Surgical Treatment of Zenker's Diverticulum

Zenker's diverticulum (ZD), also known as hypopharyngeal diverticulum, is an acquired protrusion through the weak area between inferior constrictor muscle and cricopharyngeus muscle. The disorder was first reported in 1767 by Ludlow [1]. It was not formally named until Zenker and Von Ziemssen [2] comprehensively and precisely described this condition in 1877. The incidence of ZD was estimated to be 1-2 per 100,000 patients/year and twice as common in males [3]. Over the past several decades, the pathophysiology of this entity has evolved significantly. The most popular pathogenesis theories are structural abnormalities [4], increased hypopharyngeal pressure, increased cricopharyngeal tone and gastroesophageal reflux [5]. Since the first successful resection of ZD by Wheeler [6] in 1886, different surgical techniques have been indicated for the management of ZD, including diverticulectomy, diverticulopexy, diverticular inversion (all with or without cricopharyngeal myotomy), and myotomy alone. Traditionally, these operations were carried out openly through a left cervical incision. In consideration of the fact that ZD is a disease of the elderly and some life-threatening complications are inevitable, minimally invasive techniques have gained some kind of popularity in recent years. After decades of refinements since the first report by Mosher [7] in 1917, endoscopic therapy has become increasingly reliable for the management of ZD. To date, the optimal treatment approach for ZD remains debatable. We therefore conducted this review in order to gain a profound understanding of the current trend and evidence in surgical management of ZD.

Medline and PubMed were searched to identify studies on surgical intervention of ZD. Zenker's diverticulum, pharyngeal pouch, hypopharyngeal diverticulum, and surgery were used as keywords for the search and cross-references were added and reviewed to complete the reference list. Studies published between January 1990 and March 2011 in English were examined by reading abstracts and full articles. Only case series with more than 20 subjects or comparative studies were considered eligible for inclusion. In addition, we excluded studies which failed to provide important information, such as complications and their respective incidences. Data was extracted from each report using a pre-designed form. The complications were restricted to major surgical complications, such as perioperative death, mediastinitis, wound infection, fistula or perforations, bleeding, pneumomediastinum, vocal cord paralysis, aspiration pneumonia, dental injury, esophageal mucosal tears, among others. Minor complications, such as short-term fever, pain, throat burn were not included. Patients with no symptoms or who were significantly improved following one single treatment session were considered as having successful symptom resolution. Recurrence was defined as an increase of the severity of symptoms to the preoperative level or to an even worse condition which required repeated treatment.

We identified 6,915 patients from 93 studies evaluating the effect of the surgical treatment for ZD, including 52 studies [8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59] (n = 3,336) on endoscopic surgeries, 22 [60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81] (n = 2,204) on open surgeries, and 19 [3,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99] (n = 1,375) on both approaches. The lack of objective measurements in long-term follow-up was a common shortcoming in most studies. There was no randomized controlled trial comparing one technique with another.

It is well established that the treatment choice of ZD should be surgical intervention. Conventionally, an open left cervical incision is usually performed under general anesthesia. The diverticulum is then freed and followed by cricopharyngeal myotomy. Further, the diverticulum is either resected (diverticulectomy), or more conservatively, suspended and fixed on the hypopharyngeal wall (diverticulopexy) or invaginated into esophagus itself (diverticular inversion). It was documented that open surgery can effectively resolve symptoms in 90-95% of the patients with a mortality rate of 0-2.3% and morbidity rate of 2.5-46% [100].

Our literature review yielded 2,826 patients from 41 studies assessing the outcomes of the open surgery for ZD. Of these, 7 different open surgical techniques were evaluated (table 1A). The surgical complications for open surgery are listed in table 2. Overall morbidity was 10.5% (297/2,826) and mortality was 0.6% (17/2,826). When reviewing the trend of the last 20 years, 1,059 out of 1,877 patients (56.4%) were treated with diverticulectomy plus cricopharyngeal myotomy (absence of information in some studies) which represents the most preferred open surgical approach for ZD. Yet, there was no sound evidence that favors one approach over another. The optimal treatment for ZD continues to be controversial. The ongoing debates about open surgery generally come down to the necessity and extent of myotomy, resection, suspension or inversion of diverticulum.

