Preoperative and postoperative esophageal acid exposure in the reflux and dysphagia groups. P value with regard to the difference between the preoperative and postoperative values in the reflux group was <.001.
Intrathoracic migrated fundoplication and telescoping in a patient with recurrent gastroesophageal reflux disease after conventional Nissen fundoplication.
Furnée EJB, Draaisma WA, Broeders IAMJ, Smout AJPM, Gooszen HG. Surgical Reintervention After Antireflux Surgery for Gastroesophageal Reflux DiseaseA Prospective Cohort Study in 130 Patients. Arch Surg. 2008;143(3):267-274. doi:10.1001/archsurg.2007.50
Surgical reintervention after antireflux surgery for gastroesophageal reflux disease is required in 3% to 6% of patients. The subjective outcome after reintervention has been reported in several studies, but objective results after these subsequent operations have rarely been published. The purpose of this study was to assess the symptomatic and objective outcomes in patients who underwent subsequent operation because of recurrent reflux symptoms or troublesome dysphagia after primary antireflux surgery.
Prospective cohort study.
University medical center.
Between January 1, 1994, and March 31, 2005, 130 patients (mean [SD] age, 48.4 [14.1] years) undergoing surgical reintervention after antireflux surgery for gastroesophageal reflux disease were prospectively studied.
Main Outcome Measures
Symptomatic outcome was determined by questionnaires. Esophageal manometry and 24-hour pH monitoring were performed to assess the objective outcome.
A total of 144 reinterventions were performed in 130 patients, for recurrent reflux in 94 patients (65.3%) and for troublesome dysphagia in 50 patients (34.7%). Belsey Mark IV fundoplication through a left-sided thoracotomy was performed in 78 (54.2%) and a subsequent Nissen or partial fundoplication during 66 reinterventions (45.8%), including 16 laparoscopic procedures. After a mean (SD) follow-up of 60.1 (37.2) months, symptoms were absent or significantly improved in 70.3% of patients and esophageal acid exposure was normalized in 70.2% of patients after surgery. Postoperative complications occurred after 14 subsequent operations (9.7%).
Surgical reintervention after antireflux surgery for gastroesophageal reflux disease yielded good symptomatic and objective results in 70% of patients in this prospective cohort study. Since the morbidity of this type of surgery is far from negligible, the expectations should be discussed in detail before additional operation.
Antireflux surgery is an effective surgical intervention for gastroesophageal reflux disease (GERD). Satisfactory results have been reported in 85% to 90% of patients up to 5 years after surgery.1- 4 Revisional surgery is necessary in 3% to 6% of patients.5- 10 Indications for surgical reintervention can be subdivided into recurrent reflux symptoms and troublesome dysphagia.11 The cause of failure may be technical errors, wrap migration or telescoping, an incorrect primary diagnosis, or selection of a wrong procedure in patients with severe esophageal motility disorders.5 Frequently, however, no cause of failure can be demonstrated, and surgical reintervention after failed fundoplication is known to be technically demanding.
Several studies10,12- 27 have been published on the outcome of additional operation. Because of the relative scarcity of these procedures, most of these studies are small. Furthermore, most studies have been directed toward the feasibility of and the subjective outcome after reintervention, whereas objective data on postoperative results are rarely available. Therefore, the effect of revisional antireflux surgery on functional outcome remains unclear. This prospective study aimed to assess the subjective and objective outcomes in patients who underwent subsequent operation for either recurrent reflux symptoms or troublesome dysphagia in a tertiary care referral center.
A total of 130 patients who underwent surgical reintervention for recurrent reflux symptoms or troublesome dysphagia after 1 or more failed previous antireflux procedures between January 1, 1994, and March 31, 2005, were included in this prospective cohort study. Patients with a paraesophageal hernia without documented GERD as the indication for primary surgery and patients with achalasia were excluded.
