- Email: [email protected]
Epiphrenic diverticulum: Results of surgical treatment
Epiphrenic Diverticulum: Results of Surgical Treatment Joseph C. Benacci, MD, Claude Deschamps, MD, Victor F. Trastek, MD, Mark S. Allen, MD, Richard C. Daly, MD, and Peter C. Pairolero, MD Section of General Thoracic Surgery, Mayo Clinic and Mayo Foundation, Rochester, Minnesota
From 1975 to 1991,112 patients (64men and 48 women) were found to have an epiphrenic diverticulum. Symptoms were absent or minimal in 71 patients and incapacitating in 41. All patients with minimal symptoms were managed conservatively; 35 were available for follow-up, which ranged from 1 to 25 years (median, 9 years). None of these 35 patients had clinically significant progression of symptoms. Surgical repair was done in 33 patients with incapacitating symptoms. Achalasia was present in 8 of the surgical patients (24.2%), diffuse esophageal spasm in 3 (9.1%1, hypertensive lower esophageal sphincter alone in 1 (3.0%), and nonspecific motor abnormalities of the esophageal body in 7 (21.2%). Diverticulectomy and esophagomyotomy were performed in 22 patients, diverticulectomy alone in 7, esophageal resec-
tion in 3,and esophagomyotomy alone in 1. Concomitant hiatal hernia repair was done in 6 patients. Complications occurred in 11 patients; 6 had esophageal leaks. There were three operative deaths (9.1%), all occurring in patients with abnormal manometry. Follow-up was complete in 29 patients and ranged from 4 months to 15 years (median, 6.9 years). Long-term results were excellent in 14 patients (48.2%), good in 8 (27.6%), fair in 5 (17.2%), and poor in 2 (6.9%). We conclude that operation has significant risks and is not warranted in patients with minimal symptoms because progression is unlikely. Surgical treatment, however, is advisable in patients with incapacitating symptoms because most operative survivors will have long-term symptomatic palliation. (AWI rhorac surg z993;55m09-~4)
Diverticulum of the lower esophagus, or epiphrenic diverticulum, is rare. Most are acquired and occur as a mucosal pouch protruding through the muscularis wall. These pulsion pseudodiverticula are thought to be caused by an increase in intraesophageal pressure and may arise anywhere within the thoracic esophagus but are most common in the lower 10 cm [l, 21. An associated distal esophageal motility disorder is often present; the causative relationship with the diverticulum, however, is unclear [2, 31.
these concerns, we retrospectively reviewed our recent experience with epiphrenic diverticulum.
For editorial comment, see page 1067. The clinical manifestations of epiphrenic diverticulum are variable and unpredictable. No correlation exists between the size of the diverticulum and the severity of symptoms [2,4]. Likewise, distinguishing between symptoms caused by the diverticulum and those caused by an underlying motility disorder is difficult [5]. Consequently, management of patients with this condition is controversial [6]. Our past experience favored operative treatment only for symptomatic patients [l, 7-91. A more recent report, however, recommended operation for all patients [lo]. Controversy also exists as to whether an esophagomyotomy and antireflux procedure should be performed at the time of operative intervention [ll, 121. To address Presented at the Thirty-ninth Annual Meeting of the Southern Thoracic Surgical Association, Wesley Chapel, FL, Nov 5-7, 1992. Address reprint requests to Dr Descharnps, Mayo Clinic, 200 First St SW, Rochester, MN 55905.
0
1993 by The Society of Thoracic Surgeons
Material and Methods The records of all patients with an epiphrenic diverticulum documented by either barium swallow or esophagoscopy seen at our institution between November 1, 1975, and October 31, 1991, were reviewed. We analyzed each record for age, sex, type and duration of symptoms, diagnostic evaluation, method of management, operative and pathologic findings, and long-term functional results. Results were considered excellent if the patient was eating a general diet without symptoms, good if minimal symptoms were present and neither medication nor esophageal dilation was required, fair if symptoms were improved but either medication or esophageal dilation was required, and poor if symptoms were unchanged or worse. Operative mortality included those patients who died within the first 30 days after operation and those who died later but during the same hospitalization. The Fisher exact test was used to determine differences between groups (p < 0.05 was considered significant).
Clinical Findings There were 112 patients with epiphrenic diverticulum (64 men and 48 women). Forty-seven patients were asymptomatic at the time of our initial evaluation. Most of these diverticula were found as incidental roentgenographic abnormalities on upper gastrointestinal contrast studies 0003-4975/93/$6.00
1110
BENACCI ET AL EPIPHRENlC DIVERTICULUM
performed for other reasons. Twenty-four other patients had minimal symptoms, mostly intermittent dysphagia. Nine of these 24 patients had regurgitation, but only 1had aspiration. All 71 patients with absent or minimal symptoms were eating a general diet. Because these patients had minimal symptoms, further investigation was not done and repair of the diverticulum was not recommended. The remaining 41 patients all had incapacitating symptoms. All were evaluated by a surgeon. Three patients were considered high risk because of advanced age, cardiac status, or cancer and were not considered surgical candidates. Another 5 patients refused operation. The remaining 33 patients had surgical repair.
