Advanced therapy in thoracic surgery - part 10 docx

Substernal placement may be required in cases of severe scarring of the posterior mediastinum as occurs after gastric conduit leak or prior esophageal resection.. Wong points out that th

gastroepiploic artery Using double clips, ties, or

ultra-sonic scalpel, short gastric vessels are taken, larger vessels

being clamped and tied with 3-0 silk sutures Elevation of

the spleen with one or two moist lap pads may aid in

exposure of the short gastric vessels Dissection

proxi-mally on the greater curvature extends up to the hiatus.

Following this, the dissection is carried distally, taking

great caution in staying at least 2 cm lateral and inferior

to the gastroepiploic artery.

The greater curvature is lifted anteriorly and

adhe-sions between the pancreas and stomach are divided

using electrocautery The left gastric artery is located, and

celiac nodal tissue is swept onto the specimen Using a

30 mm vascular stapler the origin of the left gastric artery

is clamped After verification of an excellent pulse in the

gastroepiploic arcade, the stapler is fired (Figure 41-5).

The remaining portion of the gastrohepatic ligament is

divided using a combination of cautery and staples.

Using a combination of careful electrocautery and

blunt dissection, a Kocher maneuver is performed

mobi-lizing the duodenum to the midline A pyloromyotomy

or Heineke-Mikulicz pyloroplasty is performed If a

pyloroplasty is performed it is closed with interrupted

3-0 silk sutures, in one layer, carefully incorporating

mucosa and muscular wall.

A 6 cm left neck incision is made along the anterior

border of the sternocleidomastoid muscle starting at the

sternal notch Dissection continues medial to the carotid

sheath but lateral to the strap muscles and thyroid gland.

The middle thyroid vein and omohyoid muscles are

divided Blunt dissection is directed toward the spine,

and the Penrose drain, which had been placed around the

cervical esophagus from the chest, is grasped The

esoph-agus is gently and bluntly mobilized The nasogastric

tube is withdrawn proximal to a point of division of the esophagus and a linear cutter 75 mm stapler is used to divide the cervical esophagus A #2 silk suture is secured

to the distal end of the divided esophagus prior to its division The esophagus is delivered into the abdomen with its attached silk suture The other end of the silk suture remains in the neck to aid in passage of the gastric conduit into the neck.

A gastric tube is created by resecting the phageal junction and lesser curve of the stomach with a series of linear 75 mm staplers (Figure 41-6) The tube should be kept relatively narrow to aid in gastric empty- ing The line of division is directed to a point on the lesser curvature just proximal to the crow’s foot of veins The right gastric artery and associated tissue are ligated and divided at this point on the lesser curve in order to prevent “bow-stringing” of the stomach and aid in converting the gastric tube from a curved structure into a straight conduit

gastroeso-A final check for hemostasis in the abdomen is made prior to pulling the conduit into the neck The gastric

494 / Advanced Therapy in Thoracic Surgery

FIGURE 41-4 The second Penrose drain is knotted and left in the

abdomen for retrieval during the abdominal phase of the operation

FIGURE 41-5 After verification of an excellent pulse in the

gastro-epiploic arcade, the stapler is fired

FIGURE 41-6 A gastric tube is created by resecting the

gastro-esophageal junction and lesser curve of the stomach

conduit must be mobilized to the neck with minimal

trauma and without twisting One reliable atraumatic

method is pulling the conduit up in a plastic endoscopic

camera bag The silk tie, which had been brought into the

abdomen through the mediastinum from the neck, is tied

to the valved end of a 30 mL Foley catheter (Figure 41-7).

The balloon is secured in one end of the bag The entire

conduit is placed in the bag, and the excess bag is trimmed.

The valved end of the Foley catheter is brought through to

the neck, suction is applied to the catheter and bag, and

the gastric conduit is delivered into the neck The assistant

must guide and push the conduit into the mediastinum,

ensuring that no rotation occurs The bag is then cut away

from the conduit in the neck The pylorus should sit at the

hiatus.

The anastomosis may be created in a side-to-side

functional end-to-end using a linear 75 mm stapler

followed by an endoscopic 30 mm stapler Prior to

closure of the anterior wall of the anastomosis using a

TA-30 stapler, the nasogastric tube is delivered into the

neoesophagus with its tip at the hiatus A hand-sewn

anastomosis can be created by using interrupted 3-0 silk

sutures in a single layer Care must be taken to

incorpo-rate full thickness bites Either a Penrose or Jackson-Pratt

drain is placed posterior to the anastomosis It is usually

left in for 5 days The platysma is closed using a running

2-0 Vicryl, and the skin is closed with staples A j-tube is

inserted at a point 40 cm distal to the ligament of Treitz if

one was not placed preoperatively The abdominal fascia

is closed with a running #2 synthetic monofilament, and

the abdominal skin is closed using staples.

Transhiatal Esophagectomy

The abdominal phase of the procedure is carried out in a

manner identical to that described above Placement of

an upper hand retractor greatly aids in visualization The

use of malleable handheld metal retractors at various

points in the operation is also useful The phrenic attachments are divided using cautery, and the distal esophagus is encircled with a Penrose drain The esophageal hiatus is dilated to allow entry of the sur- geon’s hand A rim of diaphragmatic muscle is incorpo- rated in cases of lower esophageal tumors Using the Penrose drain as a handle, the esophagus in the lower chest is mobilized bluntly Larger arterial branches dir- ectly from the aorta are clipped Keeping the fingertips against the esophagus, the prevertebral plane is devel- oped into the upper chest A key aspect of blunt esopha- geal dissection is remaining very close to the esophagus, with the palmar aspect of the fingers directly against the esophagus This accomplishes two objectives The first is the avoidance of adjacent structures such as membra- nous trachea and the azygous vein and its branches The second is to disrupt only smaller branches of the esophageal arterial supply after they have branched approximately l cm off the esophagus.

esophago-At this point a lower cervical incision is made as described in the previous section on tri-incisional esophagectomy Care must be taken in getting around the cervical esophagus not to injure the recurrent laryngeal nerve The right recurrent laryngeal nerve lies some distance from the lower cervical esophagus and should be out of harm’s way if dissection is kept immediately on the esophagus and cautery is avoided Metal self-retaining retractors can also cause injury to the recurrent nerve in the tracheoesophageal groove and should not be used A Penrose drain is used to encircle the cervical esophagus, and blunt dissection is used into the upper chest with two fingertips against the esophagus Care must be taken not to injure the membranous trachea Dissection of the posterior esophagus along the prevertebral fascia is con- tinued using a sponge stick to connect to the surgeon’s other hand inserted through the hiatus (Figure 41-8) Dissection is then carried out anterior to the esophagus

in similar fashion, reversing the direction of the hand and keeping the fingertips against the esophagus (Figure 41-9) Extreme care must be taken as the surgeon approaches the carina A combination of anterior dissec- tion from above and below are required until the fingers and sponge stick inserted from above meet The remain- der of circumferential dissection of the upper esophagus

is performed via the neck.

That portion of dissection lateral to the esophagus that cannot be directly visualized is performed bluntly The surgeon’s right hand is advanced through the hiatus anteriorly until it reaches that portion of the upper esophagus that has been circumferentially mobilized The first and second fingers surround the esophagus and press against the spine, and the lateral attachments are avulsed as the hand is drawn back into the abdomen

FIGURE 41-7 The silk tie is tied to the valved end of a 30 mL Foley

catheter

less well to mechanical dilatation The hand-sewn

anas-tomosis may be one or two layers Whatever the method

of anastomosis, the incidence of leakage must be close to

zero, as intrathoracic leakage carries a near 50%

mortal-ity Lack of adherence to surgical principles of lack of

tension on the anastomosis, preservation of blood

sup-ply, meticulous suture placement and tying, and gentle

handling of tissues will become apparent in an increased

leak and mortality rate.

Left Thoracoabdominal Approach

The left thoracoabdominal approach is most useful for

benign strictures involving the distal esophagus or for

proximal gastric cancers where a small margin of normal

esophagus is needed It has limited utility in management

of distal esophageal cancers, because of the need for

adequate margins It is more difficult, though possible, to

resect the esophagus high in the left chest than in the

right because of the heart and aortic arch The patient is

placed in the right lateral decubitus position, and the

abdomen, chest, and, if total esophagectomy is

contem-plated, left neck are prepped and draped in the field A

left anterolateral thoracotomy is made and the chest is

typically entered in the sixth or seventh interspace for a

distal esophagectomy If exposure of the upper esophagus

is needed, then the incision can be extended posterior to

the tip of the scapula midway between the scapula and

the spine, and a separate entrance into the chest in the

fourth interspace may be made The distal esophagus is

encircled with a Penrose drain and dissected free as

described in the previous sections.

Although the abdomen can be explored through the

chest across the diaphragm, better exposure can be

obtained by extending the thoracotomy incision across the costal margin and obliquely into the abdominal wall.

In both instances the diaphragm is taken down radially

2 cm from the chest wall so that adequate tissue remains for diaphragmatic closure Mobilization of the stomach and a Kocher maneuver can easily be performed with this exposure Exposure of the short gastrics and spleen is excellent through this incision If only a limited segment

of distal esophagus is resected, it may not be necessary to completely mobilize the gastric conduit One must not sacrifice a tension-free anastomosis, however, in order to save time on gastric conduit mobilization The esophagus should not be dissected more than a few centimeters proximal to the proposed site of division, in order to pre- serve blood supply to the esophagus.

If additional proximal margin is needed, the gus may be dissected behind and above the aortic arch Great care must be taken in this area to avoid injury to the left recurrent laryngeal nerve and thoracic duct The thoracic duct crosses from the right chest to the left chest

esopha-at approximesopha-ately the level of the carina Dissection may

be continued to the neck and total esophagectomy may

be completed by a counterincision in the left neck as described in the section on tri-incisional esophagectomy The anastomosis may be performed in hand-sewn, two-layered manner,6or in a single layer.9The anastomo- sis may also be stapled, although use of the EEA stapler may result in a higher incidence of stricture Omentum may be placed over the anastomosis Some anchor the proximal gastric conduit to pleura with several inter- rupted sutures, although it is questionable whether these sutures placed in pleura in fact take tension off the anas- tomosis.

The abdomen is closed in layers and the costal sion is carefully and securely secured by either wire or a heavy nonabsorbable monofilament figure of eight suture The chest incision is closed in standard fashion.

inci-Colon as Esophageal Replacement

Left Colon

As stated previously, the stomach is the preferred conduit for esophageal replacement Its advantages over colon include a reliable blood supply infrequently affected by atherosclerosis, low level of bacterial colonization, need for only a single anastomosis, and a mucosa that is rela- tively resistant to ulceration At times, because of intrin- sic peptic ulcer disease, previous gastric surgery, or involvement with tumor, the stomach may not be suit- able for use as esophageal replacement In this instance, colon becomes the preferred conduit, with left colon preferred over the right The arterial anatomy is more consistent, the diameter smaller, and usable length is

Strategies for Esophageal Replacement and Reconstruction / 497

FIGURE 41-11 Anastomosis is typically carried out high in the chest

at the level of the azygous vein

usually greater with the left colon If colon is to be used,

preoperative angiography is important in individuals

over 40 years of age and in anyone at risk for

atheroscle-rotic disease A barium enema or colonoscopy is

performed to rule out neoplasia or extensive diverticular

disease A mechanical and antibiotic bowel preparation is

required.