The indication for the initial procedure and subsequent operation, the type of procedures and treatment with proton pump inhibitors, and the use of histamine2-receptor antagonists and prokinetic drugs before the subsequent operation were assessed. Intraoperative anatomical findings that suggested a possible cause of failure for the previous antireflux procedure and intraoperative and postoperative complications were documented.
Information on symptomatic outcome was obtained by sending a standardized symptom questionnaire that included the Gastroesophageal Reflux Disease Health Related Quality of Life (GERD-HRQoL) score to all patients. This questionnaire contains 9 GERD-specific items that had to be scored according to a system that combined severity and frequency, eventually yielding a score between 0 and 45.28 Furthermore, a Visick score (symptoms resolved, improved, changed, or worsened) and preoperative and postoperative general quality of life using a visual analog scale were assessed.
Stationary esophageal manometry and ambulatory 24-hour esophageal pH monitoring were performed before and after subsequent operative antireflux procedures. Written informed consent was obtained from all patients who agreed with both investigations.
Use of all medication that could affect the results was stopped before esophageal manometry and 24-hour pH monitoring were performed. Manometry was performed using a water-perfused system with a multiple-lumen catheter that had an incorporated sleeve sensor (Dentsleeve Pty Ltd, Adelaide, Australia). The sleeve sensor was positioned at the level of the lower esophageal sphincter (LES). Mean end-expiratory LES pressure was determined in response to 10 wet swallows (5-mL water bolus). The end-expiratory gastric baseline pressure served as the zero-reference point. An LES pressure of 4.5 to 26.3 mm Hg was defined as normal. Thereafter, esophageal pH monitoring was performed by positioning a pH glass electrode (model LOT 440 M3; Medical Instruments Corporation, Solothurn, Switzerland) 5 cm above the manometrically determined proximal margin of the LES. The pH signals were recorded with a digital portable data recorder using a sample frequency of 1 Hz (Orion; Medical Measurements Systems BV, Enschede, the Netherlands). During 24-hour recording of intraluminal pH, patients registered the episodes of reflux-related symptoms and the periods of supine and upright positions. Reflux parameters were calculated with an analysis program (Medical Measurements Systems BV). Reflux was defined as abnormal if the pH was lower than 4 for more than 5.8% of the total time of pH recording and with the patient in the upright and supine positions during more than 8.2% and more than 3.5% of the time, respectively.29 Moreover, the symptom index and the symptom association probability were calculated. Symptoms were considered to be reflux related if the symptom index was larger than 50% or the symptom association probability was more than 95%.30,31
The surgical approach depended on the localization of the wrap. An abdominal approach was performed if (the remnants of) the fundoplication were below the diaphragm. In the case of recurrent GERD, Nissen fundoplication was performed. In patients who underwent a subsequent operation for dysphagia, additional Nissen fundoplication was performed if other causes than a too-tight wrap could be held responsible for the dysphagia. In all other circumstances, the 360° Nissen fundoplication was transformed into a partial posterior fundoplication. Left-sided thoracotomy was chosen in case of migration of the wrap to the posterior mediastinum, and a 270° wrap according to Belsey Mark IV was performed independent of the indication for subsequent operation. All abdominal additional operations were performed by 2 surgeons (H.G.G. and I.A.M.J.B.) with adequate experience in upper gastrointestinal tract procedures, and all thoracic approaches were performed by 1 surgeon (H.G.).
Patients were analyzed and compared according to the indication of subsequent operation (ie, recurrent reflux symptoms or troublesome dysphagia) and accordingly assigned to the reflux group (RG) or dysphagia group (DG). Patients who underwent more than 1 additional operation were analyzed depending on the indication of their last additional operation. Furthermore, intraoperative and postoperative complications, quantity of blood loss, and duration of operation and hospital stay were compared between abdominal (conventional and laparoscopic) and thoracic subsequent operations.
Symptomatic outcome was defined as successful if postoperative Visick scores on symptoms were marked as resolved or improved. Successful objective outcome was defined as the absence of abnormal esophageal acid exposure.