Surgical Evaluation Median age for the 33 patients (20 men and 13 women) undergoing surgical repair was 65 years (range, 37 to 82 years). Duration of symptoms ranged from 1 month to 30 years (median, 4 years). Dysphagia was present in 30 patients (90.9%), regurgitation in 27 (81.8%), repetitive episodes of aspiration in 10 (30.3%), and pyrosis in 8 (24.2%). Two of the patients with pyrosis were taking antireflux medication. Weight loss of more than 10 kg was observed in 4 patients (12.1%).Five patients (15.2%)had a prior upper gastrointestinal operation: 2 had transthoracic epiphrenic diverticulectomy, 2 had an antireflux procedure, and 1 had subtotal gastrectomy for duodenal ulcer disease. All of these procedures were performed elsewhere, and specific operative indications were not available for review. Barium roentgenographic examination of the esophagus, esophagoscopy, and esophageal manometry were performed in all 33 surgical patients. Barium swallow demonstrated a single diverticulum in 26 patients (78.8%), two diverticula in 5 (15.2%),three in 1 (3.0%),and four in 1 (3.0%).A concomitant sliding hiatal hernia was demonstrated in 16 patients (48.5%). Roentgenographic motor disturbances in the body of the esophagus were noted in 15 patients and were interpreted by the radiologist during fluoroscopy as achalasia in 9 patients and indeterminate motor abnormalities in 6. Stenosis of the distal esophagus was present in 15 patients (45.5%):in 13 at the gastroesophageal junction and in 2 at the level of the diverticulum. Esophagoscopy visualized the diverticulum in 30 patients (90.9%). Ulceration or inflammation of the diverticulum was noted in 7 patients (23.3%). Sixteen patients (48.5%)had retained food within the diverticulum. Obstruction of the esophagus from distention by a filled diverticulum occurred in 6 patients. Esophagitis of the distal esophagus was visualized in 2 patients, 1 of whom had a history of pyrosis. In 1 patient, a neoplasm was identified arising within a diverticulum known to be present for 10 years. At operation the neoplasm was found to be a malignant fibrous histiocytoma. Although manometric studies were performed in all patients, distal passage of the probe into the stomach was possible in only 20. Abnormal motility was documented in 19 of the 33 patients: 13 had abnormal motility limited to the body of the esophagus, 1 had abnormal motility
Ann Thorac Surg 1993:55:1109-14
limited to the lower esophageal sphincter (LES), and 5 had both esophageal and LES abnormalities. Nine of the 18 patients with abnormal motility in the body of the esophagus had loss of peristalsis, 3 had prolonged and high-amplitude tertiary wave contraction suggestive of diffuse esophageal spasm, and 7 had nonspecific motor abnormalities (NMA). Four of the 6 patients with abnormal LES motility had a hypertensive LES and 2 had incomplete relaxation of the LES. After preoperative evaluation 8 patients (24.2%)were considered to have achalasia, 3 (9.1%) to have diffuse esophageal spasm, 1 (3.0%)to have a hypertensive LES, and 7 (21.2%) to have NMA of the esophageal body. Hiatal hernia was documented in 16 patients (48.5%); 6 had pyrosis and 1 had esophagitis visualized on esophagoscopy. Four of the 6 patients with both hiatal hernia and pyrosis had an associated motility disorder (3 NMA and 1 achalasia). Three of the 10 patients with hiatal hernia but without pyrosis had motor abnormalities (1 achalasia, 1 diffuse esophageal spasm, and 1 NMA). Two patients (including the 1 already mentioned) had neoplasms.
Surgical Procedure The surgical approach was through a left thoracotomy in 32 patients and a right thoracotomy in 1. Our operative technique has been previously described [131. Diverticulectomy with esophagomyotomy was performed in 22 patients, diverticulectomy alone in 7, and esophagomyotomy alone in 1. The 7 patients who had diverticulectomy alone all had normal preoperative motility studies. The esophagus was resected in the remaining 3 patients, in 2 for benign esophageal strictures and in 1 for malignant fibrous histiocytoma. The 2 patients with strictures each had prior esophageal operations, 1 for hiatal hernia and 1 for epiphrenic diverticulum. Multiple diverticula were resected in 4 patients; the remaining 3 patients with multiple diverticula had only the largest diverticulum resected. The neck of the diverticulum was closed with a mechanical stapler in 27 patients and with sutures in 2. Concomitant hiatal hernia repair was performed in 6 patients. Indication for hernia repair was surgeon's preference in 5 patients and take-down of a functioning Nissen fundoplication at the time of diverticulectomy and esophagomyotomy in 1. Four patients had a Belsey antireflux procedure, 1 had a Nissen fundoplication, and 1 had an anatomic Allison repair. The diverticula ranged in size from 2 to 9 cm (median, 5 cm). Pathologic evaluation of the resected diverticula revealed mucosal inflammation or ulceration in 13 patients (39.4%).A leiomyoma surrounded the diverticulum in 1 patient, and 1 patient had a malignant fibrous histiocytoma arising in the wall of the diverticulum.
Results After operation all patients underwent a water-soluble contrast examination of the esophagus before resuming oral intake. If extravasation was absent, a liquid diet was started and advanced to a general diet. At hospital dis-
BENACCI ET AL EI'II'HRENIC DIVERTICULUM
Ann Thorac Surg 1993;55:1109-I4
charge, 22 patients were eating a general diet and 8 were eating a soft diet. The median length of hospitalization was 13 days (range, 6 to 36 days). Complications occurred in 11 patients (33.3%)and included esophageal leak in 6, pneumonia in 2, atrial fibrillation in 2, and central line sepsis in 1. All six leaks occurred in patients who had both diverticulectomy and esophagomyotomy. The stump of the diverticulum in these 6 patients was closed with staples in 4 and with sutures in 2. Four of the 6 patients had associated motility disorders. Three patients also had concomitant antireflux procedures, 1of whom had abnormal motility. Esophageal leaks occurred in 4 of 19 patients (21.1%) with abnormal motility and in 2 of 14 patients (14.3%) who had normal motility ( p = not significant). Excluding the 3 patients who had esophagogastrectomy, leaks occurred in 3 of 6 patients (50.0%) who had hiatal hernia repair and in 3 of 24 patients (12.5%)who did not have hernia repair ( p = 0.075). The esophageal leaks in 4 patients were asymptomatic and detected on routine postoperative roentgenographic examination. Each of these four leaks resolved spontaneously when oral feeding was withheld. Two of the asymptomatic leaks occurred in patients who had abnormal motility (diffuse esophageal spasm and NMA of the esophageal body). The remaining 2 patients with symptomatic leaks are described below. There were three perioperative deaths (operative mortality, 9.1%). All deaths occurred in patients with abnormal manometric findings. Each patient had both diverticulectomy and esophagomyotomy, and 2 had a concomitant hiatal hernia repair. Two of the deaths occurred in patients with clinically significant esophageal leaks. The first patient was a 77-year-old man whose motility demonstrated achalasia. This patient had diverticulectomy and esophagomyotomy. Because of persistent empyema and sepsis after esophageal leak, he underwent reoperation 2 weeks after repair. At reexploration the leak was localized to the diverticular stump, which was then reclosed and the lung decorticated. Subsequent esophageal roentgenograms demonstrated esophageal narrowing but no leak. Aspiration, however, occurred at the time of esophageal study, and the patient died 5 days later of persistent sepsis. At autopsy the leak was found to be closed, but a chronic empyema was documented. The second patient was a 68-year-old man whose motility was consistent with NMA and who had diverticulectomy, esophagomyotomy, and a Belsey antireflux procedure. This patient underwent reoperation with esophageal resection and diversion for an esophageal leak 6 days after primary repair. Reconstruction was performed with stomach at a third operation 17 days later. Subsequent esophageal roentgenograms did not demonstrate a leak. The patient, however, died unexpectedly 8 days after reconstruction. Autopsy demonstrated no intestinal leak and no site of infection. Cause of death was presumed to be cardiac arrhythmia. The third death occurred in a 73-year-old woman whose motility demonstrated prolonged LES contraction and who had a Belsey antireflux repair in addition to diverticulectomy and esophagomyotomy. Her postoperative course was
1111
complicated by respiratory failure, which gradually improved. She, however, aspirated during routine esophageal roentgenograms. No leak or obstruction of the distal esophagus was documented. Progressive pulmonary infiltrates developed, and the patient died on the ninth postoperative day. Autopsy demonstrated pneumonia and no leak.
