Introduction
Zenker's diverticulum (ZD) is a saccular formation that protrudes into the posterior
wall of the pharyngo-esophageal junction in an anatomically weak area between the
inferior pharyngeal constrictor and cricopharyngeus muscles.
Conventional treatment of symptomatic ZD by open surgery (OS) to extirpate the ZD
by cricopharyngeal myotomy has an overall postoperative morbidity rate of 11 % [1], including mediastinitis (3 %), vocal cord paralysis, and pharyngocutaneous fistula,
among others, and the mortality rate is up to 3 % [2]. Postsurgical recurrence rates are between 3 % and 19 %, rising to 15 % to 35 %
when there is no cricopharyngeal myotomy [3].
In the mid-1960 s, rigid endoscopy (RE), described and abandoned in previous decades
due to its high mortality, was reintroduced using rigid diverticuloscopes. The greatest
limitations of RE are the difficult maneuverability in patients with a small ZD and
the need for maximum cervical hyperextension, with a large number of failed cases
[4]. The recurrence rate is 10 % to 12 % and the complication rate 7 % [5].
Since the first diverticulotomies with flexible endoscopy (FE) of the esophageal-diverticular
septum (EDS) were described in 1995 [6], various endoscopic techniques have been developed with good results and less morbidity
and mortality than surgery [7]. Different scalpels, monopolar coagulation forceps, laser and argon plasma have
been used, together with use of transparent caps and flexible diverticuloscopes. In
recent years, Ligasure (LS1500, Covidien, Medtronic, Minneapolis, Minnesota, United
States), a system imported from laparoscopic surgery, has shown very good results
in small series [8]
[9]
[10]. We present the largest series of patients with ZD treated by endoscopic diverticulotomy
with Ligasure (EDVL) with a long follow-up.
Patients and methods
We conducted 79 EDVLs between June 2010 and March 2018 in 69 patients with ZD in the
Digestive Service, Río Hortega University Hospital, a tertiary health center. We prospectively
collected sociodemographic, surgical, complications and follow-up data.
EDVLs were performed by three endoscopists on an outpatient basis under deep sedation
with propofol controlled by the endoscopist, prophylactic antibiotic therapy and insufflation
with CO2.
Ligasure is a bipolar device in the shape of a rotatable clamp that releases high
current/low-voltage energy to the tissue trapped between its blades, producing tissue
dissection and sealing of blood vessels ≤ 7 mm, which ensures hemostasis prior to
cutting the tissue.
The ZD was identified by gastroscopy, and a guidewire (Jagwire HPG, Boston Scientific,
Natick, Massachusetts, United States) was placed in the gastric cavity before the
gastroscope was retired. The external end of the guidewire was inserted through a
hole made through an 18G catheter at the end of the long blade of the plastic diverticuloscope
(PD) (Cook Medical, Limerick, Ireland) ([Fig. 1]). Subsequently, the PD was introduced into the oral cavity and a gastroscope (GIFQ
180, Olympus Optical Co. Tokyo, Japan) was placed inside to provide endoscopic vision.
The PD was advanced until the long blade fit into the esophageal lumen and the short
blade into the diverticulum, allowing the EDS to be isolated between the two blades
([Fig. 2]), and then the gastroscope was removed and the Ligasure (5 mm × 35 mm), connected
to the generator, inserted within the PD in parallel with a 4.5-mm gastroscope (GIF-XP160N
Olympus) to guide its advance and permit the ZD to be measured.
Fig. 1 The external end of the guidewire inserted through a hole made through an 18G catheter
at the end of the long blade of the plastic diverticuloscope (Cook Medical, Limerick,
Ireland).
Fig. 2 The esophageal-diverticular septum was isolated between the two blades after gentle
advance of the diverticuloscope until the long blade fit into the esophageal lumen
and the short blade into the diverticulum.