Colon interposition may be used after any of the

methods of esophageal resection The length of colon

needed may be estimated by placing an umbilical tape

between the proposed lines of colon transection and

along the margin of the colon This tape is then placed

from the gastric remnant or jejunal loop along the skin of

the chest to the proposed point of anastomosis in the

neck or chest This approximates the length of colon

needed for a subcutaneous conduit and is always

suffi-cient length for conduit placed in the posterior

medi-astinum The peritoneal attachments of the left colon are

mobilized along the white line of Toldt Using

transillu-mination, the left colic artery, marginal artery of

Drum-mond, and middle colic arter y are inspected and

palpated Clamping of the middle colic artery should

leave the proximal portion of conduit with a palpable

pulse in the marginal artery of Drummond After

deter-mining that the conduit is of good quality, the esophageal

resection is completed The greater omentum is

sepa-rated off the transverse colon The middle colic artery is

divided and the mesocolon is divided using

electro-cautery, a safe distance parallel to the marginal artery

(Figure 41-12) The colocolonic anastomosis is

per-formed, and the mesocolonic defect is closed The left

colon is then mobilized to the neck or chest in the same

atraumatic method as used for a gastric conduit in an

isoperistaltic orientation The posterior mediastinal route

is preferred over substernal placement It is a more direct

route to the neck and requires less length and results in

less conduit angulation and a lower risk of obstruction.

Substernal placement may be required in cases of severe scarring of the posterior mediastinum as occurs after gastric conduit leak or prior esophageal resection.

We believe that performing the distal anastomosis first

is better for determining the necessary length of colon The distal anastomosis can be constructed using a large EEA stapler or in a functional end-to-end manner using

a linear stapler The conduit is draw up through the chest, and necessary length is determined prior to performing the proximal anastomosis The esophago- colonic anastomosis may be constructed by a one- or two-layered hand-sewn method with the end of the esophagus to the side of the antimesenteric tenia It may also be stapled using an EEA stapler or functional end- to-end stapling techniques The conduit should be inspected for arterial insufficiency or venous engorge- ment The crus should be sutured to the conduit to prevent migration of excess colon conduit into the chest

or herniation of abdominal viscera into the chest The anastomoses should be constructed over a nasogastric tube positioned in the distal stomach to allow for postop- erative decompression of the conduit (Figure 41-13).

Right Colon

If the left colon is deemed unacceptable owing to sive diverticular disease or previous surgery, the right colon may be used The right colon is first inspected for any pathology of the colon or mesentery The greater omentum is removed from the proximal two-thirds of the transverse colon The right colon conduit will be supplied by the middle colic artery via the marginal artery, and the ileocolic and right colic artery are clamped with a soft clamp to assess for adequate collat- eral circulation through the marginal artery If perfusion

exten-is deemed adequate, the right colon exten-is harvested and an appendectomy is performed Harvesting of a short segment of distal ileum for use in the proximal conduit is

FIGURE 41-12 The middle colic artery is divided and the mesocolon

is divided using electrocautery

FIGURE 41-13 Anastomoses are constructed over a nasogastric tube

positioned in the distal stomach to allow for postoperative pression of the conduit

decom-controversial Proponents argue that the size match of

the ileum is better for the esophagus, and the presence of

the ileocecal valve may help prevent reflex esophagitis.

Opponents note that reflux high in the neck is rare, and

the ileocecal valve adds bulk to the conduit and may

contribute to antegrade obstruction After mobilizing the

right colon, adequate length is confirmed by placing the

conduit on the skin of the chest wall and bringing it to

the neck From the surgeon’s perspective the right colon

is swung counterclockwise around the axis of the middle

colic artery and rotated to the neck to preserve an

isoper-istaltic orientation The colon ideally is brought to the

neck through the esophageal hiatus and posterior

medi-astinum in the orthotopic if it is being used for primary

esophageal replacement It should be brought to the neck

in atraumatic fashion as described in the section on

tri-incisional esophagectomy In certain situations the

poste-rior mediastinum may be scarred, usually because of

esophageal or gastric conduit leak and mediastinitis In

this instance the substernal route should be used This

distal anastomosis is performed first, the necessary length

of conduit is confirmed, and the proximal anastomosis is

then performed.

Jejunum as Esophageal Replacement

Jejunum is generally considered the third alternative for

esophageal replacement, because it generally cannot

reach the cervical esophagus and thus cannot be used as a

total esophageal substitute except as a free graft Jejunum

may be used as a Roux-en-Y loop, a pedicled graft, or as a

free jejunal graft A Roux-en-Y loop of jejunum will

reach the lower esophagus easily but not the upper

esophagus A point of division is chosen at a comfortable

distance from the ligament of Treitz and several proximal

jejunal branches of the superior mesenteric artery are

divided close to the base of the mesentery Test clamping

with a soft, noncrushing vascular clamp may aid in

assessment of the extent of arterial branches that may be

sacrificed The Roux-en-Y loop is tunneled through the

retrocolic mesentery For distal esophageal replacement,

the midline abdominal incision is taken across the left

costal margin The anastomosis between the esophagus

and jejunum is constructed in end-to-side fashion near

the end of the Roux limb, either stapled or hand-sewn.

The distal anastomosis may be performed end-to-side

hand-sewn or side-to-side with a stapler (Figure 41-14).

Jejunum may also be used as a substitute for a short

segment of esophagus in the treatment of benign

stric-ture Because of the limit on the length of jejunum that

can be mobilized to the chest or neck, jejunum has a

limited role in esophageal replacement for malignant

disease For distal esophageal strictures, pedicled

proxi-mal jejunum can be swung behind the colon and ach and into the lower chest for replacement of the distal esophagus (Figure 41-15) The best application may be in the treatment of peptic distal esophageal strictures that have failed antireflux measures and repeated dilations Distal esophageal excision with gastric pull-up may predispose one to gastroesophageal reflux as disruption

stom-of the hiatus and negative intrathoracic pressure draws gastric secretions into the chest Resection of the stric- ture, leaving the stomach in the abdomen and interpos- ing a short segment of isoperistaltic jejunum, may be the best method of preventing recurrent gastroesophageal reflux disease and stricture.

For isolated benign cervical esophageal disease, free jejunal interposition is an option (Figure 41-16) The diameter of jejunum closely approximates the diameter

of the esophagus, and its lack of bulk is optimally suited

to the neck This technique, first described by Jurkiewicz involves harvesting a short segment of jejunum including its vascularized pedicle in atraumatic fashion.16After dissection of the cervical esophagus and jugular and carotid vessels, the abdomen is explored and a section of jejunum at least 20 cm past the ligament of Treitz is harvested This segment of free jejunum is brought to the neck and the proximal and distal esophageal anastomoses are completed An operating microscope is then used toStrategies for Esophageal Replacement and Reconstruction / 499

FIGURE 41-14 Distal anastomosis may be performed end-to-side

hand-sewn or side-to-side with a stapler

FIGURE 41-15 Pedicled proximal jejunum are swung behind the

colon and stomach and into the lower chest for replacement of thedistal esophagus

the esophagus They found that 20% of patients with

margins less than 5 cm and 8% of patients with margins

less than 10 cm developed local recurrence No patients

with margins greater than 10 cm developed local

recur-rence.22A 10 cm margin is optimal in esophageal resection.

Wong points out that the average tumor is 6.3 cm in

length, and thus even if a near-total esophagectomy is

performed, a 10 cm margin is possible only with mid or

distal esophageal tumors.23 The additional proximal

margin gained from a cervical versus intrathoracic

anasto-mosis has been estimated to be about 2 to 3 cm A longer

margin can be obtained if the anastomosis is made in the

apex of the right chest using an EEA stapler than with a

hand-sewn anastomosis in the mid-high right chest Using

an EEA stapler smaller than 33 mm, however, is associated

with the development of strictures.23Thus, use of a cervical

anastomosis in mid and upper esophageal cancers is useful

in obtaining optimal proximal margins Proponents of an

intrathoracic anastomosis argue that intraoperative

confir-mation of margins using frozen section and the ability to

convert from an Ivor Lewis to cervical anastomosis if

needed makes their approach equally as sound.

The importance of the lymph node dissection, and the

extent of lymph node dissection required, is debated.

Skinner has long been a proponent of radical

esophagec-tomy with en-bloc resection of pleura, pericardium,

thoracic duct, and all regional nodes His group reported

an overall 4-year survival of 37% with en-bloc resection

versus no survivors after standard resection in stage III

esophageal cancer patients.24 This study was

retrospec-tive, nonrandomized, and subject to selection bias but

nonetheless shows a dramatic difference in equally staged

patients To date, no prospective, randomized US studies

have shown a survival advantage to lymphadenectomy in

esophageal cancer patients The study by Goldminc

men-tioned previously, which randomized patients to

transhi-atal versus transthoracic resection, did not show a

significant difference in survival, although there was a

trend favoring survival in patients undergoing

transtho-racic resection with lymph node dissection It is clearly

difficult to assemble a large randomized prospective trial

comparing more radical with conservative resections of

the esophagus, and the issue may never be resolved

completely.

The incidence of certain complications and their

severity varies between techniques of esophageal

resec-tion The most dreaded complication in esophagectomy

is an intrathoracic leak An intrathoracic leak carries a

high mortality rate, and even when survived, it carries a

tremendous burden in intensive care unit and hospital

stay and delayed recovery The incidence of intrathoracic

leak varies greatly but is generally considered to be in the

range of 5 to 10% Early reports described a 70 to 90%

mortality from an intrathoracic leak.10 More recent reports describe a mortality rate of 20 to 44%.25–27With earlier diagnosis and improved treatment in high-volume centers, survival from an intrathoracic leak continues to improve The one thing that is clear regarding an intra- thoracic anastomosis is that it must be done with a low incidence of leak.