Values were expressed as mean (SD). Data were analyzed using SPSS statistical software version 12.0.2 for Windows (SPSS Inc, Chicago, Illinois). The paired-sample t test was used for statistical analysis of continuous preoperative and present values and the t test for independent samples for statistical analysis of continuous values between groups. Statistical analysis of categorical values between groups was performed with the Pearson χ2 test and statistical analysis of categorical preoperative and present values with the McNemar test. The Spearman coefficient was used to determine correlations. Preoperative and present differences and differences between groups were considered statistically significant at P <.05.
A total of 130 patients (65 men and 65 women; mean [SD] age, 48.4 [14.1] years) participated in the study. Nissen fundoplication was the most performed primary antireflux operation (110 patients [84.6%]). Of these, 44 patients (40.0%) were operated on laparoscopically. In 8 patients (6.2%), partial fundoplication served as the primary antireflux operation. Two of these operations were performed laparoscopically, 3 through a left-sided thoracotomy, and 3 through an upper abdominal midline incision. The other primary interventions were gastropexy (5.4%), implantation of an Angelchik prosthesis (2.3%), and 2 other procedures (1.5%), all performed through an upper abdominal midline incision. Forty-seven of the primary interventions (36.2%) were performed in our center, and the remaining ones in other hospitals in the Netherlands. Reflux-related symptoms were the indication for primary intervention in all but 1 patient, who was primarily operated on for dysphagia. Before referral, 16 patients (12.3%) had undergone another antireflux procedure because of failure of the primary antireflux operation.
Overall, a total of 144 reinterventions were performed in 130 patients. Fourteen patients underwent 2 reinterventions. The mean (SD) interval between the first antireflux procedure and the subsequent operation was 52.9 (70.6) months. The indication for subsequent operation was recurrent reflux symptoms in 94 operations (65.3%) and troublesome dysphagia in 50 (34.7%). The indication for the first additional operation in those patients who underwent 2 reinterventions was recurrent reflux symptoms in 11 and dysphagia in 3 patients. In the RG, 94 subsequent operations were performed: 55 thoracic and 39 abdominal. In the DG, 50 subsequent operations were performed: 23 thoracic and 27 abdominal. An abdominal approach was used in 66 subsequent operations (45.8%): 54 Nissen fundoplications (43 open [29.9%] and 11 laparoscopic [7.6%]) and 12 partial wraps (7 open [4.9%] and 5 laparoscopic [3.5%]). The other 78 additional operations (54.2%) were Belsey Mark IV fundoplications. Collis gastroplasty was added in 9 patients: 7 in addition to a Belsey Mark IV procedure and 2 during open Nissen fundoplication.
Several causes of failure of the previous antireflux procedure were identified during subsequent operation (Table 1). At least 1 anatomical abnormality was found during 131 (91.0%) and 2 or more during 39 subsequent operations (27.1%). Complete or partial disruption of the wrap and cephalad slippage of the gastroesophageal junction and stomach through the wrap were the most common findings, followed by migration of the wrap into the chest and the presence of a paraesophageal hernia. Disruption of the wrap was more frequently found during subsequent operation performed for recurrent reflux symptoms, whereas the absence of any intraoperative anatomical abnormality was more frequent during subsequent operation for dysphagia.
Intraoperative complications are listed in Table 2. Splenectomy for injury of the spleen was required during 2 subsequent abdominal operations (3.0%). Injury of the liver and (contralateral) pneumothorax occurred more frequently during subsequent thoracic operations (P = .04 and P = .001, respectively). Three subsequent laparoscopic operations (18.8%) were converted to laparotomy. Subsequent thoracic operations were associated with significantly greater blood loss (mean [SD], 539  vs 221
 mL; P <.001) and longer operating times (mean [SD], 161  vs 106  minutes; P <.001).