When the Ligasure reached the middle zone of the EDS, the blades were opened and closed
on the tissue ([Fig. 3]), the clotting-tissue sealing function was activated and the cut was made ([Fig. 4]). This twin maneuver was repeated as often as necessary, depending on the size of
the septum, to leave a residual wall < 5 mm to avoid perforations. A clip (Resolution
Clip, Boston Scientific, Natick Massachusetts, United States) could be inserted preventively
at the discretion of the endoscopist ([Fig. 5], [Video 1]). Post-procedure, patients were monitored and discharged to home in about 60 minutes,
with a report on the intervention and written recommendations including initiating
liquid intake at 8 hours post-intervention, soft foods in the following 24 hours and
a progressively normal intake during the first week, and instructions to visit the
emergency room if necessary. Patients received a telephone call from the endoscopy
team at 24 hours, 1 week, 1 month and 6 months and, where possible, clinical or telephone
follow-up was maintained for up to 4 years. An objective dysphagia scale could not
be used because a large part of the follow-up was by phone and many patients were
elderly and did not understand the numerical scale we wished to employ.
Fig. 3 When the Ligasure reached the middle zone of the EDS, the blades were opened and
closed on the tissue.
Fig. 4 Appearance of the septum after cutting with Ligasure.
Fig. 5 After cutting, a clip (Resolution Clip, Boston Scientific, Natick Massachusetts,
United StatesA) can be inserted, depending on the preference of the endoscopist, to
avoid perforations.
Video 1 Procedure of Zenker diverticulotomy.
The statistical analysis was made using the Stata program (StataCorp 2013, Stata Statistical
Software Release 13, StataCorp LP, College Station, Texas, United States). Categorical
variables were described as percentages. Continuous variables were presented as mean
and standard deviation if the distribution was normal and as median, interquartile
range (IQR) and/or range when the distribution was non-normal.
Results
Between June 2010 and March 2018, 72 patients with symptomatic ZD were referred to
our Endoscopy Unit for EDVL. There were 45 men and 24 women, with a mean age of 73.4
years (SD: 11.1). One patient had previously undergone surgery for ZD. Patients reported
symptoms for a mean of 1.5 years (IRQ: 1 – 2) before referral for treatment. The most
frequent symptom was dysphagia in 73.9 % of the cases, which was very severe, preventing
fluid intake, in 11 cases, severe for solid foods in 40 and mild or intermittent in
18. Forty-five patients had food regurgitation from the diverticulum and there were
19 episodes of choking, 16 halitosis, 9 nocturnal episodes of cough or regurgitation,
8 cases with weight loss, and aspiration pneumonia in 10 cases ([Table 1]).
Table 1
Patient data n (%).
|
Total number of patients
|
n = 69
|
|
Female
|
24 (34.8 %)
|
|
Male
|
45 (65.2 %)
|
|
Age, years, mean (range)
|
73.4 (46 – 93)
|
|
Symptom duration before treatment (years)
|
1.5 (1 – 2)
|
|
Symptoms
Dysphagia:
-
Very severe (fluids)
-
Severe (solid foods)
-
Mild or intermittent
Regurgitation
Choking
Halitosis
Nocturnal cough or regurgitation
Weight loss
Aspiration pneumonia
|
11 (15.9 %)
40 (58 %)
18 (26.1 %)
45 (65.2 %)
19 (27.5 %)
16 (23.2 %)
9 (13 % )
8 (11.6 %)
10 (14.5 %)
|
Technical success was defined as EDS section using EDVL. Clinical success was defined
as patient-reported disappearance of symptoms or significant symptomatic improvement
1 week after EDVL. Recurrence was defined as the reappearance of daily or weekly symptoms
in asymptomatic patients or occasional symptomatology after the intervention.