The first concern relates to cricopharyngeal myotomy. The rationale of myotomy is that upper esophageal sphincter (UES) dysfunction could be a pathogenesis factor for the diverticulum. The effects of cricopharyngeal myotomy include normalization of the opening size of the UES [101], reduction of resting UES pressures and the decrease of intrabolus pressures [102]. Some early small-sample case series reported good clinical outcomes when performing myotomy as the sole treatment [78,103,104]. Furthermore, Skinner and Zuckerbraun [105] reported their experience in dealing with recurrent ZD, who found that most patients with early recurrence did not have adequate cricopharyngeal myotomy, and all of the 8 patients who were treated with sufficient myotomy gained persistent symptom relief within a mean 53-month follow-up. A more recent study by Gutschow et al. [92] in 2002 compared open diverticulectomy alone (n = 34) with other open approaches along with myotomy (n = 67). The results showed no difference in terms of ‘good' or ‘excellent' symptomatic outcomes. Of interest, they noticed that the diverticulectomy without myotomy might predispose to the development of postoperative fistula and the recurrence, which may result from persistent high hypopharyngeal pressure. Nowadays, more and more surgeons believe that dividing the cricopharyngeus muscle is a crucial step in the treatment of ZD, and it has been a standard procedure in many major centers. In the 41 studies of open surgeries, only 2 studies did not perform myotomy [80,95], and 2,526 patients out of 2,826 treated with open surgery for ZD received either myotomy alone or along with other approaches, which reflects that most surgeons consider the dysfunction of UES as an etiologic factor for ZD. The necessity of the myotomy for ZD is still debatable however. The largest case series involving 888 patients from Mayo Clinic failed to show any improvement following the additional application of myotomy to diverticulectomy [79]. A retrospective nonrandomized study published in 2003 by Colombo-Benkmann et al. [74] included 79 ZD patients who underwent diverticulectomy alone or diverticulectomy plus myotomy when hypertrophic transverse fiber of cricopharyngeus muscle was clearly visible. No significant impact on the postoperative outcomes was observed after a 60-month follow-up. Some other authors also argue that considerable anatomical variations of the cricopharyngeal muscle exist and the cricopharyngeus muscle may not present in some ZD patients. Consequently, they advocated that myotomy should be performed selectively [74]. Another concern about myotomy was the risk of aspiration pneumonia when the protection barrier of UES is compromised by the myotomy [106]. From our literature review though, the postoperative aspiration pneumonia in open surgery is very rare (0.3%, 9/2,826). Besides, a prospective study enrolling 10 patients, using pH monitoring, revealed that cricopharyngeal myotomy did not increase the risk of esophagopharyngeal acid regurgitation [107].

In the absence of a definitive study, the effect of myotomy in ZD remains obscure. Large, well-designed randomized controlled studies with a sufficiently long-term follow-up and reliable functional measurements are needed to settle this issue in the future.

When conducting a cricopharyngeus myotomy, the extent of myotomy needs to be clarified. Anatomically, the closing muscles of UES are composed of inferior pharyngeal constrictor, cricopharyngeus muscle and cervical esophagus. Only the cricopharyngeus, the primary muscle which exerts the physiological function of UES [108], contracts and relaxes during all the physiological states [109]. Theoretically, in order to achieve complete myotomy, except for the cricopharyngeus muscle, the inferior pharyngeal constrictor and cervical esophagus should also be dealt with, but there is no general agreement on the exact length of the myotomy, ranging from 2 to 6 cm [60,62,110]. Comparing ZD patients with healthy controls, Lerut et al. [81] demonstrated that cricopharyngeal muscle and cervical esophageal striated muscle have the same pathological changes. Therefore, they recommended a long (4-5 cm) myotomy for better outcomes. Except for the commonly employed linear myotomy, Duranceau and colleagues [69,111] proposed a flap-resected myotomy which could prevent subsequent healing and reconnecting of the muscle fiber.