Long-Term Follow-up Follow-up was complete in 29 of the 30 postoperative survivors (96.7%)and ranged from 4 months to 15 years (median, 6.9 years). Follow-up was obtained through telephone calls, letters to the patient's physician, and personal interviews. Diagnostic testing was performed when warranted by the patient's symptoms. There were six late deaths, none related to esophageal complications. Cause of death was pneumonia in 2 patients, congestive heart failure in 2, cancer in 1, and complications from rheumatoid arthritis in 1. All patients did well immediately after repair. All were eating a general diet. Weight gain of 11and 23 kg occurred in 2 of the 4 patients who presented with preoperative weight loss; no change in weight occurred in the other 2. Recurrent diverticulum was not known to have developed in any patient. Dysphagia subsequently developed in 5 patients (17.2%), 4 of whom had an associated motility disorder. All 5 patients required dilation 3 to 12 years after repair. Complete resolution of dysphagia occurred in 3 of these patients, each after a single dilation. The fourth patient had complete relief of dysphagia for the last 2 years of his life after four dilations. The last patient continues to have persistent dysphagia after six dilations. Nine patients had pyrosis (31.0%) during follow-up, 3 of whom had the symptom preoperatively. Antireflux medication was required in 4 of these patients. Five patients had diverticulectomy and esophagomyotomy alone, 2 had diverticulectomy alone, 1 had esophagomyotomy and hiatal hernia repair, and 1 had diverticulectomy, esophagomyotomy, and hiatal hernia repair. Pyrosis occurred in 2 of 5 patients (40.0%) who had both esophagomyotomy and hiatal hernia repair and in 5 of 17 patients (29.4%) who had esophagomyotomy without hiatal hernia repair ( p = not significant). One of the 9 patients had associated dysphagia. Overall, long-term results were excellent in 14 patients (48.2%),good in 8 (27.6%), fair in 5 (17.2%),and poor in 2 (6.9%)(Table 1).
Nonoperative Findings Seventy-one patients with epiphrenic diverticulum had absent or minimal symptoms and did not undergo surgical repair. Twenty-seven of the 47 patients who were asymptomatic when the epiphrenic diverticulum was first diagnosed were not seen again at our institution. The remaining 20 patients were followed up for a median of 4.0 years (range, 1 to 17 years). All remained asymptomatic during follow-up. Twenty-four patients had mild symptoms when initially evaluated, 6 of whom were found to have the diverticulum before symptoms developed. The median
1112
BENACCI ET AL EPIPHRENIC DIVERTICULUM
Ann Thorac Surg 1993;55:1109-14
Table 1 . Long-Term Results of 29 Patients Who Had Surgical Repair of an Epiphrenic Diverticulum Between 1976 and 1991 Procedure
Excellent
Good
Fair
Poor
Esophagomyotomy &
7
5
4"
2
Diverticulectomy alone Esophagomyotomy
4
2
lb
...
...
lb
...
...
3
...
...
...
diverticulectomy
alone
Esophageal resection
Two of 4 had concomitant hiatal hernia repair. hernia repair.
a
Concomitant hiatal
length of time the diverticulum was present before symptoms occurred in these 6 patients was 12.5 years (range, 1 to 21 years). Nine of the 24 patients were never evaluated again at our institution; the remaining 15, however, were followed up for a median of 11years (range, 1 to 25 years). All were serially evaluated with either upper gastrointestinal roentgenograms or esophagoscopy during this period. Symptoms persisted but were unchanged in severity in all 15 patients. No incapacitating symptoms developed. Altogether, 35 patients with epiphrenic diverticula who had either absent or minimal symptoms were followed up for a median of 6.9 years (range, 4 to 15 years). No patient had clinically significant progression of symptoms. One of the 8 patients presenting with incapacitating symptoms who either was not offered or declined surgical repair was lost to follow-up. The remaining 7 patients were followed up from 1 to 11 years (median, 5 years). All had worsening of their symptoms. Two patients died at 2 and 11 years, both of metastatic prostate cancer. Two other patients had subsequent resection of their diverticula at 5 and 10 years and now are asymptomatic. The remaining 3 patients required dilation at 1, 2, and 9 years.