In three ZD ≤ 1.5 cm, EDVL was not possible because the PD could not be placed. There
was no failure due to inability to manipulate the Ligasure. EDVL was performed in
69/72 patients, with a technical success of 95.83 %. Mean ZD size was 2.8 cm (SD 1.1 cm)
and median number of cuts was two (IQR: 1 – 2), depending on the length of the EDS. In
49 cases (71 %) a clip was placed before retiring the overtube. The mean time of EDVL,
from introduction of the gastroscope to extraction of the PD from the oral cavity,
was 22 minutes (11 – 48).
Ten patients continued with or presented dysphagia after EDVL and underwent a second
EDVL. Therefore, duringn the study period, we performed 79 EDVLs in 69 patients.
Immediate complications were two case (2.5 %) of mild, self-limited intraprocedural
bleeding after cutting the septum.
Most early complications were mild: odynophagia (21.5 %) and cervical pain (5 %) that
resolved in < 72 hours with first-level analgesia. However, two males (2.5 %) presented
iatrogenic esophageal microperforation in the first hours after treatment, requiring
hospital admission for 8 and 12 days, respectively. These patients reported thoracic
or interscapular pain and low-grade fever with leukocytosis, and small peridiverticular
air bubbles were evident on computed tomography, despite placement of a clip in both
cases. Neither required surgical or endoscopic procedures, and evolved well with conservative
treatment that included fasting and antibiotic therapy. One patient, with a 2-cm ZD,
who received a single cut, had ankylosing spondylitis with significant deformity and
cervical-thoracic rigidity that made maneuvers difficult during the intervention.
The other patient, with a 3.5-cm ZD, was treated with two cuts, without abnormalities
being detected during the intervention.
The only late complication (1.3 %) was a case of iatrogenic digestive hemorrhage,
10 days after EDVL, without hemodynamic repercussions, treated endoscopically with
a clip ([Table 2]).
Table 2
Technical results and complications.
|
Successful septotomy, number of cases
|
69/72 (95.83 %)
|
|
Procedure time, minutes, mean (range)
|
22 (11 – 48)
|
|
Septum size in cm, mean (range)
|
2.8 (1 – 5)
|
|
Number of cuts, mean (range)
|
2 (1 – 4)
|
|
Used clips (number of cases )
|
49 (71 %)
|
|
Immediate complications (intraprocedure)
|
2 (2.5 %)
|
|
Early complications (24 – 48 h after procedure)
|
21 (30 %)
15 (21.5 %)
4 (5 %)
2 (2.5 %)
|
|
Late complications (> 72 h after procedure)
Hemorrhage
|
1 (1.3 %)
|
The 69 patients or their relatives were contacted by telephone the day after EDVL
and subsequently, with a progressive loss of cases: 61 (85.5 %) at 1 week, 51 (73.9 %)
at 1 month, 36 (52.2 %) at 6 months and 56 (81.2 %) at 34.6 months on average (24 – 64
months).
At 8 hours, 53/69 patients (77.8 %) had started a liquid diet, and the 61 located
at 1 week, except for the two with microperforation, tolerated a solid or semi-solid
diet, indicating a tolerance to more solid food than before the intervention, with
disappearance or improvement of other symptoms such as regurgitation or cough. Therefore,
the clinical success rate was 96.7 % (59/61).
One month after EDVL, 44/51 (86.3 %) patients had no dysphagia and 7/51 (13.7 %) reported
difficulty in eating some solids (especially fibrous meats or bread).
At 6 months, 33 of 36 (91.6 %) reported remaining asymptomatic and 3 of 36 (8.3 %)
reported dysphagia for solid food, with one patient with little initial improvement
and very frequent dysphagia. This patient was considered as “insufficient treatment”
and underwent a second EDVL with a very good evolution.
Of the 56 patients (81.2 %) followed for a mean of 34.6 months, 47 (84 %) had no dysphagia
(including the patient with a second early EDVL), four had dysphagia only for fibrous
foods, three dysphagia for semisolids and regurgitation, and two severe dysphagia.