For dealing with the diverticulum sac itself, three options are mainly used: resection, suspension or inversion. The primary advantage of the diverticulopexy and diverticular inversion over diverticulectomy is leaving mucosa intact without breach. Logically, the incidence of leakage could decrease. Other merits include shorter hospitalization, earlier resumption of oral intake, and shorter antibiotic treatment [73]. A comparative study by Gutschow et al. [92] indicated that diverticulectomy plus myotomy involved a significantly longer postoperative stay and fasting food period. The fistula rate in the resection group and suspension group was 1/12 and 2/47, respectively. The symptomatic outcomes were not different between the two groups. A more recent study compared myotomy plus diverticulectomy in 14 patients with myotomy plus diverticulopexy in 36 patients, revealing that diverticulectomy required a longer hospital stay and fasting period. Four out of 36 patients with diverticulopexy and 2 out of 14 patients with diverticulectomy developed recurrence of dysphagia within 6 months, but all of them were spontaneously symptom-free within 1 year after operation [60].

As another alternative approach, diverticular inversion itself has not drawn much attention, even though it was shown in some earlier studies to be effective and to have fewer complications than excision [112,113]. Since 1990, there have been very few studies reporting diverticular inversion. The only one, reported by Morton and Bartley [72], compared 15 cases of diverticulectomy with 18 cases of invagination, and showed better results regarding complications, operation time, and length of hospital stay. Of note, their cases were free of recurrence after 56 months of follow-up.

Strong evidence in favor of one approach over others is still lacking, and it is almost impossible to obtain a conclusion from ‘head-to-head' comparisons among a variety of surgical approaches. When taking a look at the trend of open surgery in the last 20 years, myotomy plus diverticulectomy has gained wide popularity (table 1). Considering the pathogenesis of the ZD, it seems that resection of diverticulum and concurrent cricopharyngeus myotomy is a more logical strategy. It addresses not only anatomical abnormality but also dysfunction of UES. The favorable results of this approach can be seen in some large-sample series [75,88]. With the assistance of a stapler technique the risk of leakage after diverticulectomy has been reduced substantially, and the fistula rate in the studies of Jougon et al. [61] and Bonavina et al. [88] was 1% (1/92) and 1.7% (2/116), respectively. Besides, resection of diverticulum showed another advantage in the very small risk of carcinoma genesis from diverticulum [73,114].

The first successful endoscopic treatment of ZD was reported as early as 1917 by Mosher [7], who described esophagodiverticulotomy by sharply dividing the esophagodiverticular wall endoscopically. He later abandoned this technique because of the high risk of mediastinitis and death. It was not until 1960 when the endoscopic approach took a leap forward, that Dohlman and Mattsson [115], using electrocautery in endoscopic diverticulotomy, reported a very low risk of mediastinitis and only 7% recurrence in 100 patients. Since then, the endoscopic approach for ZD has evolved over time. The endoscopic approach can generally be accomplished by two means - a flexible or a rigid endoscope.

The procedures are carried out under general anesthesia and patients are placed supine with their neck overextended. A special diverticuloscope is used to visualize and expose the common septum separating the diverticular sac from the esophageal lumen. After transoral positioning the diverticuloscope in the esophagus, the common wall containing cricopharyngeus muscle fiber between the diverticular sac and esophagus lumen can be exposed with connection to a video system. Division of the septum can then be performed by one of the following techniques: electrocautery, CO₂ laser, KTP/532 laser, stapler, needle-knife, or Harmonic Ace [100].

From the literature research, four techniques were introduced in rigid endoscopic diverticulotomy during the period of 1990-2011, including electrocautery, CO₂ laser, stapler, and Harmonic Ace (table 1B).