The incidence of leak from a cervical anastomosis is generally higher than with an intrathoracic anastomosis, typically in the 10 to 15% range.8

The mortality from a leak in the cervical position is much lower than an intrathoracic leak and is in the range of 2 to 6% A leak in the neck can be diagnosed radiographically on barium swallow or clinically by fever, elevated white blood cell count, localized redness, and tenderness Treatment requires opening the incision at the bedside with local anesthesia, with or without sedation The platysma is opened, and a finger is passed deep to the anastomosis in order to break up any loculated collections The patient is maintained on jejunal feeds, nothing by mouth, until the fistula is closed If the leak is large, placement of a T-tube stent may help control drainage and allow for stenting of the anastomosis Rarely, leakage from a cer vical esophageal anastomosis may enter the thorax This may result from either retraction of the gastroesophageal anastomosis into the high thorax or leak of enteric contents in dependent fashion into the mediastinum This leak must be treated like any life-threatening intrathoracic leak, with thoracotomy, drainage, and diversion, if necessary Provided there is no leakage into the mediastinum, the main sequelae of a cervical esophageal leak are limited to patient inconvenience and

a higher incidence of stricture associated with leak The stricture rate associated with a gastroesophageal anastomosis is higher in the cervical region than in the intrathoracic position and may be almost double (28% vs 16%).28This is commonly attributed to increased tension

on the anastomosis, decreased arterial perfusion in the long conduit, and increased venous congestion of the conduit possibly owing to swelling in a tight cervical compartment Honkoop and colleagues studied the risk factors for late stricture and compared a single-layer continuous synthetic anastomosis to a circular EEA cervi- cal anastomosis.29They found that previous cardiac disease, EEA stapler, and postoperative leak increased the risk for stricture These strictures can usually be treated by bougienage dilation, although repeat dilations may be necessary Dilation of an anastomosis made with an EEA stapler may not be possible In our series, using a combi- nation of hand-sewn and stapled cervical anastomoses, the stricture rate requiring more than two dilations was 9%.9Recurrent laryngeal nerve injury is also higher with an anastomosis in the cervical position versus the intratho-

502 / Advanced Therapy in Thoracic Surgery

racic position A review of 5,483 esophagectomy patients

showed a recurrent nerve palsy rate of 4.8% with the Ivor

Lewis approach versus 11.2% with the transhiatal

approach.28Initially, the Brigham and Women’s Hospital

rate of recurrent laryngeal nerve injury following

tri-incisional esophagectomy with radical mediastinal lymph

node dissection was 17% After modification of surgical

technique including isolation of the cervical esophagus

from within the vagus nerve high in the mediastinum the

rate of recurrent nerve injury dropped to 7%.9Recurrent

nerve palsy may affect quality of life and result in

aspira-tion pneumonia and sepsis Recognized early and treated

aggressively, however, the risk of aspiration pneumonia

and sepsis may be minimized At the Brigham and

Women’s Hospital, any patient who is noted

postopera-tively to have hoarseness or an ineffective cough

under-goes immediate laryngoscopy If a vocal cord paralysis is

noted, it is injected and medialized This aggressive mode

of treatment has limited the incidence of

postesophagec-tomy pneumonia at that institution to 5%.9

The effect of the thoracotomy incision itself is

mani-fested in patient discomfort and splinting Surprisingly,

the incidence of pulmonary complications has not been

shown to be any lower with a transhiatal esophagectomy

than when a thoracotomy is used This may be because of

several factors The widespread use of epidurals, limited

thoracotomy incisions, and muscle-sparing approaches

all aid in limiting post-thoracotomy discomfort and

splinting The increased incidence of recurrent nerve

injury with cervical as opposed to intrathoracic

anasto-mosis with its attendant risk of aspiration may account

for the significant incidence of pulmonary complications

seen with transhiatal esophagectomy, despite the absence

of a thoracotomy incision.

Although both vagus nerves are divided during

esophagectomy, reflux esophagitis in the remaining

portion of esophagus can be a major problem This may

be a result of continued acid production and reflux or

reflux of bile into the conduit It is theorized that

nega-tive intrathoracic pressure draws bile or acid into the

chest Construction of a narrow gastric conduit is also

believed to aid in emptying of the conduit and

minimiza-tion of reflux While it is unclear whether construcminimiza-tion of

a cervical anastomosis carries a lower risk of reflux

esophagitis than with a high intrathoracic anastomosis,

most believe that a distal esophagectomy with low

intrathoracic anastomosis is at high risk for

complica-tions from reflux esophagitis Turnball and Ginsberg

found that 20% of patients undergoing distal

esophagec-tomy needed repeat surgical procedures to relieve

symp-toms from reflux or stricture.30

Conclusion

In summary, the esophageal surgeon must be skilled in a variety of techniques of esophageal resection and recon- struction A surgeon skilled in only a single technique of esophageal resection cannot adequately care for the wide spectrum of patients presenting with esophageal cancer Although in esophageal surgery, as in medicine in general, there are no absolutes, there are general guide- lines that should be followed A 10 cm proximal margin

is desirable, and thus cancers of the middle and upper esophagus should undergo complete resection with anas- tomosis in the neck If an Ivor Lewis resection is used for middle-third lesions, the surgeon must be prepared to convert to a tri-incisional esophagectomy if the initial frozen margin is close or positive In locally advanced middle-third tumors, thoracotomy and dissection under direct vision is desirable and improves the safety of the operation The same may apply to middle-third tumors receiving neoadjuvant therapy For tumors of the distal esophagus, transhiatal, tri-incisional, and Ivor Lewis resection are probably equally as safe and probably result

in equivalent long-term survival.

An intrathoracic anastomotic leak is disastrous, ing a mortality rate of up to 50% Any surgeon that performs esophagectomy with an intrathoracic anasto- mosis must do so with a low incidence of leakage, certainly under 5% The use of an EEA stapler smaller than 33 mm in performing this anastomosis is associated with a higher rate of stricture and should probably be avoided Finally, a skilled esophageal surgeon must be prepared to employ jejunal and colonic conduits if stom- ach is not available for replacement and should use these conduits primarily if a distal esophageal resection is performed for peptic stricture.

carry-References

1 Earlam R, Cunha-Melo JR Oesophageal squamous cell noma: I A critical review of surgery Br J Surg 1980;67:381

carci-2 Ellis FH Carcinoma of the esophagus Cancer 1983;33:264

3 Posthlethwait RW Complications and deaths after tions for esophageal carcinoma J Thorac Cardiovasc Surg1983;85:827

opera-4 Giuli R, Gignoux M Treatment of carcinoma of the agus Retrospective study of 2400 patients Ann Surg1980;192:44

esoph-5 Muller TM Surgical therapy of oesophageal carcinoma Br

J Surg 1990;77:845m

6 Mathisen DJ Transthoracic esophagectomy : a safeapproach to carcinoma of the esophagus Ann Thorac Surg1988;45:137

7 Mitchell RL Abdominal and right thoracotomy approach

as standard procedure for esophagogastrectomy with low

morbidity J Thorac Cardiovasc Surg 1987;93:205

8 Orringer MB, Marshall R, Iannettoni MD Transhiatal

esophagectomy: clinical experience and refinements Ann

Surg 1999;230:392–403

9 Swanson SJ, et al Transthoracic esophagectomy with

radi-cal mediastinal and abdominal lymph node dissection and

cervical esophagogastrostomy for esophageal carcinoma

Ann Thorac Surg 2001;72:1918–25

10 Urschel J Esophagogastrostomy anastomotic leaks

compli-cating esophagectomy: a review Am J Surg 1995;169:634–9

11 McKeown K Total three-stage oesophagectomy for cancer

of the oesophagus Br J Surg 1976;63:259

12 Swanson S, Sugarbaker D The tri-incisional

esophagec-tomy—the Brigham and Women’s Hospital Approach

Chest Surg Clin N Am 2000;10:531–52

13 Orringer MB Technical aids in performing transhiatal

esophagectomy without thoracotomy Ann Thorac Surg

1984;38:128–32

14 Lewis I The surgical treatment of carcinoma of the

esopha-gus with special reference to a new operation for growths of

the middle third Br J Surg 1946;34:18–31

15 Law S, Fok M, Chu KM, Wong J Comparison of hand-sewn

and stapled esophagogastric anastomosis after esophageal

resection for cancer — a prospective randomized

controlled trial Ann Surg 1997;776:169–73

16 Jurkiewicz MJ Vascularizcd intestinal graft for

reconstruc-tion of the cervical esophagus and pharynx Plast Reconstr

Surg 1965;36:509–17

17 Stark SP, Delcore R, et al Transhiatal versus transthoracic

esophagectomy for adenocarcinoma of the distal esophagus

and cardia Am J Surg 1996;172:478–82

18 Pac M, Keles M, et al Transhiatal versus transthoracic

esophagectomy for esophageal cancer J Thorac Cardiovasc

Surg 1993;106:205–9

19 Hankins JR, McLaughlin JS, et al Carcinoma of the

esopha-gus: a comparison of the results of transhiatal vs

transtho-racic resection Ann Thorac Surg 1989;47:700–5

20 Goldminc M, Launois B, et al Oesophagectomy by a shiatal approach or thoracotomy: a prospective randomizedtrial Br J Surg 1993;80:367–76

tran-21 Chu KM, Wong J, et al A prospective randomized son of transhiatal and transthoracic resection for lower-third esophageal carcinoma Am J Surg 1997;174:320–4

compari-22 Tam PC, Cheung TIC, Ma L, et al Local recurrences aftersubtotal esophagectomy for squamous cell carcinoma AnnSurg 1987;205:189–94

23 Wong J Esophageal resection for cancer: the rationale ofcurrent practice Am J Surg 1987;153:18–24

24 Altorki NK, Girardi T, Skinner DB En bloc esophagectomyimproves survival for stage III esophageal cancer J ThoracCardiovasc Surg 1997;114:948–56

25 Sauvanet A Diagnosis and conservative management ofintrathoracic leakage after oesophagectomy Br J Surg1998;85:1446–9

26 Ellis HF, Krasna MJ Esophagogastrectomy for carcinoma ofthe esophagus and cardia: a comparison of findings andinsults after standard resection in three consecutive eight-year intervals with improved staging criteria J ThoracCardiovasc Surg 1997;113:836–46

27 Fahn HJ, Wang LS Leakage of intrathoracic ceral anastomoses in adenocarcinoma of the gastric cardia:changes in APACHE III scores and their prognostic signifi-cance Eur J Surg 1997;163:345–50

oesophagovis-28 Rindani R, Martin C, Cox M Transhiatal versus Ivor Lewisoesophagectomy: is there a difference? Aust N Z J Surg1999;69:187–94

29 Honkoop P, et al Benign anastomotic strictures after hiatal esophagectomy and cervical esophagogastrostomy:risk factors and management J Thorac Cardiovasc Surg1996;111:1141

trans-30 Turnball AM, Ginsberg RJ Options in the surgical ment of esophageal carcinoma Chest Surg Clin N Am1994;4:315–29

Modern esophageal surgery dates to 1913, when Torek

was the first to successfully remove the thoracic

esopha-gus Torek made no attempt to reestablish cervical

esophageal continuity with the gastrointestinal tract.

Since that time numerous approaches to esophageal

resection and reconstruction have been described The

most commonly performed procedures today are the

transhiatal esophagectomy popularized by Orringer and

the combined thoracic and abdominal esophageal

resec-tion.1Although the transhiatal approach is believed by

some to be associated with fewer pulmonary

complica-tions, regardless of the surgical approach, esophagectomy

remains one of the most physiologically stressful surgical

procedures performed today.