Comment Epiphrenic diverticula are rare. However, the exact prevalence of this condition is unknown because asymptomatic patients are usually not discovered. Most of these diverticula are found in middle-aged or elderly patients, and male patients have a slight predominance. The relative incidence of epiphrenic to pharyngoesophageal diverticula at the Mayo Clinic during the past four decades has been 1 to 5 [13]. Mondiere [14], as early as 1833, postulated that pulsion diverticula were mucosal herniations occurring through the muscularis wall associated with some form of obstruction to swallowing. Although it is not surprising that symptoms were attributed solely to the saccular abnormality of the distal esophagus in the past, the role of esophageal motility disorders in the genesis of this condition was not implicated until the 1930s [15]. With the advent of manometric studies it is now increasingly evident that functional obstruction of the distal esophagus may be not only the cause of the diverticulum, but also a
major cause of symptoms. However, motility disorders are not found in every patient, and, when present, both the type of manometric disturbances and the severity of symptoms vary [l6-18]. Our current experience is consistent with these past observations, namely, that the cause of symptoms is multifactorial. Symptoms in patients with epiphrenic diverticula are variable. Many are asymptomatic and others have only mild dysphagia that is readily managed with simple methods such as thorough mastication and adequate fluids at mealtime. The majority of our patients (63%) were in this category, and the diverticulum was an incidental finding on barium swallow for unrelated reasons. Still other patients, however, have worsening and frequently incapacitating symptoms, most often severe dysphagia, chest pain, food retention, regurgitation, and subsequent aspiration. These latter symptoms may become life-threatening, as repeated episodes of pneumonia may result in progressive destruction of lung parenchyma. In our series, the ratio of patients with absent or only minimal symptoms to those with incapacitating symptoms was 1.7 to 1. All patients with suspected epiphrenic diverticulum should have barium upper gastrointestinal roentgenographic examination. Barium swallow provides proof of diagnosis, serves as a baseline if the patient is asymptomatic, provides clues to any associated motility disorder, and may detect other lesions, such as cancer or stricture, that are causing symptoms. Those patients with incapacitating symptoms should have further evaluation with both esophagoscopy and esophageal manometry. Esophagoscopy allows careful evaluation of the esophageal mucosa for esophagitis and the rare presence of cancer. However, as our patient with leiomyoma demonstrates, endoscopy is not 100% sensitive. Esophagoscopy may also be of value in removing retained debris from the sac before operation in patients with severe retention and regurgitation. Manometry is mandatory to define associated motility disorders. These manometric findings may help determine the length of esophagomyotomy required to relieve functional obstruction. However, manometry may underestimate the extent of abnormal motility because of the difficulty in passing the probe into the stomach, as was the situation in nearly 40%of our surgical patients. If gastroesophageal reflux is suspected, a 24hour pH study can also be performed [ll, 191 to confirm reflux before proceeding with an antireflux procedure on clinical findings alone. If not confirmed, the symptoms thought to be related to reflux may be caused by other conditions, such as abnormal motility or regurgitation of diverticular contents. The decision to proceed with surgical repair can be difficult. The surgeon must balance the risk of the procedure against the potential benefit when selecting surgical candidates. Although only half of our asymptomatic or mildly symptomatic patients had long-term follow-up available for review, progressive symptoms did not develop in any of them. Thus, we believe that patients with minimal symptoms should be managed conservatively and followed up at regular intervals. If symptoms are
BENACCI ET AL EPIPHRENIC DIVERTICULUM
Ann Thorac Surg 1993;55:1109-14
incapacitating, a s was t h e situation in more than one-third of o u r patients, an operation should be advised, if the patient is otherwise i n good health. Neither size nor d e p e n d e n t location of t h e diverticulum correlated with symptoms in o u r patients. O u r current treatment of choice is diverticulectomy combined with a long esophagomyotomy. The sac is mobilized and t h e diverticulectomy performed longitudinally over a 50F dilator. W e prefer t o use a stapling device and to close t h e muscular wall over t h e diverticular s t u m p . The esophagomyotomy should be performed opposite t h e site of t h e diverticulectomy and should be carried onto t h e stomach for a few millimeters a n d extended cephalad through all regions of the esophagus documented t o have abnormal motility. If motility is normal, t h e esophagomyotomy should be carried to a level above the diverticulum, which is usually between t h e inferior pulmonary vein a n d t h e arch of the aorta. Some surgeons have suggested that all patients undergoing an esophagomyotomy should have a concomitant antireflux procedure [ll,20, 211. W e do not routinely a d d a n antireflux procedure i n t h e absence of preoperative gastroesophageal reflux or hiatal hernia. If either is present, a less obstructive antireflux procedure, ie, Belsey mark IV, should be performed. The 18.2% incidence of postoperative esophageal leak in o u r series is cause for concern. All leaks occurred in patients who had both diverticulectomy a n d esophagomyotomy, and t h e leak rate was greater i n patients with abnormal motility. Although no significant difference in leak rate was apparent with t h e different types of surgical procedures, t h e rate was higher for patients who had concomitant hiatal hernia repair (50.0%) than for patients without hiatal hernia repair (12.5%). Despite this, we still believe that all patients should have an esophagomyoto m y regardless of manometric findings and that antireflux procedures should be used selectively. O u r mortality of 9.1% is high and associated with both abnormal motility and esophageal leak. Others have reported similiar findings [22]. W e infer from these data that every effort should be m a d e to identify a n d correct associated esophageal conditions to minimize postoperative complications. These data also reinforce o u r conclusion that epiphrenic diverticula should be surgically managed only in patients who h a v e incapacitating symptoms. In summary, s y m p t o m s i n patients with epiphrenic diverticula are variable. Patients with absent o r minimal symptoms should b e managed conservatively as progression of s y m p t o m s is unlikely. Patients with incapacitating s y m p t o m s should u n d e r g o diverticulectomy a n d esophagomyotomy. Operative risks, however, are substantial a n d portend the difficulties in performing multiple concomitant procedures on t h e esophagus. Nonetheless, long-term results are acceptable a n d durable. W e hope that further understanding of the pathophysiology of epi-
1113
phrenic diverticula will allow better selection of patients, reduced morbidity, and improved long-term results.
References 1. Habein HC, Kirklin JW, Clagett OT, Moersch HJ. Surgical treatment of lower esophageal pulsion diverticula. AMA Arch Surg 1956;72:1018-24.
2. Debas HT, Payne WS, Cameron AJ, Carlson HC. Physiopathology of lower esophageal diverticulum and its implications for treatment. Surg Gynecol Obstet 1980;151:593-600. 3. Bruggeman LL, Seaman WB. Epiphrenic diverticula: an analysis of 80 cases. Am J Roentgenol 1973;119:26&76. 4. Conrad C, Nissen F. Giant epiphrenic diverticula. Eur J Radio1 1982;2:48-9. 5. Ravich WWJ. Esophageal dysphagia. In: Groher ME, ed. Dysphagia: diagnosis and management, 2nd ed. Stoneham: Butterworth-Heineman, 1992:85-6. 6. Montesani C, DAmato A, Citone G, Procacciante F, Narilli P, Ribotta G. Surgical treatment of epiphrenic diverticula of the esophagus: is diverticulectomy always necessary? A report of two cases. ltal J Surg 1985;15:6%73. 7. Allen TH, Clagett OT. Changing concepts in the surgical treatment of pulsion diverticula of the lower esophagus. J Thorac Cardiovasc Surg 1965;50455-62. 8. Habein HC, Moersch HJ, Kirklin JW. Diverticula of the lower part of the esophagus: a clinical study of 140 nonsurgical cases. Arch Intern Med 1956;97:768-77. 9. Harrington SW. The surgical treatment of pulsion diverticula of the thoracic esophagus. Ann Surg 1949;129:606-18. 10. Altorki NK, Sunagawa M, Skinner DB. Thoracic esophageal diverticula: why is operation necessary? J Thorac Cardiovasc Surg 1993;105:260-4. 11. Evander A, Little AG, Ferguson MK, Skinner DB. Diverticula of the mid- and lower esophagus: pathogenesis and surgical management. World J Surg 1986;10:820-8. 12. Duda M, S e q 2, Vojaeek K, Rotek V, Rehulka M. Etiopathogenesis and classification of esouhageal diverticula. Int Sura "
-1985;70:291-5.