The electrocautery technique, also known as Dohlman's technique, represents the classic method for endoscopic diverticulotomy. Our literature search yielded 9 studies involving 485 patients, with an overall complication rate of 7.8% (38/485) and mortality rate of 0.2% (1/485). The two most common complications were subcutaneous emphysema (14/485, 2.9%) and mediastinitis (10/484, 2.1%). The largest series about Dohlman's procedure for ZD was reported in 1994 by Van Overbeek [56], who performed Dohlman's technique in 328 patients and endoscopic diverticulotomy in 216 cases. The complication rate was 9.5% in the electrocoagulation group and 6.6% in the other. The ‘highly satisfied' rate was 90.6%, with a minimal follow-up period of 10 months. Using the same technique, Von Doersten and Byl [57] reported 18% of 40 patients developed complications and 92.5% reached symptom relief without recurrence in the averaged 42-month follow-up.

In the last 20 years, Dohlman's technique has been largely replaced by the CO₂ laser and stapler technique. The high energy, high focus beam, and less tissue trauma were the advantages over electrocautery [48]. It also provides a better visualization of the diverticular bridge and easier control of the operation [55]. Our literature search identified 1,060 patients from 19 studies who underwent CO₂ laser procedures for ZD. The overall complication rate was 9.3% and mortality was 0.2% (2/1,060). The common complications were subcutaneous emphysema (3%, 32/1,060), mediastinitis (1.3%,14/1,060), fistula (1.1%,12/1,060) and bleeding (1%,11/1,060). Hoffmann et al. [50] reported 119 patients with ZD undergoing endoscopic CO₂ laser treatment. The complication rate was 3.4% (1 mediastinitis, 1 fistula, and 2 dental injuries), and in 1-3 years of follow-up, 93% of the patients were completely symptom-free and 3.9% were suspected to experience recurrence. Van Overbeek [56] used the CO₂ laser technique in 216 patients and found 6.0% of the patients developed complications but without death. Combining with the 328 patients undergoing electrocoagulation treatment, the overall highly satisfied rate was 90.6% and fairly satisfied rate was 8.6%. He also noticed that patients with CO₂ laser experienced less pain and took food more readily during the first postoperative day. The favorable result of the CO₂ laser was also reported by Lippert et al. [97], Krespi et al. [41] and Kos et al. [48].

In the last 10 years, endoscopic stapling technique has become increasingly popular. It possesses a theoretical advantage over the other techniques. It can simultaneously cut and seal the wound edge, contributing to a lower incidence of perforation and bleeding. In addition, it is also free of thermal damage to the recurrent laryngeal nerve. We identified 44 studies reporting this technique and involving 1,800 patients. The overall complication rate was 7.1% (127/1,800) and mortality was 0.3% (5/1,800). The common complications were dental injuries (2.0%, 36/1,800), esophageal mucosal damages (1.6%, 29/1,728), and perforations (1.6%, 29/1,728) (table 3). The outcomes varied significantly in different studies. The largest series using the endoscopic stapling approach was reported by Bonavina et al. [88]. They performed endoscopic stapling in 181 cases with ZD, and there were only 2 cases of dental injury and 1 case of mucosal tear in their series (complication rate: 1.7%). They also reported conversions to open surgery in 8 cases, in whom 7 were due to poor exposure and 1 due to mucosal tear during the endoscopy. Of the patients undergoing the endoscopic stapling approach, 92% were symptom-free after a mean follow-up of 27 months. Another study by Chang et al. [15] also evaluated the endoscopic stapling technique in 150 ZD patients, and they reported 9 cases who were converted to open surgery because of poor exposure of diverticulum. The complication rate was 12.7%, but only 2.0% were considered as major complications (1 each of perforation, aspiration pneumonia, and transient vocal cord paralysis). In a mean follow-up of 32.2 months, 88% of the patients gained either an improved or complete relief of symptoms and 10.9% developed recurrence of symptoms.