Open esophagectomy is associated with high rates of

morbidity (60 to 84%) and mortality rates ranging from

1 to 4% in specialized centers—mortality increases to

over 10% in less experienced hands.2The transhiatal

approach to esophagectomy may limit the pulmonary

morbidity of the procedure but has been criticized as

violating the surgical oncologic principles of en bloc

resection Recent studies have shown that the most

important issue regarding morbidity and mortality for

either approach is that it can be reduced in high-volume

centers.3In an effort to limit the physiologic stress of

esophagectomy while preserving the principle of en bloc

resection, our group at the University of Pittsburgh

Medical Center (UPMC) has developed a minimally

invasive approach to esophageal resection.

Minimally invasive approaches to gastrointestinal

surgery have been shown to protect the patient to some

degree from the physiologic impact of the procedure In

general, most patients experience less pain, fewer wound

complications, less blood loss, and a quicker return to

normal activity However, some complications have been reported to be more common with minimally invasive procedures, including iatrogenic bowel and vascular injuries In addition, the cost of disposable minimally invasive equipment and longer operative times needs to

be considered Since the introduction of the laparoscopic Nissen fundoplication for reflux disease4

there has been a dramatic increase in the types of esophageal disease approached using minimally invasive techniques and in the number of centers performing these procedures Minimally invasive approaches have been described for the treatment of achalasia,5paraesophageal hernia,6and reflux associated with a shortened esophagus7and for the staging and treatment of esophageal cancer.8Our investi- gations in animal and cadaver models documented the feasibility of laparoscopic and thoracoscopic esophageal resection Numerous clinical reports have also confirmed the feasibility and potential advantages of minimally invasive esophagectomy (MIE).

DePaula and colleagues,9 and Swanstrom and Hansen10were the first to report a total minimally inva- sive approach to esophagectomy Their procedure included a laparoscopic transhiatal esophagectomy oper- ation similar to the open transhiatal technique described

by Orringer Subsequently, our UPMC group reported the totally MIE,11which consisted of a thoracoscopic esophageal mobilization procedure as described by Law and colleagues12 and laparoscopic gastric mobilization with a cervical anastomosis as described by De Paula and coworkers and Swanstrom and Hansen The rationale for thoracoscopic esophagectomy was to allow a better view

to facilitate accurate dissection of the esophagus, perform

en bloc lymphadenectomy, and provide hemostasis out the morbidity of a thoracotomy Although total

with-minimally invasive esophagectomy has been shown to be

feasible and safe in experienced hands, controversy still

exists concerning indications, optimal approach, and

clinical benefits This chapter addresses these issues and

summarizes our own experience.

Indications

The indications for the laparoscopic or thoracoscopic

esophagectomy in our early experience were stricture

(1%), lymphoma (0.5%), perforation (0.5%),

tracheo-esophageal fistula (1%), Barrett’s syndrome with

high-grade dysplasia (18%), and carcinoma (79%) In our

initial experience we avoided large bulky tumors and

patients with extensive lymph node metastases As we

gained experience and confidence in the minimally

inva-sive approach, we included patients who had received

neoadjuvant chemotherapy and, in some cases, radiation.

Surgical Approach

In August 1996 we performed our first MIE at UPMC

using an approach similar to that described by DePaula

and Swanstrom using a laparoscopic transhiatal

esophagectomy.9,10 The advantages of the transhiatal

approach include no repositioning of the patient or

single-lung ventilation The limitations of this approach

include the small working space through the esophageal

hiatus Laparoscopy provides only limited access to

thoracic lymph nodes, and there is significant difficulty

gaining mobilization of the middle and upper thirds of

the esophagus Because of these limitations our standard

approach evolved to include thoracoscopy to facilitate

esophageal mobilization and lymph node dissection.8

Initially, patients primarily with Barrett’s high-grade

dysplasia13were offered MIE As our experience has

increased, most patients with resectable lesions are now

included The approach now most commonly used at

UPMC is the combined laparoscopic and thoracoscopic

esophagectomy.

Video-Assisted Thoracic Surgery

Mobilization

The patient is prepared with a double-lumen

endotra-cheal tube for single-lung ventilation and then placed in

the left lateral decubitus position The operating surgeon

is positioned on the patient’s right with the assistant on

the left Port placement is crucial to facilitate a minimally

invasive approach to any complex procedure Four

thora-coscopic ports are placed (Figure 42-1) Initially, the

camera port is placed at the anterior axillary line and

seventh intercostal space The second port (10 mm) is the

working port for the operating surgeon and is in the

posterior axillar y line and the eighth interspace Additional ports are placed posterior to the tip of the scapula and another at the anterior axillary line and fourth intercostal space The final port is for retraction of the lung and countertraction during the esophageal dissection The patient’s body habitus will determine optimal port placement and the surgeon’s flexibility in placement greatly influences the ease of the operation.

An elevated diaphragm sometimes obscures exposure

of the distal esophagus In this case, a single retracting suture (O-Surgitek, US Surgical, Norwalk CT) is placed near the central tendon of the diaphragm and brought out of the inferior anterior chest wall through a 1 mm skin incision This will provide downward traction on the diaphragm to allow good exposure of the distal esophagus.

Dissection of the thoracic esophagus is primarily accomplished with the ultrasonic scalpel Mobilization begins with division of the inferior pulmonary ligament and incision of the mediastinal pleura up to the level of the inferior pulmonary vein Paraesophageal lymph nodes are swept toward the specimen and the subcarinal lymph nodes are dissected with the specimen exposing the left and right main stem bronchi The azygos vein is divided using the endo-GIA stapler (US Surgical) The pleura is then incised lateral to the esophagus Aortoesophageal vessels are divided, and clips are used liberally in the area of the thoracic duct To facilitate exposure a Penrose drain is placed around the esophagus (Figure 42-2) and used as a traction device Following mobilization of the esophagus from the thoracic inlet to the diaphragmatic reflection, a single 28 F chest tube is inserted through the camera port, and an intercostal block using 0.5% Marcaine is administered Ports are closed, and the patient is turned to the supine position.

FIGURE 42-1 Video-assisted thoracoscopic surgical port sites.

ment of Treitz to the anterior abdominal wall using the

Endostitch A needle catheter kit (Compact Biosystems;

Minneapolis, MN) is placed percutaneously into the

peri-toneal cavity under direct laparoscopic vision and

direct-ed into the loop of jejunum The guide wire and catheter

are threaded into the loop of jejunum The jejunal

punc-ture area is tacked completely to the anterior abdominal

wall for a distance of several centimeters.

Neck Dissection

A 4 to 6 cm horizontal neck incision is made just above

the suprasternal notch and the cervical esophagus

exposed Finger dissection is continued down into the

mediastinum until the thoracic dissection plane is

encountered The cervical esophagus is divided and the

esophagogastric specimen pulled out of the neck incision.

As traction is applied to the specimen in the neck, the

assistant guides the specimen in its proper alignment into

the mediastinum Care is taken to maintain orientation of

the gastric tube The specimen is removed from the field.

An anastomosis is performed between the esophagus and

gastric tube The completed reconstruction is shown in

Figure 42-5.

Technical Considerations

Several technical details will facilitate minimally invasive esophagectomy For example, the ultrasonic coagulating shears have been timesaving in our minimally invasive experience The entire mobilization of the esophagus can

be done without the tedium of clipping and dividing or tying The left gastric vessels are safely and rapidly divided using the endoscopic stapling device with a vascular load To date, we have had no bleeding compli- cations associated with reliance on these hemostatic tech- niques In our initial experience, early opening of the phrenoesophageal ligament led to loss of insufflating gas into the chest during laparoscopy, and we now perform this as a last step Dissection of the posterior esophagus

in the area of the thoracic duct can lead to thoracic duct leaks, which we reported early on in our experience; liberal use of endoclips during the lateral esophageal thoracic dissection has eliminated this problem.

The use of thoracoscopy has several technical tages It allows better visualization for nodal clearance, a controlled dissection of the esophagus to avoid injury of mediastinal structures, and a decrease in operative blood loss Although we did not encounter any major intraop- erative emergencies during thoracoscopic dissection, several investigators have reported major complications such as bleeding, aortic injury, and tracheal laceration.13,14

advan-We found early on in our experience that the total laparoscopic esophagectomy did not allow us the visual- ization that thoracoscopy allows.

Minimally invasive esophagectomy is a technically demanding surgery requiring a thorough knowledge of the open procedure, appropriate instrumentation, and advanced laparoscopic skills, including suturing and complex tissue dissection Despite the introduction of instruments such as ultrasonic coagulating shears, endo- scopic stapling devices, and bipolar scissors, which make the procedure easier, it remains complex because of the extensive and careful dissection and multiple steps Because of the variability in minimally invasive surgical skills, it is difficult to define a learning curve or to estimate how much the operative time will decrease with experi- ence However, in our experience, the operation can be performed routinely in less than 5 hours for routine cases.

Oncological Perspectives

The appropriateness of MIE for any malignancy always raises the question of adherence to oncologic principles Currently, there is not a consensus on this issue for esophageal cancer regarding even the best open approach Advocates of transhiatal resection promote it

as a less morbid surgery with comparable, although poor, survival rates when compared with more extensive resec-

Minimally Invasive Esophagectomy / 507

FIGURE 42-5 Completed laparoscopic–thoracoscopic esophagectomy.

tions Advocates of more extensive resection claim an

improved long-term survival rate despite the increased

morbidity of the added thoracotomy.

Theoretically, combined thoracoscopic and

laparo-scopic esophagectomy offers the advantage of radical

surgical resection with the lessened physiologic insult,

which should result in shorter recovery periods Many

surgeons remain justifiably concerned about the possible

dangers represented by minimally invasive treatments of

cancer These include missed lesions, inadequate staging,

compromised margins, appropriate nodal dissections,

port-site cancer recurrence, and possible tumor

dissemi-nation.

The margins of resection of all patients in our

experi-ence who underwent combined thoracoscopic and

laparoscopic esophagectomy were free of tumor or

Barrett’s metaplasia at the time of frozen section In our

first 77 patients, we did encounter 3 who had a final

pathology report with positive microscopic advential

margins or lymphatic invasion; this is not different from

our open experience We use intraoperative endoscopy

liberally to identify the proximal and distal extent of

tumor involvement before surgical resection Endoscopy

helps to identify patients with tumor involvement of the

gastroesophageal junction; a more extensive resection of

the gastric cardia can be performed The mean number

of lymph nodes retrieved from combined thoracoscopic

and laparoscopic esophagectomy is similar to the number

retrieved from our previous open procedures.