Postoperative complications occurred after 14 subsequent operations (9.7%) (Table 2). Pulmonary complications were more frequent after thoracotomy (P = .03). Two patients (1.4%) died, both after a thoracotomy. One patient died of acute cardiac arrest, and postmortem evaluation did not reveal a cause directly related to the surgical procedure. The other patient had septic shock and postoperative bleeding from the thoracic aorta with a low-grade local sepsis and fistula to the esophagus due to foreign material implanted during the first operation. Hospital stay was shorter after subsequent abdominal operations (mean [SD], 7.9 [2.9] vs 14.4 [14.0] days; P <.001).
A total of 118 patients (90.8%) were available for follow-up. Eight patients were lost to follow-up, 2 patients refused collaboration, and 2 patients died during follow-up (Figure 1). The mean (SD) time to follow-up for the subjective outcome was 60.1 (37.2) months.
Daily heartburn was documented by 15 patients (20.0%) and daily symptoms of dysphagia by 9 (12.0%) in the RG. Nine patients (20.9%) in the DG still experienced dysphagia every day at follow-up. Heartburn on a daily basis was present in 13 of these patients (30.2%). No significant differences were found between the RG and DG with regard to the presence of daily heartburn and dysphagia, GERD-HRQoL score, self-rated change in symptoms, and general quality of life (Table 3). Overall, 36 patients (30.5%) had heartburn and/or dysphagia on a daily basis at follow-up. No significant correlation was found between heartburn and the presence of abnormal acid exposure (r = −0.01; P = .92). Use of antisecretory drugs decreased significantly in the RG (P = .001; Table 3). In the DG, 13 patients (30.2%) used antisecretory drugs but had no reflux symptoms before subsequent surgery. This finding represents a trend toward increased use during follow-up (P = .08). The use of antisecretory drugs was significantly correlated with heartburn (r = 0.40; P = .001) and the presence of abnormal acid exposure (r = 0.31; P = .003) at follow-up.
Sixty-three patients (75.9%) in the RG underwent esophageal manometry and 66 patients (79.5%) underwent 24-hour pH monitoring as part of their preoperative workup before subsequent operation. Manometry was performed in 35 patients (74.5%) and 24-hour pH monitoring in 28 patients (59.6%) before surgery in the DG. Five patients were lost to follow-up, and 1 patient was excluded for objective assessment because measurement was unreliable owing to subtotal gastric resection after Belsey Mark IV fundoplication. Moreover, 29 patients refused manometry, pH monitoring, or both. Eventually, 102 patients (78.5%) underwent stationary esophageal manometry, and 94 (72.3%) underwent 24-hour pH monitoring during the follow-up period (Figure 1). No significant difference in frequency of daily heartburn and/or dysphagia was found between patients who refused both investigations and patients in whom objective data were obtained. The mean (SD) time to follow-up for the objective outcome was 18.6 (30.5) months. Mean (SD) preoperative end-expiratory LES pressure was significantly higher in the DG than the RG (12.8
[7.6] mm Hg vs 9.1 [6.1] mm Hg; P = .009). Esophageal motility was preoperatively abnormal (hypertensive LES pressure, incomplete relaxation of the LES, and/or low-amplitude peristaltic contractions of the distal part of the esophagus) in 30 patients (63.8%) in the DG. At least 1 unsuccessful dilation before surgery was performed in 22 patients (46.8%) in this group. Mean (SD) LES pressure decreased significantly in the DG to 8.5 (6.2) mm Hg at follow-up (P = .03) but not in the RG (9.1 [5.5] mm Hg; P = .83). This difference in LES pressure in the DG was not correlated with the incidence of postoperative symptoms of daily dysphagia in patients who had a subsequent operation through a left-sided thoracotomy. For patients with a subsequent abdominal operation for dysphagia, however, a significant correlation was detected between decrease in LES pressure and postoperative dysphagia (r = 0.56; P = .03). In half of the patients who had daily symptoms of dysphagia at follow-up, esophageal motility was abnormal.