I
"
13. Trastek VF. Pavne WS. EsoDhaeeal diverticula. In: Shields 14. 15. 16.
17.
18.
TW, ed. Geneial thoracic &gky. Philadelphia: Lea and Febiger, 1989:989-1001. Mondiere JT. Notes sur quelques maladies de I'oesophage. Arch Gen Med Paris 1833;3:28-65. Vinson PP. Diverticula of the thoracic portion of the esophagus: report of 42 cases. Arch Otolaryngol 1934;19:508-13. Bontempo I, Corazziari E, Mineo TC, Tosoni M, Anzini F, Ricci C. Esophageal motor activity in patients with esophageal diverticula. In: DeMeester TR, Skinner DB, eds. Esophageal disorders, pathophysiology and therapy. New York: Raven Press, 1985:427-9. Dodds WJ, Stef JJ, Hogan WJ, Hoke SE, Stewart ET, Arndorfer RC. Dstribution of esophageal peristaltic pressure in normal subjects and patients with esophageal diverticulum. Gastroenterology 1975;69:58&90. Kaye MD. Oesophageal motor dysfunction in patients with diverticula of the mid-thoracic oesophagus. Thorax 1974;
29:66&72. 19. Viard H. Sala JJ, Favre JP, Bernard A, Cougard P. Le
traitement chirurgical des diverticules de pulsion de I'oesophage. J Chir (Paris) 1987;124:658-62. 20. Belsey R. Functional disease of the esophagus. J Thorac Cardiovasc Surg 1966;52:1W8. 21. Little AG, Soriano A, Ferguson MK, Winans CS, Skinner DB. Surgical treatment of achalasia: results with esophagomyotoma and Belsey repair. Ann Thorac Surg 1988;45:489-94. 22. Fekete F, Vonns C. Surgical management of esophageal thoracic diverticula. Hepatogastroenterology 1992;39:97-9.
1114
BENACCI ET AL EFIPHRENIC DIVERTICULUM
Ann Thorac Surg 1993;55:1109-14
DISCUSSION DR JOSEPH I. MILLER, JR (Atlanta, GA): I want to thank the authors for the opportunity to review the manuscript before presentation. As usual, Dr Benacci and Dr Deschamps are to be congratulated on their presentation. We have all come to expect that from the group at the Mayo Clinic. I have several comments. We would agree that all of their patients have been adequately worked up with appropriate manometrics, barium swallow, and endoscopy; however the complication rate is quite high, with six leaks, or 18% of the total group, with three deaths having resulted. This would actually be six leaks in 30 patients who underwent the repair excluding the 3 who had esophagogastrectomy. Also noted is that only 22 of their 33 patients actually had a myotomy, although I have the feeling that the Mayo Clinic must have changed recently, in that I noticed in their conclusions myotomy is recommended in all patients. Also, only 6 patients had a hiatal hernia repair. Having been at the Mayo Clinic 1 am well aware of their tendency not to do a concomitant hiatal hernia repair when a myotomy is done even in their patients with achalasia. At Emory University, Dr Mansour and I have had a smaller experience with only approximately 22 patients operated on with epiphrenic diverticulum; however, we have performed a diverticulectomy plus a modified Heller myotomy and modified Belsey repair in all patients in the last 17 years. We have had no leaks and no operative mortality in this group of patients. It is our impression that all patients should undergo a myotomy with their combined diverticulectomy, and in addition we would recommend performance of a modified Belsey at the time of repair. My questions to Dr Benacci are, do you still think that a concomitant hiatal hernia repair is contraindicated in this group of patients, in the majority of the patients? Perhaps the cause of your increased leak rate is an inadequate myotomy without complete relief of the increased pressure zone in the distal esophagus. This could be caused by inadequate mobilization at the area of the gastroesophageal junction. DR MILTON V. DAVIS (Kaufman, TX): This is a very beautiful report, and I agree completely with what Dr Miller said except for one thing. I would do today a different kind of reflux-preventing operation. This report is flawed, but not because they have reported their results honestly, which we appreciate. The fact is that the epiphrenic diverticulum is like the part of the iceberg that sticks out of the water. It is not the problem, and it should not be treated as the problem. It is the result of the problem. Another point 1 want to make is that there are terrific problems when surgeons allow the gastroenterology service to select these patients. Now, if that is the only place you are going to get your patients, you get them wherever you can, and I do not fault that. But you said you evaluated them by gastrointestinal studies, by scoping, and by manometry. What about a history? A history is what you need to evaluate these patients; it is what you need to decide on their treatment. I have one last point. The word stricture is kicked around in our literature time and time again. Our guest speaker from Switzerland, Dr Naef, reported before this Society when Hawley Seiler invited him first to be a guest speaker about the failure of adequate diagnosing of esophageal strictures. I will not go into it any farther, but that is my point.
DR KAMAL A. MANSOUR (Atlanta, GA): This is a nice and really important report for the practicing general thoracic surgeon. One is, however, struck with the high 24% incidence of esophageal leak postoperatively. I have a statement and two questions for you. The statement is, although these diverticula present more frequently on the right side, it is our practice, as it is yours also, to perform the diverticulectomy, the distal myotomy, and the Belsey mark IV through a left thoracic approach. Failure to adhere to this so-called triple-treat principle led one of the thoracic surgeons in our community to perform only diverticulectomy through a right thoracotomy, which was followed by massive and persistent esophageal leak for 7 weeks postoperatively before it was repaired through a left thoracic approach and by adding the myotomy and the Belsey mark IV. Now, I have two questions. Did you find any relationship between the incidence of leak postoperatively in patients who underwent diverticulectomy alone as compared with those who had myotomy in addition to excision of their diverticulum? Second I noticed that you worked up your patients with barium swallow, endoscopy, and manometry. How about 24-hour pH studies; are you going to d o them in the future? DR THOMAS M. EGAN (Chapel Hill, NC): I enjoyed your report. 1 think Dr Mansour asked my question, but occasionally problems provide you with opportunities, and you have an opportunity here to analyze your esophageal leak rate. Specifically, was there a relationship between esophageal leak and the absence of a myotomy, or were there any other factors that you identified that contributed to your esophageal leak rate? DR BENACCI: Doctor Miller, our patients spanned 16 years, and the operative procedure undoubtedly evolved over this time. We certainly d o not believe that hiatal hernia repair is contraindicated, but we do not perform routine concomitant hiatal hernia repair. Those patients who have a sliding hiatal hernia documented by upper gastrointestinal studies or those who have a compatible history of gastroesophageal reflux supported by 24hour pH monitoring should have a fundoplication. Because an associated motility disorder is invariably present, we favor an incomplete wrap, such as a Belsey repair. We also had no evidence that the cause of leak was an inadequate esophagomyotomy. Quite the contrary, the highest leak rates occurred in patients who had diverticulectomy, esophagomyotomy, and hiatal hernia repair. The esophagomyotomy should extend onto the stomach for 0.5 to 1 cm. Doctor Davis, most patients with epiphrenic diverticulum are referred through our gastroenterologists. We agree entirely that history is extremely important. If the patient is asymptomatic, the odds are great that the diverticulum will not be discovered. We also believe that only symptomatic patients should have their diverticulum repaired. As you stated, the diverticulum is not the cause of the problem, and symptomatic patients need further investigation as an underlying motility disorder is often present. Doctor Mansour, diverticulectomy alone was only done in 7 patients; leaks did not develop in any of them. All of our leaks occurred in patients who had both diverticulectomy and esophagomyotomy. Our current practice is to perform a 24-hour pH study only in those patients who have a history suggestive of pyrosis or who have a diaphragmatic hernia.