Oncologic principles are followed during combined

thoracoscopic and laparoscopic esophagectomy Surgical

dissection and direct manipulation of the tumor mass is

minimized Thoracic trocars were used at all chest port

sites to prevent direct contact of the dissecting

instru-ment with the chest wall We removed the resected

speci-men through the cervical incision instead of through an

abdominal incision, as one mechanism for port-site

cancer recurrence is direct contact of the tumor at the

limited extraction site In our first 77 patients, at a

median of 3 years’ follow-up, there has been no port-site

metastasis observed after combined thoracoscopic and

laparoscopic esophagectomy.6

MIE appears to be as safe as conventional

esophagec-tomy when performed in a center with extensive

mini-mally invasive and open surgical esophageal experience.

In our experience, advanced fellowship training in

mini-mally invasive surgery enhances the safety of complex

laparoscopic or thoracoscopic procedures Applications

of thoracoscopy and laparoscopy to esophagectomy

elim-inate the thoracotomy and laparotomy incisions and

therefore further reduce operative trauma and enhance

postoperative recovery Further clinical trials will be

necessary to confirm this single institution experience.

3 Patti MG, Corvera C, Glasgow RE, et al A hospital’s annualrate of esophagectomy influences the operative mortalityrate J Gastrointest Surg 1998;2:186–92

4 Dallemagne B, Weerts JM, Jehaes C, et al LaparoscopicNissen fundoplication; preliminary report Surg Endosc1991;3:138–43

5 Patti MG, Pelligrini CA, Horgan S, et al Minimally invasivesurgery for achalasia: an 8 year experience 168 patients.Ann Surg 1999;4:587–93

6 Luketich JD, Schauer PR, Christie NA, et al Minimallyinvasive esophagectomy Ann Thorac Surg 2000;70:906–11

7 Johnson AB, Oddsdottir M, Hunter JG Laparoscopic Collisgastroplasty and Nissen fundoplication A technique for themanagement of esophageal foreshortening Surg Endosc1998;12:1055–60

8 Luketich JD, Schauer P, Landreneau R, et al Minimallyinvasive surgical staging is superior to endoscopic ultra-sound in detecting lymph node metastases in esophagealcancer J Thorac Cardiovasc Surg 1997;114;817–21

9 DePaula AL, Hashiba K, Ferreira EAB, et al Laparoscopictranshiatal esophagectomy with esophagogastroplasty SurgLaparosc Endosc 1995;5:1–5

10 Swanstrom L, Hansen P Laparoscopic total esophagectomy.Arch Surg 1997;132:943–7

11 Nguyen NT, Schauer PR, Luketich JD Combined scopic and thoracoscopic approach to esophagectomy J AmColl Surg 1999;118;328–32

laparo-12 Law S, Fok M, Chu KM, et al Thoracoscopic tomy for esophageal cancer Surgery 1997;122:8–14

esophagec-13 Bemelman WA, Taat CW, Slors FM, et al Delayed ative emptying after esophageal resection is dependent onthe size of the gastric substitute J Am Coll Surg1995;180:461–4

postoper-14 Millikan KW, Silverstein J, Hart V, et al A 15-year review ofesophagectomy for carcinoma of the esophagus and cardia.Arch Surg 1995;130:617–24

The optimal technique for esophageal resection has been

hotly debated for the past 25 years Part of this ongoing

dialogue has involved the method used for esophageal

anastomosis—should the anastomosis be hand sewn or

stapled, and should it be located in the chest or the neck?

Despite this intensive investigation, anastomotic leaks

following esophageal resection continue to plague

thoracic surgeons This chapter examines the differing

methods of esophageal anastomosis and discusses the

management of anastomotic leaks when they occur.

Esophageal Anastomotic Techniques

Selection of Conduit

The clear majority of esophageal anastomoses are

constructed with the use of a gastric conduit, and hence

the following comments pertain largely to

esophagogas-tric reconstruction Alternative conduits, such as colon or

jejunum, allow for a variety of techniques analogous to

esophagogastric anastomoses.

Location of Anastomosis

The location of the anastomosis is dependent largely on

the operative approach to esophagectomy Use of a

thora-coabdominal or an Ivor Lewis (combined right

thoraco-tomy and laparothoraco-tomy) approach will place the

anastomosis within the thoracic cavity; a transhiatal and

so-called “three-hole” (modified McKeown) technique

depend on a cervical anastomosis Both intrathoracic and

cervical anastomoses have certain advantages, but they also expose the patient to risks particular to each anasto- motic site.

From an oncologic point of view, the cervical motic location offers the potential benefit that a larger amount of esophagus is resected, and thus the proximal margin is more substantial This is particularly important for tumors involving the middle and upper third of the esophagus and has been linked to a lower incidence of local recurrence.1–3For gastroesophageal junction tumors this difference is less crucial, especially if the intratho- racic anastomosis is placed high in the chest cavity Further, it is unlikely the location of the anastomosis per

anasto-se has an impact on overall survival Although a study examining survival based on anastomotic location has not been published, comparison of survival rates follow- ing transthoracic and transhiatal esophagectomies (a very rough comparison of intrathoracic vs cervical anas- tomoses) show no difference in outcome.4–8 Survival following esophageal resection for carcinoma is depen- dent on tumor stage rather than the operative technique The incidence and consequences of an anastomotic leak following esophagectomy are clearly different between an intrathoracic and cervical anastomosis Intrathoracic anas- tomoses tend to leak less frequently, with a reported inci- dence from 0 to 7%, and often less than 2%.4,7–15 In contrast, cervical anastomoses have a higher reported inci- dence of leak, usually ranging from approximately 8 to 15%.4,7,8,13,16–18In 1999, Orringer and colleagues published

updated figures for the University of Michigan experience

with transhiatal esophagectomy in 1,085 patients, noting a

cervical leak rate of 13% despite a variety of anastomotic

techniques.19 However, the same authors have recently

described a newly devised cervical anastomotic technique

with a reported anastomotic leak rate of 2.7% in 111

oper-ative survivors.20Heitmiller and colleagues reported their

cervical anastomotic leak rate of just 0.8% in 262

consecu-tive patients in whom a two-layer hand-sewn technique

was used.21

These latter studies would seem to suggest that

a low leak rate is clearly achievable with a cervical

anasto-mosis The incidence of lethal mediastinitis is rare

follow-ing a cervical anastomotic leak but can exceed 60%

following an intrathoracic leak.9,22 More recent studies

suggest the mortality rate following an intrathoracic leak

to be approximately 20 to 35%.23–25

Several other factors should be weighed when

consid-ering anastomotic location following esophagectomy.

Benign stricture at the anastomotic site has been linked

to a prior history of anastomotic leak20,26and thus may be

encountered more frequently following cervical

anasto-moses, although this incidence is also dependent on the

anastomotic technique employed Recurrent laryngeal

nerve injury occurs in 7 to 14% of patients following

esophagectomy incorporating cervical dissection18,19and

likely contributes to the aspiration syndromes

occasion-ally seen after surgery Patients with an intrathoracic

anastomosis are more prone to gastroesophageal reflux,

at times severe These same patients, because of the

oblig-atory thoracic incision, can experience a higher incidence

of pulmonary complications than do transhiatal patients

in the postoperative period.8,22

Anastomotic Technique

Anastomoses to reconstruct the alimentary tract

follow-ing esophageal resection may be constructed with a

hand-sewn method, with the use of a stapling device, or

with a combination of both techniques The technique

chosen is largely dependent on the individual surgeon’s

experience and preference A hand-sewn technique is

more versatile, particularly if an alternative conduit

(colon or jejunum) is used and if performed correctly

promotes mucosa-to-mucosa apposition leading to less

stricture formation Stapled anastomoses are completed

faster, are more reproducible, and are perhaps a bit easier

to construct, thus tending to appeal more to the

occa-sional esophageal surgeon The use of stapling

instru-ments also may facilitate construction of the anastomosis

in select areas, such as the apex of the chest.

There are several features of a successful anastomosis

that should remain constant regardless of the exact

tech-nique chosen (Table 43-1) Conduit ischemia remains a

major factor in the development of anastomotic

compli-cations and should be minimized at all costs.13,27 For esophagogastric anastomoses, the proximal 20% of the gastric fundus (with sacrifice of the left gastric vessels) is supplied through a microscopic network of collateral vessels fed primarily from the right gastroepiploic system.28This fact has led some investigators to preserve

as much of the stomach as possible, rather than fashion a slim gastric tube, in order to preserve the collateral blood supply.20The degree of ischemia of the translocated conduit may or may not be clinically apparent and most commonly results from over aggressive mobilization of the conduit resulting in arterial or venous insufficiency Occasionally, twisting or external compression of the conduit results in further vascular compromise at the proximal anastomotic site These factors may play a role

in the higher rate of anastomotic complications mented with cervical anastomoses Recently, efforts have been made to measure fundal tissue perfusion intraoper- atively with oximetry, photoplethysmography or laser Doppler flowmetry to minimize ischemic complications

docu-at the anastomotic site.29–31Although esophageal ischemia

is less of an issue, excessive dissection of the proximal esophagus should be discouraged given the segmental nature of the blood supply to the organ.10The gentle handling of tissues at the anastomotic site will further reduce mechanical trauma leading to ischemic compro- mise Finally, measures to reduce anastomotic tension and promote conduit decompression in the early postop- erative period should be adopted to diminish mechanical forces that can lead to anastomotic disruption.

Hand-Sewn Anastomoses

For many esophageal surgeons, questions remain ing the optimal approach to a hand-sewn esophageal anastomosis: single-layer or two-layer closure, running or interrupted suture, absorbable or nonabsorbable suture There is a paucity of carefully controlled studies examin- ing the relative benefits of the various methods Zieren and colleagues conducted a prospective, randomized trial

regard-of single- versus two-layer closure techniques for cervical gastroesophageal anastomoses, noting a significantly higher stricture rate in two-layer closures.32 This study was marred by an anastomotic leak rate approaching 20% in each arm Bardini and colleagues found no differ-

TABLE 43-1 Essential Elements of a Successful EsophagealAnastomosis

Preserve conduit blood supplyAvoid conduit torque or compressionAvoid excessive esophageal dissectionHandle tissues gently