Another technique using the so-called harmonic scalpel was recently introduced for endoscopic treatment of ZD. In a study that included 25 patients treated with an endoscopic harmonic scalpel, 1 developed aspiration pneumonia and 1 cervical subcutaneous emphysema after surgery. After an average follow-up of 11.3 months, only 1 patient suffered from recurrence [34]. In the other study, Sharp et al. [31] performed endoscopic treatment of ZD in 52 patients, where 20 patients were treated with a harmonic scalpel, 28 received a stapling technique and 2 underwent both techniques. Complications in 5 patients (17.9%) relating to the stapler technique and 1 (5%) to the harmonic scalpel were noted, but the diverticulum size in the harmonic group was smaller compared to the stapler group. Unfortunately, no long-term follow-up information is available for this study. More studies are needed to further assess this technique.

A main drawback of the rigid endoscopic diverticulotomy is that the exposure of diverticulum may be comprised by some anatomic reasons or small diverticulum, under which circumstances a conversion to open surgery might be required. The reported conversion rate could be as low as 0% [28,50] or as high as 30% [52]. Combining all the data, we found that 18 cases out of 1,060 patients (1.7%) with the CO₂ approach and 101 out of 1,800 (5.6%) patients with the stapling technique underwent conversion surgery. The most common reason for unsatisfying exposure was anatomic limitation (87/119), followed by intraoperative esophageal mucosal injury (12/119) and small/large diverticulum (12/119). Other relatively rare conditions included obesity, cancer, bleeding, abnormal thyroid gland and complex diverticulum. The reported anatomic factors preventing adequate exposure consist of prominent teeth, a short neck, a retrognathic mandible, retrognathia, large tongue, and rigid cervical kyphosis. A prospective study assessing the factors predicting endoscopic exposure of ZD showed that a short neck length, shorter hyomental distance and higher BMI were correlated significantly with the failure of endoscopic exposure [52].

Flexible endoscopic diverticulotomy was not reported until 1995 [53,54], which can be performed under conscious sedation without general anesthesia and neck extension - the two major advantages over the rigid endoscopic approach. Using a variety of accessories (nasogastric tube, hood, endoscopic cap, and overtube), the septum between the diverticulum and esophageal lumen can be successfully visualized and stabilized without overextension of the neck. The incision was then made thorough different techniques including needle-knife, hook-knife, argon plasma coagulation (APC), and monopolar forceps [116]. Some authors prefer multiple sessions in order to avoid perforation and mediastinitis. The mean number of treatment sessions varied from 1 to 3 in different studies [9,13], and usually the operation can be repeated easily at an interval of 1 day to several weeks [13,53,54]. From our literature search, there were 472 patients from 12 studies undergoing flexible endoscopic diverticulotomy (table 1B). The overall complication rate for flexible endoscopic diverticulotomy was 15% (71/472) and for mortality it was 0%. The common complications included cervical emphysema (27/472), perforation (19/472) and bleeding (15/472). The recurrence rate varied from 0% (0/28; 20-month follow-up) [17] to 35% (11/31; 26-month follow-up) [14]. Better understanding about the flexible endoscopic technique calls for more studies in the future.

Using the needle-knife technique, Hashiba et al. [35] performed flexible endoscopic diverticulotomy on 47 patients under sedation. Postoperative bleeding (2%) and subcutaneous emphysema (12.8%) were reported in their study. 96% of the patients experienced marked symptom relief after the first session, but most of the patients needed more than one session (mean 2.2 sessions). During their 1-day to 1-year follow-up period, 2 (4%) patients had recurrence of symptoms. Costamagna et al. [10] compared two different assistant techniques for flexible needle-knife diverticulotomy: the cap-assisted technique versus the diverticuloscope-assisted technique. The operation time in the cap group was significantly longer than in the diverticuloscope group, and the complication rate was 32% (9/28) in the cap group and 0% (0/11) in the diverticuloscope group. The recurrence rate was 29% in the cap group and 9% in the diverticuloscope group within a 3- to 60-month follow-up period. Recently, Repici et al. [17] compared the flexible needle-knife technique (n = 28) with the rigid stapling technique (n = 30) and found a similar length of hospital stay, symptom relief percentage and complication rate (1/28 vs. 2/30), however the procedure time in the rigid stapling group was significantly longer than that in the flexible needle-knife group (63 vs. 42 min). Only 1 patient with the endoscopic stapling procedure developed recurrence in the mean 20-month follow-up period. In 1995, Mulder et al. [53] were the first to describe flexible endoscopic diverticulostomy in 20 ZD patients using monopolar biopsy forceps. A good symptomatic response was seen after a mean of three treatment sessions and there were no major complications. During the 6- to 7-month follow-up, 17 patients remained asymptomatic and the other 3 died of unrelated causes. Mulder [24] later reported their experience of application of the APC technique in 125 cases. Symptom improvement was achieved in all 125 patients after a mean 1.8 treatment sessions, and the overall complication rate was 19.2% (24/125).

