Gastrographin swallow demonstrating an acute postoperative paraesophageal herniation
A correlation between the cause of recurrence and the interval between the primary procedure and the symptomatic failure has been reported. In the first three years, recurrence is generally due to wrap slippage while wrap breakdown is usually the most frequent cause beyond three years . The probability of failure significantly increases over time. After 20 years, the probability of failure requiring remedial surgery is estimated at 30% .
Although the laparoscopic approach in revisional surgery for failed antireflux [16, 17] first met skepticism, results obtained after the initial learning curve have established minimally invasive surgery as the first management option. From a technical standpoint, the key points for an adequate surgical revision include the identification of complex anatomy, the evaluation of failure mechanisms, and the choice of the appropriate surgical procedure to correct the problems. Herniation and slippage are usually induced by an inadequate crural closure and/or excessive traction on the gastroesophageal junction and an unrecognized short esophagus. Repair under tension, which is axial on the esophagus and radial on the crura, is probably the main cause of failure. Regarding crural closure, some groups advocate the use of a mesh. However, mesh use is still a matter of debate, with heterogeneous results and complications rates [18–20]. The reality of shortened esophagus is still discussed .
In our series of laparoscopic reoperations, recurrent GERD (41.7%) and dysphagia (16.6%) were the most frequent indications for redo, followed by a combination of reflux and dysphagia (4%) and gas bloating. Most frequent anatomical patterns of failure were herniation and slippage, followed by a valve or cruroplasty, which was too tight, and a twisted or bilobed stomach .
17.3 Reoperation Techniques
The strategy in reoperations must be tailored to the patient’s symptoms, patterns of failure, type and quality of the previous procedure, and number of previous operations. Recent reports on long-term outcomes provided additional information on the sustainability of different types of redo procedures.
One fundamental requirement in redo surgery is to dismantle the first fundoplication and restore the esophagogastric anatomy  (Fig. 17.3). The second step is to understand the mechanism of failure, which, besides surgical mistakes, is mostly related to some mechanical tension on the elements of the repair, fundoplication wrap, esophagus or crural muscles. The third and most important step is to select the appropriate redo procedure.
Dismantling of fundoplication
Recurrent GERD caused by a disruption of the wrap is probably the easiest symptom to treat. Indeed, it is usually associated with a poor surgical technique at the primary operation, in terms of suturing capabilities and/or of the suture material used. Tension on the fundoplication is usually associated with a wrap created without gastric fundus mobilization. The redo procedure has to take into account these elements and usually consists in a refundoplication using a Nissen or a partial wrap, which provides good long-term results . A valve under tension is also frequently observed in the new onset of persisting dysphagia, and the treatment is also focused on a tension-free fundoplication, with excellent results. A partial valve is usually recommended, as secondary esophageal motility disorders (pseudo-achalasia) can develop in long-standing esophageal obstruction (Fig. 17.4). Recurrent GERD, whether associated or not with dysphagia, due to wrap slippage can be more challenging. In its simplest form, it is caused by a wrong valve placement at the primary procedure and it requires a new, well-positioned fundoplication. In its more complex form, it can be related to a progressive cephalad slippage of the esophagogastric junction (EGJ) through the wrap. The main causing factor is most commonly the esophageal tension, which pulls the EGJ upwards, or more rarely the fixation of the wrap on the EGJ. In the first case, a high intramediastinal dissection or esophageal lengthening procedure may allow for a tension-free repositioning of the EGJ in an infra-diaphragmatic position and for the creation of a new fundoplication  (Fig. 17.5). A careful intraoperative assessment of the anatomical position of the endoscopic EGJ is mandatory (Fig. 17.6). Indeed, visual assessment is unreliable, due to tissue alterations. The third most common failure pattern is the intrathoracic migration of the fundoplication, or recurrent hiatal hernia, which can be associated with slippage (Fig. 17.7). Reoperation is complex as it has to take into account axial and radial tension on the antireflux repair. Axial tension can be managed similarly to slippage. Radial tension is challenging since crural repair addresses crural muscles which have been previously manipulated, sutured or reinforced with a mesh in the most difficult cases.
Tight fundoplication with esophageal outlet obstruction
Collis lengthening procedure. The first staple line defines the appropriate level of transection of the great curvature (a). The last staple line complete the partial fundectomy (b)
Intraoperative control of the GEJ position
Intrathoracic migration of the valve
Dealing with more than one prior surgery is even more challenging, and the quality of the last repair is fundamental. If it was performed following the standard recommendations, one should consider alternative methods, such as Roux-en-Y bypass, fundoplication dismantling, esophageal myotomy or esophagectomy in the worst-case scenario .
In our experience , herniation was detected in 50 patients out of 129 laparoscopic redo surgeries. Early postoperative paraesophageal herniation required immediate laparoscopic redo, with hernia reduction and a new crural repair. In case of late presentation, one Collis-Nissen procedure was performed, while the remaining procedures varied from redo Nissen (39 patients) to partial wrap in four cases. Sutured crural repair was performed in 58% of patients, reinforced with pledgets or a mesh in other patients. In case of slipped wrap (n = 45), the previous fundoplication was taken down along with esophageal mobilization. Crural repair and a new fundoplication (32 Nissen, 8 Toupet, and 3 Collis-Nissen procedures) were performed. In the remaining two patients, the fundoplication was completely dismantled.