Minimize anastomotic tensionPerform postoperative conduit decompression

ences in outcome between running and interrupted

suturing techniques in single-layer cervical anastomoses,

although they felt the running technique was cheaper,

faster, and easier to teach to resident staff.33Others have

found that a running suture technique was associated

with a higher rate of anastomotic leak.34

Beyond these reports, information concerning

hand-sewn anastomotic techniques comes largely from

single-institution studies detailing outcomes following

esophagectomy Perhaps the most impressive results

come from the Massachusetts General Hospital, using a

technique first described by Churchill and Sweet in 194235

and most recently reported in modified form by

Mathisen and colleagues in 1988.10A two-layer,

inter-rupted suture technique was used where gentle tissue

handling, precise suture placement, and careful mucosal

apposition were emphasized (Figure 43-1) Silk suture

was used in both layers, although some surgeons

incor-porating this technique substitute a synthetic absorbable

suture for the inner layer Mathisen reported a 0%

anas-tomotic leak rate and a 5% stricture rate with this

method in 104 consecutive patients undergoing

transtho-racic esophagectomy The ability to teach this method

was also emphasized, with over 90% of these cases

performed by the resident staff Heitmiller and colleagues

at Johns Hopkins Hospital applied this anastomotic

tech-nique to cervical gastroesophageal anastomoses,

describ-ing a remarkable 0.8% leak rate in 262 consecutive

patients following esophagectomy.21

Stapled Anastomoses

Stapling instruments to facilitate esophageal anastomoses

have been employed for approximately 25 years,

primar-ily through the use of an end-to-end anastomotic device

(EEA stapler, US Surgical) This instrument is inserted

through an anterior gastrotomy, with the leading center

rod passed back through the stomach wall at the desired

anastomotic site Following attachment of an anvil to the

center rod, the anvil is passed into the distal esophagus,

secured by tightening a previously placed purse-string

suture As the stapler is fired, the two anastomotic ends

are brought together, placing the anvil in close

approxi-mation with the gun (within the stomach) carrying the

staples Two circular rows of staples are fired, and the

compressed portions of intervening stomach and

esoph-agus are excised automatically This produces two intact

“doughnuts” of tissue that should be inspected to assure

adequate completion of the anastomosis The EEA device

is relatively easy to use, provides reproducible results, and

may facilitate completion of anastomoses in select areas

where a hand-sewn technique is cumbersome (such as

the apex of the chest) Conversely, this technique does

not work well for cervical anastomoses The rate of

anas-tomotic leak noted in a meta-analysis of 1,964 cases of esophagogastrectomy using the EEA stapler was 13.9%,22although reports from experienced esophageal centers detail a leak rate of 0 to 4%.36–39Another drawback of the EEA stapler is the higher rate of late anastomotic stric- ture formation seen with the device, which may approach

30 to 40%.26,38,40,41

These strictures may relate to the lack of mucosal to mucosal apposition inherent in the anasto- moses constructed by these devices.13The rate of stricture formation appears to correlate with the size of the stapler used, with the smaller diameter devices causing the higher stricture rate.42,43

Frustrated by a persistent cervical leak rate of 10 to 15% despite using a variety of anastomotic techniques, Orringer and colleagues recently described their results with a new method incorporating the use of a 3 cm long Endo-GIA stapler (US Surgical) to construct a functional side-to-side esophagogastric anastomosis (Figures 43-2

to 43-8).20Approximately three-fourths of the sis is completed through a single firing of the stapler, with the remaining open gastrotomy and esophageal lumen closed manually in two layers Care is taken to avoid the gastric staple line when constructing the anas- tomosis to minimize the risk of ischemia of the gastric tip and anastomotic site This technique offers several advantages in that it is simple to complete, is adaptable to differing conduits and anastomotic locations, and provides a widely patent (at least 3 cm) anastomosis that

anastomo-is less susceptible to subsequent stricture formation Using this method, Orringer and coworkers were able to reduce their anastomotic leak rate to 2.7% and also noted

a reduction in the number of postoperative anastomotic dilations required.

Stapled or Hand-Sewn?

Although previous reports have described superior results with stapled as compared with hand-sewn anasto- moses,13,44 this bias has not been confirmed in random- ized studies Beitler and Urschel pooled results from four randomized studies comparing stapled and hand-sewn anastomotic techniques following esophagogastrectomy and found no difference in the incidence of anastomotic leak (stapled 9%, hand-sewn 8%).41In the same compari- son, stapled anastomoses were associated with a higher rate of stricture formation (27%) compared with hand-

sewn anastomoses (16%, p < 02) Further, although in

nonrandomized studies stapled anastomoses were ated with a lower leak rate (6% vs 11%) than were hand- sewn anastomoses, this difference evaporated when only prospective data was evaluated In these comparative studies the stapling techniques utilize the circular EEA stapler, rather than the linear stapler used in the method described more recently by Orringer.

associ-Esophageal Anastomotic Techniques and the Management of associ-Esophageal Anastomotic Leaks / 511

nal route associated with a higher rate compared with placement of the conduit through the bed of the resected esophagus.32,45One may theorize that this is because the substernal route is longer and may lead to increased compression of the conduit at the level of the thoracic inlet Other intraoperative variables that may contribute

to a higher anastomotic leak rate include excessive blood loss3 4 and use of a colonic as opposed to a gastric conduit.45 The presence of an anastomotic margin posi- tive for residual carcinoma has been inconsistently linked

to an increased incidence of leak.46–48

A number of other variables have been postulated to play a role in the development of esophageal anastomotic leaks Hypoalbuminemia (as a surrogate for malnutri- tion) has been identified as a risk factor for anastomotic complications by some,34,47and not by others.48 The majority of patients undergoing esophagectomy for benign or malignant disease have some degree of malnu- trition, and most surgeons would agree that in severe cases this should be addressed prior to surgery The use

of induction chemotherapy, radiotherapy, or combined chemoradiation does not appear to affect the complica- tion rate after esophageal resection,9,34,47,49–52 although higher anastomotic leak rates have been reported in patients undergoing “salvage esophagectomy” following definitive chemotherapy and radiotherapy.53 The pres-

FIGURE 43-6 The Endo-GIA stapler (30 mm in length, 35 mm staples)

is introduced into the stomach and esophagus A, The insertion is

facilitated by the previously placed traction sutures and by rotating

the stapler so that the anvil is inserted into the stomach B, Great care

is taken to align the stomach and esophagus in a parallel fashion,

avoiding the previously completed staple line of the gastric conduit

Reproduced with permission from Orringer MB.20

FIGURE 43-7 A, The jaws of the stapler are approximated, allowing

the surgeon to rotate the stapler to confirm appropriate anastomotic

location and to place additional seromuscular suspension sutures B,

The stapler is fired and removed The staple line is inspected forhemostasis, and the nasogastric tube is positioned under direct vision.Reproduced with permission from Orringer MB et al.20

FIGURE 43-5 The stapled proximal end of the esophagus (transected

obliquely using an Endo-GIA stapler) is removed Two stay sutures are

placed, one lifting the anterior tip of the cut esophagus while the

other approximates the lower esophageal edge to the midpoint of the

gastrostomy Reproduced with permission from Orringer MB et al.20

Although it is not as sensitive as the standard

esopha-gogram, computed tomography (CT) can occasionally be

helpful in patients where a high degree of suspicion of a

leak persists despite equivocal testing The presence of

mediastinal air or loculated fluid can suggest the

pres-ence of an occult leak In select cases, CT scanning can

guide therapy via percutaneous catheter placement.

Esophagoscopy using a flexible endoscope is also less

sensitive than contrast esophagography in detecting

anastomotic leaks However, endoscopy can be quite

helpful in evaluating the proximal conduit for evidence

of vascular insufficiency The finding of ischemia in

conjunction with signs and symptoms of anastomotic

disruption should dispel uncertainty and prompt the

surgeon to action.

Management: Fulminant Leaks

Initial management for an early, catastrophic leak includes

intravenous fluid resuscitation, broad-spectrum

antibi-otics, inotropic support as indicated, and immediate

operative intervention Some degree of conduit necrosis is

usually encountered, and the anastomosis is taken down.

The ischemic portion of the conduit is resected with the

remainder returned to the abdomen A cervical

esophagostomy and gastrostomy is performed, and the

contaminated mediastinum is extensively debrided and

drained If the patient survives this initial insult,

gastroin-testinal continuity is reestablished at a later date, usually

with a retrosternal colonic interposition.

Management: Cervical Leaks

Many small, well-contained cervical leaks require no

specific intervention This type of leak is usually

asymp-tomatic and found at the routine contrast study 5 to 7

days following surgery Frequently, the cavity associated

with the leak decompresses (empties) back into the

alimentary tract Given the lack of sepsis, antibiotics are

usually not required and the patient is either maintained

nil by mouth or on clear liquids for several additional

days A contrast study is often performed to document

healing of the leak, although this is probably not

neces-sary if the patient remains asymptomatic.

The presence of a larger, contained cervical leak that

produces symptoms of wound erythema, fluctuance, and

drainage requires open evacuation If the cervical wound

is closed appropriately at the initial operation, this can be

accomplished at the bedside with little discomfort to the

patient Thorough inspection of the cervical anastomotic

site should result in clearance of all loculated fluid

pock-ets Occasionally, a small anastomotic defect will be noted

and amenable to direct suture closure, thereby

shorten-ing the healshorten-ing process Twice daily dressshorten-ing changes are

initiated, and antibiotics are utilized if a significant

degree of cellulitis is present The patient is maintained

on enteral nutrition via the previously placed tomy tube and kept nil by mouth, although some surgeons recommend having the patient drink several glasses of water daily to “flush” the wound.58The majority

jejunos-of leaks will close over the following 2 weeks, and it is rare that formal reoperation with anastomotic repair or tissue flap advancement is necessary Failure of the patient to respond promptly with this approach should alert the surgeon to the possibility of inadequate drainage or unsuspected conduit necrosis and the subse- quent development of a catastrophic problem, such as descending mediastinal infection or vascular erosion of the undrained abscess.59

Management: Intrathoracic Leaks

Small intrathoracic leaks that are asymptomatic and dentally discovered on routine esophagography require

inci-no specific therapy Contained moderate-sized leaks that are asymptomatic and drain spontaneously back into the esophagus can also occasionally be treated expectantly, although this approach should be used cautiously as fistulization into vital mediastinal structures can occur If symptoms intervene or there is concern regarding erosion into adjacent organs, these contained fluid collec- tions should be drained In these cases, institution of broad-spectrum antibiotics should be considered Some surgeons have attempted internal drainage by placement

of a nasogastric tube (under fluoroscopic guidance) through the anastomotic defect, although this technique risks further disruption of the anastomosis.60Successful drainage of the abscess can often be established percuta- neously under CT guidance,61which leads to sponta- neous healing of the leak over 1 to 3 weeks If rapid improvement in the patient’s condition does not occur, open drainage should be undertaken This latter approach will allow direct inspection of the anastomotic site, and occasionally direct repair of the leak will be possible, buttressed with a vascularized tissue flap of muscle or pericardial fat Debridement of devitalized tissue and establishment of adequate drainage completes the procedure In cases of poorly contained or noncon- tained leaks, attempts at conservative management should be abandoned in favor of prompt reexploration.

Complications

Fistulization to the airway or a major blood vessel can occur, particularly if the anastomotic leak remains neglected for a prolonged period of time Communication with the airway typically occurs to the trachea or left main bronchus and presents with symptoms of airway contami- nation and infection Delineation of the fistula via contrast esophagography will localize the problem and help define

516 / Advanced Therapy in Thoracic Surgery

the degree of anastomotic disruption Optimal repair

includes division of the fistulous tract, repair of the airway

and anastomotic defects, and interposition of

well-vascularized tissue between the two repair sites Fistulas to

major vessels are even more problematic and are often

fatal at initial presentation despite aggressive intervention.