The nine patients who reported recurrence of dysphagia, after initially having been
asymptomatic or with significant symptom improvement, underwent a second EDVL after
a median of 12 months (IQR: 11.5 – 17, range 10 – 64) after the first procedure, with
clinical success in seven cases. No improvement in dysphagia was achieved in two elderly
women with senile dementia who, although presenting initial improvement, had progressive
worsening which did not respond to enlargement of the small residual EDS cut. These
patients presented severely diminished esophageal motility with a “corkscrew appearance”
and food remains along the esophagus. This was compatible with an age-related reduction
in esophageal motility but, due to the advanced age and inability to collaborate,
no esophageal manometry study was made. The first patient received placement of a
percutaneous endoscopic gastrostomy, which was rejected by the second patientʼs family.
Without considering the patient with a bad response to the first treatment and the
two elderly patients with progressive early worsening and no improvement after a second
treatment, primary success with recurrence, i. e. asymptomatic cases or clear improvement
after the first EDVL with reappearance of symptoms months or years later without other
reasons for dysphagia, occurred in 7 of 69 patients (10.4 %).
In patients with recurrence, the mean size of the ZD before the first intervention
was 2.8 cm (range 1 – 5, SD 1.1 cm) and the median number of cuts was 2 (IQR: 1 – 2),
with a clip placed in 60 % after the EDS section. The mean size of the ZD after the
first EDVL in the seven patients where information was available was 1.5 (range 0.9 – 2.5)
One year after the second EDVL, three patients had died, two were not contacted, three
were asymptomatic, one reported occasional dysphagia for some hard foods and one patient
still had severe dysphagia.
Discussion
Our results in 69 outpatients with ZD with FE show that Ligasure is a simple and rapid
technique with few complications and lasting efficiency.
We used a PD that permits a better vision and fixation of the EDS with more precise
control of the length of the cut and which was also very useful in directing the placement
of a clip and reducing risk of aspiration during the procedure. In three patients
with ZD < 1.5 cm we were unable to fit the EDS between the two valves of the PD because
the short blade, which measures 2.5 cm, was longer than the diverticulum. It might
be useful to design PDs with shorter blades or use a cap in a very small DZ, which
might be the only limitation of this method. However, a study comparing a cap with
PD found a similar technical success rate but significantly greater clinical success
in patients in whom PD was used, with a greater number of adverse events (AEs) and
a longer intervention time with use of the cap [11]. Once the PD was placed, there were no specific difficulties in introducing the
Ligasure or in sectioning the EDS, regardless of its thickness, as has been reported
with use of some staplers [12]. The procedure required a mean of 22 minutes, among the shortest reported times
for EF [12]
[13].
Our series had a technical success rate of 95.8 %, with 96.7 % clinical success and
84 % asymptomatic patients at 34.6 months (24 – 64 months). The recurrence rate was
10.4 % at 20.3 months (10 – 64 months) with a good response to a second EDVL at 12
months (IQR: 11.5 – 17). A second EDVL achieved disappearance or frank reduction in
symptoms in a male with poor initial improvement after EDVL and seven patients with
symptomatic recurrence. Nevertheless, two elderly women with senile dementia and age-related
severely decreased esophageal peristalsis did not improve after a second procedure
on a small residual ZD. This shows that a second, rescue EDVL is very useful in patients
with incomplete treatment or recurrence, provided they are well selected, which probably
excludes those with severe comorbidities and other disorders of swallowing or esophageal
motility, as recommended by other authors [14]. In our series, there was only one postsurgical recurrence, which was managed successfully
without complications. We believe that EDVL is suitable in this context as are other
endoscopic techniques [2]
[15]. Unfortunately, we were not able to measure improvement in symptoms using an objective
scale and had to rely on subjective assessment by the patients.
A review [1] of FE techniques, not including Ligasure found a global technical success rate of
99.4 %, a clinical success rate of 87.9 % (after 1 to 3 treatment sessions), a failure
rate due to persistence of symptoms of 10 % and a recurrence rate of 13.5 %, with
a mean/median follow-up ranging from 7 to 43 months.