17.4 Intraoperative Complications and Postoperative Outcomes
In one of the most comprehensive summary of outcomes after reoperative antireflux surgery including more than 1500 patients, perioperative mortality was nil, a significant finding considering the complexity of these operations . Intraoperative complications occurred in approximately 19% of cases, with perforation of either the esophagus or the stomach (14%) and accounting for 76% of total intraoperative morbidity. Conversion was required in 1–20% of patients, and was mainly caused by a loss of anatomical landmarks, adhesions, and intraoperative complications [7, 23–25]. In our series of 129 patients undergoing laparoscopic reoperations, the conversion rate due to excessive adhesions was 1.5%, the postoperative complication rate was 7%, and early reoperation was required in three patients due to cardia and pyloromyotomy leaks and to early wrap migration . Those findings compare favorably with the results of one of the largest series of open redo surgeries, with a 10% reoperation rate and a 25.4% postoperative complication rate .
17.5 Long-Term Outcomes
Good to excellent long-term outcomes after primary laparoscopic antireflux surgery have been reported up to 90% of patients . The success rate after reoperative antireflux surgery depends on the indication for redo (dysphagia, recurrent GERD mostly), on the anatomical failure pattern; which can make the redo procedure more or less challenging, and on the type of procedure performed. The need for re-reoperation is a good indicator of success. While the satisfaction rate after the first redo surgery for failure remains valuable at short-term follow-up (>80%), this rate tends to drop with the duration of follow-up and the number of previous operations [27–29].
At a short-term follow-up period (<2 years), good to excellent results were obtained in 74–81% of patients [7, 27, 30]. Better results for the laparoscopic approach were reported in the short run in a comparative study with the open approach, 94 versus 74% respectively, while similar results (70% success rate) were obtained in the long run with laparoscopy, laparotomy or thoracotomy .
Few studies in the literature report results with a long-term follow-up period (>3 years) [32–34] and only one at more than 5 years . Long-term results are more disappointing than short-term ones, even if a group reported an excellent 90% success rate [13, 21].
In a case series by Vignal et al. (median follow-up: 4.5 years, completed in 29 patients) , 2 out of 47 patients required further surgery after a first reoperation at 3 and 7 months respectively, in the first case for wrap herniation and in the second case for complicated recurrent reflux, requiring a total duodenal diversion with truncal vagotomy. The authors also provided a comparison between the redo surgery group and a group of patients who underwent a single antireflux procedure during the same period. The GIQLI score was higher in patients after primary surgery (103 ± 27 versus 82 ± 23 in the redo surgery group). However, the satisfaction rate and the need for proton pump inhibitors (PPIs) were equivalent.
Granderath et al.  reported a satisfaction rate of 93% in their cohort of 27 patients, up to 5 years after surgery. The GIQLI score was significantly higher than the preoperative score, and increased significantly at interval reports at 3 years and 5 years, reaching values comparable to the score of healthy populations. In terms of symptoms, heartburn recurrence was reported in two patients at 3 and 5 years of follow-up, without any evidence of anatomical or morphological failure and successfully treated with low-dose PPIs. Dysphagia persisted in seven cases of patients who underwent a Nissen fundoplication as a redo surgery.
In the extensive retrospective analysis by Awais et al. , a long-term evaluation was performed with the GERD-HRQOL test and the SF36 test in a total of 186 patients available at a median follow-up of 3.3 years. The GERD-HRQOL was excellent in 52.2%, satisfactory in 33.3%, and poor in 14.5% of patients.
Finally, in our long-term study (mean follow-up: 76 months) , resolution of dysphagia was obtained in 68% of patients while recurrent GERD was treated in 73% of patients referred for index symptoms. After a redo procedure for herniation, the probability of being free-from-failure at 5 years was 83 versus 93% in case of slippage. After 10 years, the probabilities were 37 and 50% in case of herniation and slippage respectively. Failure of the repair was found in 41% of patients who underwent redo surgery for herniation, and four of them necessitated an additional operation. After reoperation for wrap slippage, failure was demonstrated in 27% of patients, three of them requiring additional surgery.
Failure after a reoperative antireflux surgery is challenging. Almost 10% of patients failed again and required another redo procedure [16, 34]. The best illustration of this problem was put forward by Smith et al. who reported a large series of 285 patients who underwent 307 reoperations for failed antireflux surgery . Two hundred and forty-one patients underwent 1 redo, 59 patients 2 redos, 6 patients 3 redos, and 1 patient 4 redos. Transdiaphragmatic wrap migration was the main indication for multiple redo surgeries. The mean interval between the first and second redo was 24 ± 33 months and 12 ± 7 months between the second and the third redo. The failure rate was 7.1% after the first redo and 17% after the third redo. The type of procedure performed successively was iterative fundoplication. Based on this experience, the authors did not consider redoing the primary operation after three prior attempts, but instead, they planned to perform a different operation.