Rarely a “herald bleed” will suggest the fistulous

commu-nication, allowing a planned operative approach Vascular

control of the offending vessel with repair or ligation at the

fistula site, combined with vascular bypass outside the

infected field and secure, buttressed repair of the

anasto-mosis can lead to salvage in selected cases.

The presence of an anastomotic leak frequently leads

to the later development of an anastomotic stricture,

with an incidence as high as 57% in one study.26 Early

recognition of this problem and aggressive management

with serial dilations can ameliorate many if not all of the

presenting symptoms It should be remembered, though,

that not all cases of dysphagia reported by patients in the

weeks following esophagectomy represent stricture at the

anastomotic site Dysfunction owing to edema and

denervation injury to the mobilized esophagus can

spon-taneously resolve with the passage of time, leading to

effective pharyngoesophageal swallowing without

specific intervention.

References

1 Tam PC, et al Local recurrences after subtotal

esophagec-tomy for squamous cell carcinoma Ann Surg

1987;205:189–94

2 Kato H, et al Anastomotic recurrence of oesophageal

squa-mous cell carcinoma after transthoracic oesophagectomy

Eur J Surg 1998;164:759–64

3 Blewett CJ, et al Local recurrence after total or subtotal

esophagectomy for esophageal cancer J Exp Clin Cancer

Res 2001;20:17–9

4 Goldminc M, et al Oesophagectomy by a transhiatal

approach or thoracotomy: a prospective randomized trial

Br J Surg 1993;80:367–70

5 Chu KM, et al A prospective randomized comparison of

transhiatal and transthoracic resection for lower-third

esophageal carcinoma Am J Surg 1997;174:320–4

6 Pommier RF, et al Relationships between operative

approaches and outcomes in esophageal cancer Am J Surg

1998;175:422–5

7 Rindani R, Martin CJ, Cox MR Transhiatal versus

Ivor-Lewis oesophagectomy: is there a difference? ANZ J Surg

1999;69:187–94

8 Hulscher JB, et al Transthoracic versus transhiatal resection

for carcinoma of the esophagus: a meta-analysis Ann

Thorac Surg 2001;72:306–13

9 Karl RC, et al Factors affecting morbidity, mortality, andsurvival in patients undergoing Ivor Lewis esophagogas-trectomy Ann Surg 2000;231:635–43

10 Mathisen DJ, et al Transthoracic esophagectomy: a safeapproach to carcinoma of the esophagus Ann Thorac Surg1988;45:137–43

11 Page RD, et al Esophagogastrectomy via left cophrenotomy Ann Thorac Surg 1990;49:763–6

thora-12 Putnam JB, et al Comparison of three techniques ofesophagectomy within a residency training program AnnThorac Surg 1994;57:319–25

13 Urschel JD Esophagogastrostomy anastomotic leakscomplicating esophagectomy: a review Am J Surgery1995;169:634–40

14 Visbal AL, et al Ivor Lewis esophagogastrectomy foresophageal cancer Ann Thorac Surg 2001;71:1803–8

15 Shahian DM, et al Transthoracic versus extrathoracicesophagectomy: mortality, morbidity, and long-termsurvival Ann Thorac Surg 1986;41:237–46

16 Gurkan N, et al Transhiatal oesophagectomy foroesophageal carcinoma Br J Surg 1991;78:1348–51

17 Katariya K, et al Complications of transhiatal tomy J Surg Oncol 1994;57:157–63

esophagec-18 Swanson SJ, et al Transthoracic esophagectomy with cal mediastinal and abdominal lymph node dissection andcervical esophagogastrostomy for esophageal carcinoma.Ann Thorac Surg 2001;72:1918–25

radi-19 Orringer MB, Marshall B, Iannettoni MD Transhiatalesophagectomy: clinical experience and refinements AnnSurg 1999;230:392–400; discussion 400–3

20 Orringer MB, Marshall B, Iannettoni MD Eliminating thecervical esophagogastric anastomotic leak with a side-to-side stapled anastomosis J Thorac Cardiovasc Surg2000;119:277–88

21 Heitmiller RF, Fischer A, Liddicoat JR Cervical gogastric anastomosis: results following esophagectomy forcarcinoma Dis Esophagus 1999;12:264–9

esopha-22 Muller JM, et al Surgical therapy of oesophageal noma Br J Surg 1990;77:845–57

carci-23 Sauvanet A, et al Diagnosis and conservative management

of intrathoracic leakage after oesophagectomy Br J Surg1998;85:1446–9

24 Agrawal S, et al Intrathoracic anastomosis after oesophagealresection for cancer J Surg Oncol 1996;63:52–6

25 Whooley BP, et al Analysis of reduced death and tion rates after esophageal resection Ann Surg2001;233:338–44

complica-26 Honkoop P, et al Benign anastomotic strictures after shiatal esophagectomy and cervical esophagogastrostomy:risk factors and management J Thorac Cardiovasc Surgery1996;111:1141–6; discussion 1147–8

tran-27 Pierie JP, et al Healing of the cervical esophagogastrostomy.

J Am Coll Surg 1999;188:448–54

28 Liebermann-Meffert DM, Meier R, Siewert JR Vascular

anatomy of the gastric tube used for esophageal

reconstruc-tion Ann Thorac Surg 1992;54:1110–5

29 Uribe N, et al Evaluation of residual vascularisation in

oesophageal substitution gastroplasty by surface

oximetry-capnography and photoplethysmography An experimental

study Eur J Surg 1995;161:569–73

30 Pierie JP, et al Impaired healing of cervical

oesophagogas-trostomies can be predicted by estimation of gastric serosal

blood perfusion by laser Doppler flowmetry Eur J Surg

1994;160:599–603

31 Boyle NH, et al Intraoperative scanning laser Doppler

flowmetry in the assessment of gastric tube perfusion during

esophageal resection J Am Coll Surg 1999;188:498–502

32 Zieren HU, Muller JM, Pichlmaier H Prospective

random-ized study of one- or two-layer anastomosis following

oesophageal resection and cervical oesophagogastrostomy

Br J Surg 1993;80:608–11

33 Bardini R, et al Single-layered cervical esophageal

anasto-moses: a prospective study of two suturing techniques Ann

Thorac Surg 1994;58:1087–9; discussion 1089–90

34 Dewar L, et al Factors affecting cervical anastomotic leak

and stricture formation following esophagogastrectomy

and gastric tube interposition Am J Surg 1992;163:484–9

35 Churchill ED, Sweet RH Transthoracic resection of tumors

of the stomach and esophagus Ann Surg 1942;115:897–920

36 Berrisford RG, Page RD, Donnelly RJ Stapler design and

strictures at the esophagogastric anastomosis J Thorac

Cardiovasc Surg 1996;111:142–6

37 Donnelly RJ, Sastry MR, Wright CD Oesophagogastrectomy

using the end to end anastomotic stapler: results of the first

100 patients Thorax 1985;40:958–9

38 Fok M, et al Comparison of a single layer continuous

hand-sewn method and circular stapling in 580

oesophageal anastomoses Br J Surg 1991;78:342–5

39 Bardini R, et al Anastomosis World J Surg 1994;18:373–8

40 Law S, et al Comparison of hand-sewn and stapled

esopha-gogastric anastomosis after esophageal resection for cancer:

a prospective randomized controlled trial Ann Surg

1997;226:169–73

41 Beitler AL, Urschel JD Comparison of stapled and

hand-sewn esophagogastric anastomoses Am J Surg

1998;175:337–40

42 Griffin SM, et al Early and late surgical complications of

subtotal oesophagectomy for squamous carcinoma of the

oesophagus J R Coll Surg Edinburgh 1991;36:170–3

43 Muehrcke DD, Kaplan DK, Donnelly RJ Anastomotic

narrowing after esophagogastrectomy with the EEA

stapling device J Thorac Cardiovasc Surg 1989;97:434–8

44 Ferguson MK Management of esophageal anastomoticleaks In: Franco KL, Putnam JB Jr, editors Advanced ther-apy in thoracic surgery Hamilton (ON): BC Decker Inc.;

47 Patil PK, et al Cancer of the esophagus: esophagogastricanastomotic leak—a retrospective study of predisposingfactors J Surg Oncol 1992;49(Suppl):163–7

48 Peracchia A, et al Esophagovisceral anastomotic leak Aprospective statistical study of predisposing factors JThorac Cardiovasc Surg 1988;95:685–91

49 Walsh TN, et al A comparison of multimodal therapy andsurgery for esophageal adenocarcinoma N Engl J Med1996;335:462–7

50 Roth JA, et al Randomized clinical trial of preoperative andpostoperative adjuvant chemotherapy with cisplatin, vinde-sine, and bleomycin for carcinoma of the esophagus JThorac Cardiovasc Surg 1988;96:242–8

51 Kelsen DP, et al Chemotherapy followed by surgerycompared with surgery alone for localized esophagealcancer N Engl J Med 1998;339:1979–84

52 Wang M, et al Randomized clinical trial on the tion of preoperative irradiation and surgery in the treat-ment of esophageal carcinoma: report on 206 patients Int JRadiat Oncol Biol Phys 1989;16:325–7

combina-53 Swisher SG, et al Salvage esophagectomy for recurrenttumors after definitive chemotherapy and radiotherapy JThorac Cardiovasc Surg 2002;123:175–83

54 Tanomkiat W, Galassi W Barium sulfate as contrastmedium for evaluation of postoperative anastomotic leaks.Acta Radiologica 2000;41:482–5

55 Buecker A, et al Esophageal perforation: comparison of use

of aqueous and barium-containing contrast media.Radiology 1997;202:683–6

56 Ghahremani GG, Gore RM Radiologic evaluation of theesophagus In: Shields TW, LoCicero J, Ponn RB, editors.General thoracic surgery Philadelphia (PA): Lippincott,Williams and Wilkins; 2000 p 1651–61

57 Gollub MJ, Bains MS Barium sulfate: a new (old) contrastagent for diagnosis of postoperative esophageal leaks.Radiology 1997;202:360–2

58 Orringer MB Complications of esophageal resection andreconstruction In: Waldhausen JA, Orringer MB, editors.Complications in cardiothoracic surgery St Louis (MO):Mosby-Year Book; 1991 p 354–69

518 / Advanced Therapy in Thoracic Surgery

59 Iannettoni MD, Whyte RI, Orringer MB Catastrophic

complications of the cervical esophagogastric anastomosis

J Thorac Cardiovasc Surg 1995;110:1493–500; discussion

1500–1

60 Jorgensen JO, Hunt DR Endoscopic drainage of esophageal

suture line leaks Am J Surg 1993;165:362–4

61 Whooley BP, et al Critical appraisal of the significance ofintrathoracic anastomotic leakage after esophagectomy forcancer Am J Surg 2001;181:198–203

Despite the numerous advances in the early diagnosis of

gastrointestinal malignancies, the majority of patients

have unresectable esophageal cancer at the time of

pre-sentation A variety of thermal, mechanical, and chemical

methods are available for the relief of symptoms of

dysphagia for these patients, including laser and argon

plasma ablation, chemotreatment, combined

chemother-apy and radiation treatment, repeated dilations, dilation

with placement of percutaneous endoscopic gastrotomy

tube, alcohol injection, and stent placement.