Results of preoperative and postoperative 24-hour pH monitoring are presented in Table 4. Mean preoperative total esophageal acid exposure was abnormal in patients in the RG only (Figure 2) and decreased significantly after surgery. Symptom indexes (ie, symptom index and symptom association probability) also decreased significantly after surgery. No differences were detected between preoperative and postoperative 24-hour pH profiles in patients in the DG. Furthermore, no differences in acid exposure times and symptom indexes existed between the RG and DG at follow-up. Excessive esophageal acid exposure at follow-up was still present in 21 patients (33.3%) in the RG and 7 patients (22.6%) in the DG (P = .28).
Overall, 70.3% of the patients had a successful symptomatic outcome and 70.2% had a successful objective outcome at follow-up without significant differences between the RG and DG. Subtotal gastric resection after subsequent operation was performed in 3 patients. Indications were persistent reflux symptoms refractory to medication in 1 and intractable symptomatic gastroparesis due to vagal nerve damage in 2 other patients.
To our knowledge, this is the largest prospective study to date on subjective and objective outcomes of surgery for recurrent GERD and troublesome dysphagia after failed antireflux surgery. An overall successful symptomatic and objective outcome was demonstrated in 70% of patients, regardless of the surgical strategy, laparotomy, laparoscopy, or thoracotomy used.
Intraoperative and postoperative complications, including the incidence of splenectomy, were comparable to those reported in other studies,5,7,20,22 although higher after subsequent operation through a left-sided thoracotomy. The thoracic approach is prone to causing more pulmonary complications. Nevertheless, we prefer this approach in case of intrathoracically herniated wrap because it is usually fixed to the surrounding tissue, and with this approach the esophagus can be dissected free from the mediastinum to gain length. Morbidity, including postthoracotomy pain, however, is discussed in detail before subsequent operation.
Pathophysiologic features, analysis of results, and anticipated outcomes for subsequent operations performed because of recurrent GERD and troublesome dysphagia are different. In the first group of patients, the wrap subsided in the abdomen or in the thorax, resulting in an insufficient gastroesophageal antireflux barrier and subsequent recurrent symptoms and/or esophagitis. In case of dysphagia, the pathophysiologic defect is far less clear. Different mechanisms can be a factor in unsatisfactory outcome, such as a narrow wrap, tight crural repair, paraesophageal hernia through the wrap, or migration of fundoplication into the mediastinum, isolated or in combination.5 Although the surgical treatment of these complications is less straightforward, it has been consistent for years in this cohort.
An anatomical explanation for failure of antireflux surgery was found in 90.3% of subsequent operations in the present study. Other studies7,15,16,21,26 report figures ranging from 69% to 100%. Abnormalities at the gastroesophageal junction responsible for dysphagia are frequently lacking during subsequent operation, which was also recognized in this study. Intrathoracic wrap migration was the standard indication for a thoracotomy; however, it was also found during 10 subsequent abdominal operations. This finding can partially be explained by the inaccuracy of the preoperative esophagogram. During the last 3 years of the study period, the procedure was started with diagnostic laparoscopy, and if the migrated wrap could be freed, a new laparoscopic Nissen fundoplication procedure was performed. If the wrap was fixed or not enough length could be gained, the procedure was converted to a left-sided thoracotomy. Symptomatic and objective outcome after subsequent laparoscopic operations was comparable to the conventionally operated-on group. Several other larger studies12,21,27,32,33 have been published showing the feasibility of this approach, with successful outcomes in 65% to 93% of patients. However, the conversion rate was high in this study (19%), which might be explained, as mentioned herein, by the fact that we prefer to start laparoscopically in patients with an intrathoracic wrap after primary laparoscopic Nissen fundoplication. Going forward, we will consider the possibility of conversion to a left-sided thoracotomy if the wrap cannot be freed from the posterior mediastinum.