To date, there is no strong evidence to conclude whether a procedure is superior over another, and the results vary significantly in different studies because of varied inclusion criteria, sample size, length of follow-up period and the percentage lost to follow-up. A summary of the comparative studies is shown in table 4. When combining all the results reported from 1990 to 2011, the overall morbidity and mortality rate for the endoscopic approach is 8.7 and 0.2%, respectively, whereas it is 10.5 and 0.6% for open surgery (tables 2, 3). An extensive review by Chang et al. [15] in 2003 compared publications about endoscopic or open surgery for ZD from 1990 to 2002. They also found the endoscopic stapling technique had a lower major complication rate and mortality rate than open surgery (2.6 vs. 11.8% and 0.3 vs. 1.6%, respectively). It is reasonable that endoscopic treatment has less risk of recurrent nerve injury and wound infection because of the endoluminal approach. Application of the endoscopic technique also has reduced the fistula rate from 3.3% in open surgery to 0.6%. On the contrary, the endoscopic approach has increased the risk for esophageal mucosal damage, intraoperative bleeding, and dental injuries, accounting for a higher risk of conversion during the rigid endoscopic procedure. According to the comparative studies, most studies confirmed the endoscopic approach had a lower complication rate. However, it is notable that the complication rate varied significantly across studies, ranging from 0 to 46%, which reflected the marked inter-study heterogeneity. The different sample size, criteria for patient selection, surgical techniques and their retrospective study design make it difficult to draw a very sound conclusion.

As for the convalescent period, some studies reported that endoscopic approaches needed less anesthetic and operative time, allowed earlier resumption of oral intake, and required a shorter hospital stay [86,98,99]. For flexible endoscopic approaches, the operation can be performed in outpatient settings without general anesthesia, though most patients need more than one session until symptoms resolve [35,53,54]. Even some patients undergoing the rigid endoscopic procedure can also be safely managed on an outpatient basis [30]. Besides, the hospital costs for endoscopic surgery have proven to be significantly lower than open surgery [91]. It is notable that most of the studies were historical series; these results need to be further verified in randomized controlled trials.

Regarding the long-term outcomes, there is no consistent result for the endoscopic approach. Significant heterogeneity exists among studies. The success rate ranged from 63 to 100% after a single treatment session [14,36] and the recurrence rate varied from 0 to 35% [14,18,26] because of the different follow-up period and different sample size. Since the endoscopic approach for ZD is a relatively new technique, the lack of long-term follow-up is a common drawback for most studies, and the recurrence has not been uniformly defined (i.e. defined as the requirement of more than one treatment session in some studies), rendering it difficult to compare the technique with open surgery. Besides, as most of the studies relied on the subjective methods only to assess treatment effects, it is hard to obtain strong evidence from them.