From 1975 to 1991,112 patients (64men and 48 women) were found to have an epiphrenic diverticulum. Symptoms were absent or minimal in 71 patients and incapacitating in 41. All patients with minimal symptoms were managed conservatively; 35 were available for follow-up, which ranged from 1 to 25 years (median, 9 years). None of these 35 patients had clinically significant progression of symptoms. Surgical repair was done in 33 patients with incapacitating symptoms. Achalasia was present in 8 of the surgical patients (24.2%), diffuse esophageal spasm in 3 (9.1%1, hypertensive lower esophageal sphincter alone in 1 (3.0%), and nonspecific motor abnormalities of the esophageal body in 7 (21.2%). Diverticulectomy and esophagomyotomy were performed in 22 patients, diverticulectomy alone in 7, esophageal resec-
tion in 3,and esophagomyotomy alone in 1. Concomitant hiatal hernia repair was done in 6 patients. Complications occurred in 11 patients; 6 had esophageal leaks. There were three operative deaths (9.1%), all occurring in patients with abnormal manometry. Follow-up was complete in 29 patients and ranged from 4 months to 15 years (median, 6.9 years). Long-term results were excellent in 14 patients (48.2%), good in 8 (27.6%), fair in 5 (17.2%), and poor in 2 (6.9%). We conclude that operation has significant risks and is not warranted in patients with minimal symptoms because progression is unlikely. Surgical treatment, however, is advisable in patients with incapacitating symptoms because most operative survivors will have long-term symptomatic palliation. (AWI rhorac surg z993;55m09-~4)
Diverticulum of the lower esophagus, or epiphrenic diverticulum, is rare. Most are acquired and occur as a mucosal pouch protruding through the muscularis wall. These pulsion pseudodiverticula are thought to be caused by an increase in intraesophageal pressure and may arise anywhere within the thoracic esophagus but are most common in the lower 10 cm [l, 21. An associated distal esophageal motility disorder is often present; the causative relationship with the diverticulum, however, is unclear [2, 31.
these concerns, we retrospectively reviewed our recent experience with epiphrenic diverticulum.
For editorial comment, see page 1067. The clinical manifestations of epiphrenic diverticulum are variable and unpredictable. No correlation exists between the size of the diverticulum and the severity of symptoms [2,4]. Likewise, distinguishing between symptoms caused by the diverticulum and those caused by an underlying motility disorder is difficult [5]. Consequently, management of patients with this condition is controversial [6]. Our past experience favored operative treatment only for symptomatic patients [l, 7-91. A more recent report, however, recommended operation for all patients [lo]. Controversy also exists as to whether an esophagomyotomy and antireflux procedure should be performed at the time of operative intervention [ll, 121. To address Presented at the Thirty-ninth Annual Meeting of the Southern Thoracic Surgical Association, Wesley Chapel, FL, Nov 5-7, 1992. Address reprint requests to Dr Descharnps, Mayo Clinic, 200 First St SW, Rochester, MN 55905.
0
1993 by The Society of Thoracic Surgeons
Material and Methods The records of all patients with an epiphrenic diverticulum documented by either barium swallow or esophagoscopy seen at our institution between November 1, 1975, and October 31, 1991, were reviewed. We analyzed each record for age, sex, type and duration of symptoms, diagnostic evaluation, method of management, operative and pathologic findings, and long-term functional results. Results were considered excellent if the patient was eating a general diet without symptoms, good if minimal symptoms were present and neither medication nor esophageal dilation was required, fair if symptoms were improved but either medication or esophageal dilation was required, and poor if symptoms were unchanged or worse. Operative mortality included those patients who died within the first 30 days after operation and those who died later but during the same hospitalization. The Fisher exact test was used to determine differences between groups (p < 0.05 was considered significant).
Clinical Findings There were 112 patients with epiphrenic diverticulum (64 men and 48 women). Forty-seven patients were asymptomatic at the time of our initial evaluation. Most of these diverticula were found as incidental roentgenographic abnormalities on upper gastrointestinal contrast studies 0003-4975/93/$6.00
1110
BENACCI ET AL EPIPHRENlC DIVERTICULUM
performed for other reasons. Twenty-four other patients had minimal symptoms, mostly intermittent dysphagia. Nine of these 24 patients had regurgitation, but only 1had aspiration. All 71 patients with absent or minimal symptoms were eating a general diet. Because these patients had minimal symptoms, further investigation was not done and repair of the diverticulum was not recommended. The remaining 41 patients all had incapacitating symptoms. All were evaluated by a surgeon. Three patients were considered high risk because of advanced age, cardiac status, or cancer and were not considered surgical candidates. Another 5 patients refused operation. The remaining 33 patients had surgical repair.