There are few reports on Ligasure for treatment of ZD, with the first dating from
2014 [8]
[16]. Two small series [9]
[10] (n = 5 and 8) reported no recurrences after a mean follow-up of about 21 months,
including a partial clinical improvement efficiently rescued with a second EDVL at
3 months. The procedures were performed, as in our series, under deep sedation with
propofol, but patients were hospitalized for ≥ 24 hours while our patients were discharged
after the procedure.
A study [16] of 15 patients with ZD treated with Ligasure with RE under general anesthesia reported
93 % of asymptomatic patients at 2 to 8 weeks post-procedure with 13.3 % recurrences
after a mean follow-up of 9 months and found that Ligasure was much easier to manipulate
than surgical staplers. In 2017, the same group reported a 22-month follow-up of 21
patients with 57 % free from dysphagia and 38 % asymptomatic [17].
Symptomatic recurrence may be due to incomplete section of the cricopharyngeus muscle
or subsequent scarring. Risk factors associated with recurrence are a ZD ≥ 5 cm, an
EDS cut ≤ 25 mm and a post-treatment ZD ≥ 1 cm [18]. In our cases with recurrence we found no differences in ZD size or number of cuts
made compared with patients without recurrence, although the mean size of the septum
before the second treatment in the seven patients where this information was recorded
was 1.5 cm (range 0.9 – 2.5). As we did not make endoscopic controls, we cannot compare
this measurement in asymptomatic patients. Leaving the smallest septum without causing
a perforation seems to be crucial to obtaining good symptomatic control and reducing
recurrence. Variants of the usual section technique with greater resection of the
ESD [19]
[20]
[21] and its resection by submucosal tunneling [22] have recently been described, but prospective data are needed to determine the reproducibility
and long-term effectiveness of this technique.
The most frequent complications secondary to FE treatment for ZD [15] are hemorrhage (6.6 %) either intraprocedural or post-intervention, and perforation/emphysema
(5.3 %) requiring surgery in 0.9 %, with no mortality being described.
Our intraprocedural bleeding rate was 2.5 %, much lower than that described with other
endoscopic devices. An intraprocedural bleeding rate with different types of needle-knife
of 33 % has been described, requiring use of argon plasma or monopolar electrocoagulation
for resolution and sometimes preventing successful therapy due to lack of visibility
[14]. Our few episodes of mild, self-limited bleeding may probably be considered intraprocedural
events without technical or clinical repercussions rather than AEs and this is one
of the main strengths of use of Ligasure, in line with some stapler devices [12] and clearly safer, with respect to bleeding, than other scalpels. We had one case
of slight delayed bleeding 10 days post-intervention, which was resolved endoscopically.
Our perforation rate of 2.53 % is lower than the 5.3 % described with other FE techniques
without tissue sealing [15], while in some series [23]
[24] with the needle-knife it is higher than 20 %, occasionally requiring surgical treatment
or the insertion of endoscopic stents. Our two microperforations were resolved conservatively.
Placement of clips, initially described by Tang [25], which was used in 71 % of our cases, may help prevent perforations [14]
[18] and therefore is routinely recommended by some authors, although cases of microperforation
have been described despite sealing with clips [26], as occurred in our two cases. Other authors [27] prefer not to use clips to avoid scar formation but it is unclear whether this is
related to worse clinical outcomes.
Conclusion
In conclusion, our results reinforce the few data available on treatment of ZD with
FE and Ligasure, demonstrating that it is a simple, effective, and very safe ambulatory
technique that also permits resolution of the majority of recurrences in well-selected
patients. Although consensus on the best endoscopic option for treatment of ZD in
each patient would be ideal, after direct comparison between the different devices,
we can currently recommend Ligasure as one of the simplest and safest.