A detailed discussion of the above modalities is

be-yond the scope of this chapter Expandable metal stents

are increasingly being used for palliation of malignant

gastrointestinal obstruction Since the earliest reported

successful use of esophageal stent more than 100 years

ago, there has been remarkable advancement in stent

technology.1Today’s expandable metal stents are made of

superelastic alloys and are an effective palliative

treat-ment for malignant obstructive neoplasms in the

gastrointestinal tract This chapter discusses the

indica-tions, available technology, techniques, efficacy, and

complications related to the use of stents in malignant

esophageal obstruction.

History of Esophageal Stents

The word “stent” is named after Charles Stent, a dentist,

who in 1856 developed a thermoplastic material for

dental impressions subsequently used in stents The

earli-est successful esophageal stent placement found in the

literature was reported by Sir Charles Symonds of

England in 1885 He placed a 6-in tube of boxwood into

a patient’s esophagus, attached a silk thread to the

proxi-mal end of the boxwood, and wrapped it around the

patient’s ear to prevent downward migration.2 ment from boxwood to rigid tubes to tubes with flared ends and then Celestin tubes occurred in a fair time frame These rigid stents had many disadvantages includ- ing a high complication rate, thus the need for a new generation of stents, such as expandable stents, which could be delivered into the esophagus on a small diame- ter catheter, was logical The use of metal stent for malig- nant esophageal stenosis was first reported in the literature

Develop-by Frimberger in 1983.3The first through-the-scope metal stent insertion was done in 1990, with a stent that was developed for vascular use.4 The first-generation self- expanding metal stents consisted of bare metal barbs, with no membrane covering the struts of the stent The metal struts of these stents expand into the submucosa of the esophagus, and “epithelialization” occurs.5 The biggest drawback of these stents was tumor ingrowth through the mesh design of the stents Different mem- brane coverings, including silicon and polyurethane, were developed for these stents in hopes of preventing tumor ingrowth.6 Unfortunately, with the desired effect of preventing tumor ingrowth, the undesired effect of stent migration during or subsequent to stent placement necessitated further improvement in stent design Barbs projecting into the esophageal wall were added to the external surface of stents, thereby providing an anchor to prevent migration.7 The proximal and distal ends of stents were flared, the hypothesis being that wider proxi- mal and distal ends of the stents would migrate less frequently through the narrower lumen of the stenosis A new conical-shaped design (Flamingo stent) with vari- able radial force for use at the gastric cardia has been developed to prevent migration.8A partially covered stent was developed to combine the advantages of both

fully covered and uncovered wall stents It consists of

bare proximal and distal ends and covered central

portion The central portion covered by the membrane

prevents ingrowth, and the bare proximal and distal

portions of the stent become integrated into the

eso-phageal wall, thereby preventing migration A latest

vari-ation in stent design is the addition of a 7-centimeter

compressible valve (Dua antireflux valve) at the distal

end to preclude reflux when placed across the

gastro-esophageal junction (GEJ).9 In addition to the above

variations in stent design, various materials and alloys,

including Elgiloy, 316L stainless steel, and nitinol, are

used, exploiting their useful physical properties.8

Recently, a new self-expanding plastic stent has been

marketed for malignant esophageal obstruction.10

Indications for Stent Placement

Stent placement is indicated for the following:

1 palliation of dysphagia in a nonoperative

candi-date, a stent is indicated when dilation is

inef-fective or required too frequently; stent is

especially useful in patients with circumferential

narrowing caused by tumors and

extra-esophageal masses leading to compromise of the

esophageal lumen

2 malignant tracheoesophageal fistula

3 recurrence of tumors after definitive radiation,

chemotreatment of surgery

Contraindications for Stent Placement

Stent placement is absolutely contraindicated in

mori-bund patients with a very short life expectancy (less than

4 weeks) and in whom there is no prospect of improving

2 a noncircumferential stricture preventing

proper anchoring of the stent

3 angulated strictures

4 a completely obstructing stenosis that precludes

guidewire insertion prior to stent placement

5 significant airway compromise developing in

patients during dilation alone and airway

involvement with the tumor; airway stent

place-ment should be considered prior to attempted

esophageal stent placement11

6 intra-abdominal carcinomatoses when multiple

levels of intestinal obstruction are suspected

Types of Stents

Many different types of stents are available for use expanding stents are exclusively used at our institution and at most major centers across the country.

Self-Custom-Made Endoprostheses

Polymer tubing is used for making custom-fit stents Stent length is measured to be 4 to 5 cm longer than the tumor length The outside of the stent has spiral indenta- tions to make it less slippery, and the upper end is funneled to minimize leakage of liquids across the stent edge Radiopaque markers are added at either end of the stent, and in some cases the stent wall is strengthened by adding a plastic-embedded metal coil on the outside of the stent.12,13

Commercially Available Plastic Endoprostheses

There are a number of commercially available plastic stents The Wilson-Cook prosthesis (Wilson-Cook Medical, Winston-Salem, NC) is made of silicone and reinforced with metal coil This stent has an outer diameter of

16 mm and inner diameter of 12 mm and is available in varying lengths, from 4.4 to 16.4 cm.

KeyMed-Atkins prosthesis (KeyMed, Essex, UK) is made of silicon with a central nylon spiral to prevent kinking The distal end of the tube has a rounded shoul- der to prevent migration It is available in lengths from

14 to 19 cm, external diameter of 16 mm and inner diameter of 11.7 mm.

The ESKA-BUEB stent (ESKA, Germany) is made of silicon and a metal spiral is embedded within the wall to give stability It also has hooks integrated into the proxi- mal end of the stent to help reposition or retrieve the stent postdeployment The distal end of the stent completely collapses when passed through a stenosis.14

The Medoc-Celestin tube (Medoc, Tetbury, UK) is made of latex and reinforced with a nylon spiral It has a collapsible distal flange to prevent migration It is avail- able in varying lengths from 12.5 to 21 cm, with an outer diameter of 15 mm and inner diameter of 12 mm Owing to the high complication rate associated with the above-mentioned stents and the availability of effec- tive and relatively safe self-expandable stents, plastic stents are almost obsolete Following are the commer- cially available self-expandable stents.

Self-Expandable Metal Stents

Wallstent (Boston Scientific, Natick, MA) is made of Elgiloy wires woven into a tubular mesh design (Table 44-1).8The current Wallstent Esophageal II has distal flared ends, with Permalume silicone polymer covering the inner diameter of the stent, except a distal

522 / Advanced Therapy in Thoracic Surgery

centimeter on each side It is available in different lengths

(Table 44-2) In Europe, a new conical-shaped design

(Flamingo stent) is available for use at the gastric cardia.

This stent has a unique design with variable radial force

throughout the length of the stent; the proximal half has

greater radial force to help fixation of the stent in the

esophageal wall, and the distal half exerts less radial force

and greater flexibility to account for movements at the

gastric cardia (Figure 44-1).

The Ultraflex stent (Boston Scientific, Natick, MA) is

made of nitinol wires configured into circumferential

interwoven loops (see Table 44-1).1 It is available in

covered and uncovered versions; the covered version has

a single-layer translucent polyurethane membrane

cover-ing the entire length of the stent except for 1.5 cm at each

of the proximal and distal ends It is available in different

sizes and lengths (Figure 44-2).

The Z stent (Wilson-Cook Medical, Winston-Salem, NC) is made of a series of stainless steel wire cages bent into short segments in a zigzag configuration (see Table 44-1).7 It is also available in partially and fully covered versions A recent variation in the design of Z stent is the addition of a 7 cm collapsible valve at the distal end termed a “Dua antireflux valve” to prevent reflux across the GEJ.9This stent is available in different sizes and lengths (see Table 44-2).

The Esophagacoil stent (Medtronic/Instent; Prairie, MN) consists of a flattened nitinol wire in a

Eden-TABLE 44-1 Composition of Different Alloys Used in Stents

Adapted from McIntee BE.8

TABLE 44-2 Sizes and Properties of Different Expandable Stents Available in the United States

FIGURE 44-1 A, Tracheobronchial fistula B, After covered Wallstent

placement C, Radiologic view of Wallstent placement.

FIGURE 44-2 A, Tight proximal esophageal stricture B, After

Ultraflex stent placement C, Endoscopic view of Ultraflex stent in a

patient with tracheoesophageal fistula

A

B

C A

B

C

tightly wound coiled configuration (see Table 44-1) The

stent has a central section of adjacent coils and two 2 cm

flayed ends It is available in different sizes and lengths

(see Table 44-2).

Self-Expandable Plastic Stents

Recently, a self-expandable stent made of polyester weave

with silicone coating, the Polyflex stent, has been

marketed in the United States Experience with this stent

has been limited (Figure 44-3).10 It is also available in

various sizes and lengths (see Table 44-2).

Technique of Stent Insertion

Preprocedure Considerations

A thorough informed consent should include discussion

of alternatives to stent placement and risks of both the

dilatation and the stent insertion procedure (bleeding,

perforation, and adverse reaction to drugs) and

subse-quent complications related to stent (eg, chest pain,

misplacement, displacement, tumor overgrowth or

ingrowth, bleeding, perforation, and occlusion) In most

patients, intravenous sedation with a combination of

narcotics and benzodiazepines suffices In few

difficult-to-sedate patients (on oral narcotics for pain control),

monitored anesthesia care (MAC) may be needed If

possible, a barium swallow before stent placement should

be obtained for the shape, length, and exact location of the tumor and identifying tracheoesophageal fistula, if any This helps in preprocedure planning with the type and length of the stent to be used It is also important to perform the procedure with nurses and supporting personnel familiar with it.

we also mark the upper esophageal sphincter with contrast injection The distance of the stenosis is measured from the incisors using the markings on the shaft of endoscope A standard 0.035-inch or a 0.038- inch stiff guidewire (we prefer 0.038-inch) is then advanced through the endoscope and scope is removed The length of the stent is chosen so as to cover 2 to 3 cm

on each end of the stricture The stent assembly is duced over the guidewire and stent is positioned across the stricture and is deployed under fluoroscopic, combined fluoroscopic and endoscopic, or endoscopic guidance A thorough understanding of the mechanism

intro-of the stent deployment system and shortening enon is essential to ensure successful placement of esophageal endoprosthesis in a desired location If only fluoroscopic guidance is used for stent deployment, the endoscope can be reinserted to check for stent expansion.

phenom-We prefer not to pass the endoscope through the stent in the immediate postdeployment period for fear of dislodgment After stent placement, no radiographic or endoscopic study is necessary unless the patient becomes symptomatic again.

FIGURE 44-3 A, Esophageal tumor postradiation B, After Polyflex

stent placement C, Radiologic view of a Polyflex stent placed at

gastroesophageal junction

A

B

C