The treatment for ZD has evolved significantly in recent years and the optimal treatment modality for ZD still needs to be clarified. Both open and endoscopic approaches can be accomplished with many techniques, each of which has its own advantages and disadvantages (table 5). Several factors should be taken into account when determining the treatment option, e.g. doctor's preference, patients' will and general condition, and access to facility. Open surgery, as a well-established approach, allows for the direct operation on the diverticulum pouch and the complete conduction of cricopharyngeal myotomy. With the advance of perioperative management and new techniques, diverticulectomy with myotomy has become quite safe and its long-term outcomes are also satisfactory and durable. It is also technically feasible for all sizes of diverticula. Besides, it provides specimens for pathological assessment and eliminates the risk of cancer genesis from the pouch. The open approach however needs general anesthesia, which is a contraindication for some extremely debilitated patients. An alternative procedure, endoscopic diverticulotomy, was introduced as a minimally invasive approach. It may offer advantages as listed in table 5. The rigid endoscopic stapling technique ensures simultaneous dissection and sealing, reducing the risk of bleeding and mediastinitis. The endoscopic approach could however be technically difficult in some patients whose diverticular septum cannot be well exposed, such as small pouch, a stiff neck, poor mouth opening, prominent teeth, a short neck, a retrognathic mandible, retrognathia, large tongue, and rigid cervical kyphosis, etc. Flexible endoscopic treatment is a promising technique which can be conducted without the need of general anesthesia and hyperextension of the neck. However, adequate stabilization of the septum and protection of the anterior esophageal and posterior diverticular walls during dissection are more challenging [10]. This more recently developed technique needs more studies to verify its safety and effectiveness.

With special concern to small and larger diverticula, it seems the endoscopic approach has its limitations. In small diverticulum the exposure of diverticulum could be problematic [19,21,22,31] and has a higher complication rate [59]. Even if successfully visualized, the endoscopic approach could result in incomplete myotomy since the septum of the small diverticulum contains only limited cricopharyngeus muscle. This was confirmed by Bonavina et al. [88]. Gutschow et al. [92] also indicated that the percentage of patients with diverticula <3 cm who achieved an excellent or good symptomatic outcome was higher in patients receiving open treatment than the endoscopic approach. Similar results were reported by Rizzetto et al. [87]. For a very large diverticulum, to achieve a sufficient myotomy, a longer incision of the septum is always needed, in which case the sutureless techniques like laser, APC and diathermy may encounter a high chance of bleeding. The stapling technique therefore seems more suitable in this situation, but application of multiple staple rows as needed in a large diverticulum was associated with higher leak rate as well [45]. Counter et al. [18] also considered a very large diverticulum as a contraindication for the endoscopic approach where the insertion of the stapler with a sufficient distance down the endoscope could be a limiting factor. Ozgursoy and Salassa [46 ]concur by pointing out that endoscopic diverticulotomy would create an axially adynamic esophageal segment, thus they advocated a large diverticulum >7 cm should be treated by the open approach rather than the endoscopic procedure to limit the adynamic segment to the upper third (7 cm) of the esophagus, and, by doing so, the function of the lower esophagus would not be compromised.

It is also important to point out that the comparisons made in this review were based on retrospective observations and case-control studies, which limits the actual level of evidence for this review. Randomized controlled studies with long-term follow-up results are required to draw the definitive conclusion regarding the effect of each approach.

The surgical treatment options for ZD include open surgery and flexible endoscopic and rigid endoscopic therapy. Based on current evidence, traditional open surgery is suitable for all kinds of diverticula, providing satisfactory long-term outcomes and acceptable complication rates. However, it needs general anesthesia and more invasive procedures. Rigid endoscopic treatment can be done under general anesthesia and hyperextension of the neck. It might be technically difficult where the diverticular septum cannot be well exposed. Flexible endoscopic therapy can be conducted without general anesthesia or neck hyperextension, however it is only suitable for selected patients. Each treatment option has its pros and cons, but it is important to perform individualized therapy for each patient. Minimally invasive endoscopic therapy should be considered for debilitated patents with a middle-size diverticulum, and open surgery would be preferred when difficulty of diverticulum exposure is predicted. However, the poor quality of current evidence renders it difficult to establish a sound conclusion for the optimal treatment of ZD. More randomized controlled trials with long-term follow-up results are necessary to draw a firm conclusion.