). These patients had been prescribed medications by their general practitioner for a variety of symptoms unrelated to their previous reflux problem, and a further patient had developed a gastric ulcer. Only two patients were actually taking medication for reflux disease. One was taking 40 mg omeprazole per day for severe heartburn. This patient was also taking a nonsteroidal antiinflammatory agent for arthritis. The other patient was taking ranitidine for moderate heartburn symptoms that, despite the need for medication, had been markedly improved by surgery.

Table 2. ASSESSMENT OF HEARTBURN BY VISUAL ANALOG SCALE

Figures expressed as mean (95% confidence intervals).

There was no difference between the groups in the incidence or severity of dysphagia when estimated by yes/no questions, visual analog scales, or the composite dysphagia score (Table 3). Of the patients who reported dysphagia to solids when asked the yes/no question, only five described dysphagia sufficient to require major dietary modification (two in the nondivision group, three in the division group; composite dysphagia score >30/45) at late review. Of the 10 patients who had a composite dysphagia score of more than 22.5 of 45, 7 had undergone division of the short gastric vessels (P = .19).

Table 3. DYSPHAGIA ASSESSMENT

Figures are either percentage of total or mean (95% confidence intervals).

No tests for significance between groups at comparable follow-up intervals were significant (P > .05 at all follow-up intervals).

The overall outcome at late follow-up is summarized in Table 4. There was no overall difference between the groups in general patient satisfaction with the outcome of surgery. Five patients were not satisfied with the late outcome of their surgery. Four of these patients were in the nondivision group. One was not satisfied because she had required revision for a “slipped” Nissen at 9 months, and although she actually had minimal symptoms at late follow-up, she believed that the surgery had not corrected the additional symptom of a sore throat. Two patients (one in each group) offered recurrent heartburn as the predominant reason for their dissatisfaction, and two further patients (both in the nondivision group) were unhappy about postoperative dysphagia that had necessitated dilatation.

Table 4. OUTCOME, SATISFACTION, AND VISICK GRADING

No tests for significance between groups at comparable follow-up intervals were significant (P > .05 at all follow-up intervals).

DISCUSSION

Nissen fundoplication, whether performed laparoscopically or by an open technique, involves the creation of a circumferential fundal wrap around the distal esophagus. To reduce the risk of postoperative problems, it is thought that this should be constructed loosely, with many surgeons advocating routine division of the short gastric vessels to fully mobilize the gastric fundus, thereby facilitating the formation of a loose “floppy” wrap. ^10,11 However, Nissen’s original procedures and Rossetti’s subsequent modification achieve a similar effect without division of the short gastric vessels. ¹²

Findings from case series that compare outcome with data from historical trials have been used to support the argument for routine division of the short gastric vessels during both open and laparoscopic Nissen fundoplication. ^10,11 Protagonists of division argue that dysphagia rates and reoperation rates are lower if the vessels are divided. However, because outcome data from historical controls whose vessels were left intact were used to support this conclusion, ¹³ there is a significant risk that the findings of these studies were influenced by a learning curve bias, where experience-related improvement in outcomes falsely conveys benefit for vessel division. Similarly, retrospective analyses of reasons for failure after laparoscopic Nissen fundoplication have also been used to advocate vessel division. These studies have usually compared earlier experience with a later series of patients, when numerous modifications to the original procedure have been introduced with time. ^14,15

Using similar methodology, some other surgeons have shown no benefit for patients undergoing division of the short gastric vessels during Nissen fundoplication. In a recent multicenter analysis of 1,470 patients, Pessaux et al ¹⁶ compared the outcomes of Nissen (n = 655), Rossetti (n = 423), and Toupet (n = 392) procedures. Short gastric vessel division provided no clinical benefit in terms of complications, reoperation rate, or dysphagia at various time intervals up to 2 years after surgery. Many authors report good outcomes from extensive experiences with Nissen fundoplication without division of the short gastric vessels. ^3,17,18

To resolve the question of whether division of the short gastric vessels is necessary, data from prospective randomized trials are required. The outcome of these studies should not be influenced by the problem of learning curve bias. The early report from our trial ⁴ and those of Luostarinen et al ⁵ and Blomqvist et al ⁶ are the only randomized trials that have addressed this issue. Luostarinen et al randomized 50 patients undergoing an open Nissen fundoplication to have their short gastric vessels either divided or left intact. At median follow-up of 36 months, no clinical outcome differences were seen between the two techniques for the problem of dysphagia, ability to belch, amount of flatus passed, or gas bloat symptoms. However, there was a significantly increased incidence of recurrent sliding hiatus hernia, as well as a trend toward a greater incidence of a defective wrap and recurrent reflux symptoms in the group of patients who had undergone division of the short gastric vessels. Blomqvist et al randomized 99 patients and found that division of the short gastric vessels conferred no outcome benefit at 1 year after surgery. The only significant difference between both techniques was the finding of a longer operating time for vessel division. Further, two patients required urgent early reoperation for complications directly related to vessel division.