Surgical Evaluation Median age for the 33 patients (20 men and 13 women) undergoing surgical repair was 65 years (range, 37 to 82 years). Duration of symptoms ranged from 1 month to 30 years (median, 4 years). Dysphagia was present in 30 patients (90.9%), regurgitation in 27 (81.8%), repetitive episodes of aspiration in 10 (30.3%), and pyrosis in 8 (24.2%). Two of the patients with pyrosis were taking antireflux medication. Weight loss of more than 10 kg was observed in 4 patients (12.1%).Five patients (15.2%)had a prior upper gastrointestinal operation: 2 had transthoracic epiphrenic diverticulectomy, 2 had an antireflux procedure, and 1 had subtotal gastrectomy for duodenal ulcer disease. All of these procedures were performed elsewhere, and specific operative indications were not available for review. Barium roentgenographic examination of the esophagus, esophagoscopy, and esophageal manometry were performed in all 33 surgical patients. Barium swallow demonstrated a single diverticulum in 26 patients (78.8%), two diverticula in 5 (15.2%),three in 1 (3.0%),and four in 1 (3.0%).A concomitant sliding hiatal hernia was demonstrated in 16 patients (48.5%). Roentgenographic motor disturbances in the body of the esophagus were noted in 15 patients and were interpreted by the radiologist during fluoroscopy as achalasia in 9 patients and indeterminate motor abnormalities in 6. Stenosis of the distal esophagus was present in 15 patients (45.5%):in 13 at the gastroesophageal junction and in 2 at the level of the diverticulum. Esophagoscopy visualized the diverticulum in 30 patients (90.9%). Ulceration or inflammation of the diverticulum was noted in 7 patients (23.3%). Sixteen patients (48.5%)had retained food within the diverticulum. Obstruction of the esophagus from distention by a filled diverticulum occurred in 6 patients. Esophagitis of the distal esophagus was visualized in 2 patients, 1 of whom had a history of pyrosis. In 1 patient, a neoplasm was identified arising within a diverticulum known to be present for 10 years. At operation the neoplasm was found to be a malignant fibrous histiocytoma. Although manometric studies were performed in all patients, distal passage of the probe into the stomach was possible in only 20. Abnormal motility was documented in 19 of the 33 patients: 13 had abnormal motility limited to the body of the esophagus, 1 had abnormal motility
Ann Thorac Surg 1993:55:1109-14
limited to the lower esophageal sphincter (LES), and 5 had both esophageal and LES abnormalities. Nine of the 18 patients with abnormal motility in the body of the esophagus had loss of peristalsis, 3 had prolonged and high-amplitude tertiary wave contraction suggestive of diffuse esophageal spasm, and 7 had nonspecific motor abnormalities (NMA). Four of the 6 patients with abnormal LES motility had a hypertensive LES and 2 had incomplete relaxation of the LES. After preoperative evaluation 8 patients (24.2%)were considered to have achalasia, 3 (9.1%) to have diffuse esophageal spasm, 1 (3.0%)to have a hypertensive LES, and 7 (21.2%) to have NMA of the esophageal body. Hiatal hernia was documented in 16 patients (48.5%); 6 had pyrosis and 1 had esophagitis visualized on esophagoscopy. Four of the 6 patients with both hiatal hernia and pyrosis had an associated motility disorder (3 NMA and 1 achalasia). Three of the 10 patients with hiatal hernia but without pyrosis had motor abnormalities (1 achalasia, 1 diffuse esophageal spasm, and 1 NMA). Two patients (including the 1 already mentioned) had neoplasms.
Surgical Procedure The surgical approach was through a left thoracotomy in 32 patients and a right thoracotomy in 1. Our operative technique has been previously described [131. Diverticulectomy with esophagomyotomy was performed in 22 patients, diverticulectomy alone in 7, and esophagomyotomy alone in 1. The 7 patients who had diverticulectomy alone all had normal preoperative motility studies. The esophagus was resected in the remaining 3 patients, in 2 for benign esophageal strictures and in 1 for malignant fibrous histiocytoma. The 2 patients with strictures each had prior esophageal operations, 1 for hiatal hernia and 1 for epiphrenic diverticulum. Multiple diverticula were resected in 4 patients; the remaining 3 patients with multiple diverticula had only the largest diverticulum resected. The neck of the diverticulum was closed with a mechanical stapler in 27 patients and with sutures in 2. Concomitant hiatal hernia repair was performed in 6 patients. Indication for hernia repair was surgeon's preference in 5 patients and take-down of a functioning Nissen fundoplication at the time of diverticulectomy and esophagomyotomy in 1. Four patients had a Belsey antireflux procedure, 1 had a Nissen fundoplication, and 1 had an anatomic Allison repair. The diverticula ranged in size from 2 to 9 cm (median, 5 cm). Pathologic evaluation of the resected diverticula revealed mucosal inflammation or ulceration in 13 patients (39.4%).A leiomyoma surrounded the diverticulum in 1 patient, and 1 patient had a malignant fibrous histiocytoma arising in the wall of the diverticulum.
Results After operation all patients underwent a water-soluble contrast examination of the esophagus before resuming oral intake. If extravasation was absent, a liquid diet was started and advanced to a general diet. At hospital dis-
BENACCI ET AL EI'II'HRENIC DIVERTICULUM
Ann Thorac Surg 1993;55:1109-I4
charge, 22 patients were eating a general diet and 8 were eating a soft diet. The median length of hospitalization was 13 days (range, 6 to 36 days). Complications occurred in 11 patients (33.3%)and included esophageal leak in 6, pneumonia in 2, atrial fibrillation in 2, and central line sepsis in 1. All six leaks occurred in patients who had both diverticulectomy and esophagomyotomy. The stump of the diverticulum in these 6 patients was closed with staples in 4 and with sutures in 2. Four of the 6 patients had associated motility disorders. Three patients also had concomitant antireflux procedures, 1of whom had abnormal motility. Esophageal leaks occurred in 4 of 19 patients (21.1%) with abnormal motility and in 2 of 14 patients (14.3%) who had normal motility ( p = not significant). Excluding the 3 patients who had esophagogastrectomy, leaks occurred in 3 of 6 patients (50.0%) who had hiatal hernia repair and in 3 of 24 patients (12.5%)who did not have hernia repair ( p = 0.075). The esophageal leaks in 4 patients were asymptomatic and detected on routine postoperative roentgenographic examination. Each of these four leaks resolved spontaneously when oral feeding was withheld. Two of the asymptomatic leaks occurred in patients who had abnormal motility (diffuse esophageal spasm and NMA of the esophageal body). The remaining 2 patients with symptomatic leaks are described below. There were three perioperative deaths (operative mortality, 9.1%). All deaths occurred in patients with abnormal manometric findings. Each patient had both diverticulectomy and esophagomyotomy, and 2 had a concomitant hiatal hernia repair. Two of the deaths occurred in patients with clinically significant esophageal leaks. The first patient was a 77-year-old man whose motility demonstrated achalasia. This patient had diverticulectomy and esophagomyotomy. Because of persistent empyema and sepsis after esophageal leak, he underwent reoperation 2 weeks after repair. At reexploration the leak was localized to the diverticular stump, which was then reclosed and the lung decorticated. Subsequent esophageal roentgenograms demonstrated esophageal narrowing but no leak. Aspiration, however, occurred at the time of esophageal study, and the patient died 5 days later of persistent sepsis. At autopsy the leak was found to be closed, but a chronic empyema was documented. The second patient was a 68-year-old man whose motility was consistent with NMA and who had diverticulectomy, esophagomyotomy, and a Belsey antireflux procedure. This patient underwent reoperation with esophageal resection and diversion for an esophageal leak 6 days after primary repair. Reconstruction was performed with stomach at a third operation 17 days later. Subsequent esophageal roentgenograms did not demonstrate a leak. The patient, however, died unexpectedly 8 days after reconstruction. Autopsy demonstrated no intestinal leak and no site of infection. Cause of death was presumed to be cardiac arrhythmia. The third death occurred in a 73-year-old woman whose motility demonstrated prolonged LES contraction and who had a Belsey antireflux repair in addition to diverticulectomy and esophagomyotomy. Her postoperative course was
1111
complicated by respiratory failure, which gradually improved. She, however, aspirated during routine esophageal roentgenograms. No leak or obstruction of the distal esophagus was documented. Progressive pulmonary infiltrates developed, and the patient died on the ninth postoperative day. Autopsy demonstrated pneumonia and no leak.