Absence of daily reflux or dysphagia symptoms correlated well with Visick scores, and this was attained in approximately 70% of patients in both groups. As for documentation of symptomatic results, we have abandoned separate scoring of heartburn, regurgitation, and dysphagia and have now returned to the Visick grading as probably the most useful instrument to document the overall subjective outcome of antireflux surgery. At follow-up, the use of antisecretory drugs significantly decreased in patients subsequently operated on for recurrent reflux symptoms, whereas this did not differ in patients operated on for dysphagia. Nevertheless, figures at follow-up were high compared with other studies19,20,23,32,33 that reported percentages varying from 7% to 46%. Although the use of antisecretory drugs was not based on abnormal acid exposure in all patients, a significant correlation between the use of proton pump inhibitors and the presence of abnormal esophageal acid exposure after surgery was found in this study. This finding stands in contrast to earlier observations by our group.34 We did not identify the reason for this treatment in patients with physiologic acid exposure.
In the RG, preoperative pH monitoring was not performed in 20.5% of patients, as was the case for esophageal manometry in 25.5% in the DG. Patients in the RG, however, had typical symptoms and esophagitis on endoscopy, whereas the patients who were subsequently operated on for dysphagia without manometry had severe dysphagia (ie, they lost weight despite maximal use of a high-calorie liquid oral food supply) and abnormal position of the wrap, a combination that justified additional operation. It has become the standard of care to perform additional operations for such anatomical abnormalities as demonstrated during endoscopy and barium swallow with either recurrent reflux symptoms and/or dysphagia.
With regard to objective outcome, there were some unexpected findings. The decrease in gastroesophageal reflux in patients operated on for reflux was significant, but normalization was not attained in all patients. This finding has been observed in a previous study as well, and full normalization is apparently not an absolute prerequisite for a subjectively successful outcome.35 In 2 of these patients, preoperative acid exposure was not measured, and in 9 patients acid exposure did not normalize but decreased at follow-up. Further analysis of objective failure of subsequent operation was not performed.
Outcome of pH monitoring was different between patients who underwent additional operations by an abdominal and thoracic approach in both the RG and DG; however, this finding is probably not directly influenced by the surgical approach but by the indication the approach was based on. By separately analyzing either group, a decrease in the number of reflux episodes in the thoracic group was found and no change was found in the patients operated on with laparotomy or laparoscopy in the DG. This finding probably means that the mechanism of recurrent reflux is not only caused by an insufficient LES complex but also by the anatomical position of the wrap. In case of the combination of dysphagia and reflux, with telescoping of the wrap and the LES migrated to the posterior mediastinum, the part of the fundus above the wrap continues to secrete acid, which easily refluxes through the LES because of the obstruction by the telescoped wrap, in a subdiaphragmatic position (Figure 3). Subsequent operation in this subgroup was found to reduce the number of reflux episodes but not the total percentage of esophageal reflux time.
In conclusion, our study demonstrated satisfactory symptomatic and objective results in 70% of patients who underwent subsequent thoracic or abdominal surgery for recurrent reflux symptoms or troublesome dysphagia. These subsequent operations, however, are accompanied by higher morbidity and mortality than primary intervention. Patients should be informed about the outcome of additional operation, and the results of this study help to tailor this preoperative counseling to the type and anatomy of the failure of the first operation.
Correspondence: Hein G. Gooszen, MD, PhD, University Medical Center Utrecht, Department of Surgery, HPG04-228, PO Box 85500, 3508 GA Utrecht, the Netherlands (email@example.com).
Accepted for Publication: November 3, 2007.
Author Contributions:Study concept and design: Furnée, Draaisma, Broeders, Smout, and Gooszen. Acquisition of data: Furnée, Broeders, Smout, and Gooszen. Analysis and interpretation of data: Furnée, Draaisma, Broeders, Smout, and Gooszen. Drafting of the manuscript: Furnée. Critical revision of the manuscript for important intellectual content: Draaisma, Broeders, Smout, and Gooszen. Statistical analysis: Furnée and Draaisma. Administrative, technical, and material support: Smout. Study supervision: Draaisma, Broeders, Smout, and Gooszen.
Financial Disclosure: None reported.