The results of our trial, similar to those reported by the above authors, show no benefit from dividing the short gastric vessels. In our earlier report with 6 months of follow-up ⁴ we could not show any statistically significant outcome differences between the two groups, with the exception of longer operating time in the group of patients randomized to undergo vessel division. However, there was a trend toward an increased incidence of epigastric bloating (41% vs. 35%) and inability to relieve bloating (43% vs. 35%) in the division group. ⁴ At late follow-up this trend has persisted, and there is now an approximately 20% greater incidence of epigastric bloating and inability to relieve bloating in the division group. The rate of excess flatus production, which was not asked about in the original study, is high in both groups, but more so in the division group at late follow-up. We found this surprising, because most surgeons would expect that division of the short gastric vessels would guarantee a loose, floppy wrap that would be more amenable to the escape of gas, and therefore would hypothesize that vessel division would be associated with less wind retention and less flatus production per rectum. This raises the possibility that mechanisms other than the apparent “tightness” of the wrap are responsible for these postfundoplication problems.

It is known that the belch reflex originates when stretch receptors in the fundus of the stomach are stimulated. ¹⁹ One possible explanation of our findings is that division of the short gastric vessels is dividing the afferent nerves responsible for this reflex. The argument that it is not the degree of tightness of the wrap is supported by the fact that in this study there was no difference in the incidence of recurrent heartburn or dysphagia between the groups at late follow-up. The fact that only 5% of patients overall reported significant heartburn suggests that both variants of the Nissen fundoplication achieve effective long-term control of reflux.

At 6 months there was no difference between the groups for all measurements of dysphagia. This is also true at late follow-up. Interestingly, and perhaps somewhat disappointingly, there appears to have been some deterioration in the dysphagia scores for the same patient group compared with our previously reported data from 6 months of follow-up. This appears to be contrary to the belief that dysphagia after Nissen fundoplication continues to improve with time. However, the overall level of dysphagia at late follow-up in this report was not much greater than that identified before surgery in both study groups. Also, dysphagia of a degree spoiling the outcome was uncommon.

The reliability of the clinical data collected in this trial and the method of collection are important issues to consider when determining the validity of conclusions drawn from the trial. We have sought to reduce the risk of bias in data collection by ensuring that a single masked investigator asked each patient a standardized set of questions, and by applying a previously validated dysphagia score. ⁹ It could be argued that the interviews should have been undertaken in person rather than by telephone, but this was impractical given that the geographic mobility of the Australian population has resulted in at least 20% of the trial patients moving more than 1,000 km away from our hospital. Only the use of telephone interviews enabled us to achieve 100% follow-up. For these reasons, we believe that we have avoided the potential for bias inherent in incomplete follow-up, lack of blinding, and follow-up by the operating surgeon. The further potential problem of learning curve bias was minimized by commencing the trial after our overall experience with laparoscopic Nissen fundoplication exceeded 200 procedures.

It could be argued that more reliable follow-up would have been obtained if pH monitoring or esophageal manometry were performed at the 5-year follow-up point. However, we have previously reported that the results of these studies do not correlate well with clinical outcome after laparoscopic Nissen fundoplication, ²⁰ and technical success judged by these tests or judged by the operating surgeon does not always equate to clinical success as determined by the individual patient. Hence the clinical outcome, as determined by the patient, is likely to be more useful as a measure of success than the outcome of physiologic testing or other parameters.

A further area in which this trial can be criticized is the issue of posterior fundal mobilization. In our study, not all patients undergoing division of the short gastric vessels had the posterior fundal attachments divided. Nevertheless, a recent randomized trial of Nissen fundoplication with division of the short gastric vessels and posterior versus no posterior fundal mobilization has not shown any benefit for complete fundal mobilization. ²¹ Further, in the randomized study of Blomqvist et al, ⁶ where no division of short gastric vessels was compared with complete fundal mobilization, again no difference in outcome was found.

Data from our trial have now shown no benefit for division of the short gastric vessels during laparoscopic Nissen fundoplication at both short- and long-term follow-up. This study, in association with data reported previously, suggests that short gastric vessel division lengthens the procedure, adds complexity and expense, and is followed by a higher incidence of “wind” problems. On the basis of these findings, we believe it is unnecessary to divide the short gastric blood vessels routinely during laparoscopic Nissen fundoplication.