Long-Term Follow-up Follow-up was complete in 29 of the 30 postoperative survivors (96.7%)and ranged from 4 months to 15 years (median, 6.9 years). Follow-up was obtained through telephone calls, letters to the patient's physician, and personal interviews. Diagnostic testing was performed when warranted by the patient's symptoms. There were six late deaths, none related to esophageal complications. Cause of death was pneumonia in 2 patients, congestive heart failure in 2, cancer in 1, and complications from rheumatoid arthritis in 1. All patients did well immediately after repair. All were eating a general diet. Weight gain of 11and 23 kg occurred in 2 of the 4 patients who presented with preoperative weight loss; no change in weight occurred in the other 2. Recurrent diverticulum was not known to have developed in any patient. Dysphagia subsequently developed in 5 patients (17.2%), 4 of whom had an associated motility disorder. All 5 patients required dilation 3 to 12 years after repair. Complete resolution of dysphagia occurred in 3 of these patients, each after a single dilation. The fourth patient had complete relief of dysphagia for the last 2 years of his life after four dilations. The last patient continues to have persistent dysphagia after six dilations. Nine patients had pyrosis (31.0%) during follow-up, 3 of whom had the symptom preoperatively. Antireflux medication was required in 4 of these patients. Five patients had diverticulectomy and esophagomyotomy alone, 2 had diverticulectomy alone, 1 had esophagomyotomy and hiatal hernia repair, and 1 had diverticulectomy, esophagomyotomy, and hiatal hernia repair. Pyrosis occurred in 2 of 5 patients (40.0%) who had both esophagomyotomy and hiatal hernia repair and in 5 of 17 patients (29.4%) who had esophagomyotomy without hiatal hernia repair ( p = not significant). One of the 9 patients had associated dysphagia. Overall, long-term results were excellent in 14 patients (48.2%),good in 8 (27.6%), fair in 5 (17.2%),and poor in 2 (6.9%)(Table 1).
Nonoperative Findings Seventy-one patients with epiphrenic diverticulum had absent or minimal symptoms and did not undergo surgical repair. Twenty-seven of the 47 patients who were asymptomatic when the epiphrenic diverticulum was first diagnosed were not seen again at our institution. The remaining 20 patients were followed up for a median of 4.0 years (range, 1 to 17 years). All remained asymptomatic during follow-up. Twenty-four patients had mild symptoms when initially evaluated, 6 of whom were found to have the diverticulum before symptoms developed. The median
1112
BENACCI ET AL EPIPHRENIC DIVERTICULUM
Ann Thorac Surg 1993;55:1109-14
Table 1 . Long-Term Results of 29 Patients Who Had Surgical Repair of an Epiphrenic Diverticulum Between 1976 and 1991 Procedure
Excellent
Good
Fair
Poor
Esophagomyotomy &
7
5
4"
2
Diverticulectomy alone Esophagomyotomy
4
2
lb
...
...
lb
...
...
3
...
...
...
diverticulectomy
alone
Esophageal resection
Two of 4 had concomitant hiatal hernia repair. hernia repair.
a
Concomitant hiatal
length of time the diverticulum was present before symptoms occurred in these 6 patients was 12.5 years (range, 1 to 21 years). Nine of the 24 patients were never evaluated again at our institution; the remaining 15, however, were followed up for a median of 11years (range, 1 to 25 years). All were serially evaluated with either upper gastrointestinal roentgenograms or esophagoscopy during this period. Symptoms persisted but were unchanged in severity in all 15 patients. No incapacitating symptoms developed. Altogether, 35 patients with epiphrenic diverticula who had either absent or minimal symptoms were followed up for a median of 6.9 years (range, 4 to 15 years). No patient had clinically significant progression of symptoms. One of the 8 patients presenting with incapacitating symptoms who either was not offered or declined surgical repair was lost to follow-up. The remaining 7 patients were followed up from 1 to 11 years (median, 5 years). All had worsening of their symptoms. Two patients died at 2 and 11 years, both of metastatic prostate cancer. Two other patients had subsequent resection of their diverticula at 5 and 10 years and now are asymptomatic. The remaining 3 patients required dilation at 1, 2, and 9 years.
Comment Epiphrenic diverticula are rare. However, the exact prevalence of this condition is unknown because asymptomatic patients are usually not discovered. Most of these diverticula are found in middle-aged or elderly patients, and male patients have a slight predominance. The relative incidence of epiphrenic to pharyngoesophageal diverticula at the Mayo Clinic during the past four decades has been 1 to 5 [13]. Mondiere [14], as early as 1833, postulated that pulsion diverticula were mucosal herniations occurring through the muscularis wall associated with some form of obstruction to swallowing. Although it is not surprising that symptoms were attributed solely to the saccular abnormality of the distal esophagus in the past, the role of esophageal motility disorders in the genesis of this condition was not implicated until the 1930s [15]. With the advent of manometric studies it is now increasingly evident that functional obstruction of the distal esophagus may be not only the cause of the diverticulum, but also a
major cause of symptoms. However, motility disorders are not found in every patient, and, when present, both the type of manometric disturbances and the severity of symptoms vary [l6-18]. Our current experience is consistent with these past observations, namely, that the cause of symptoms is multifactorial. Symptoms in patients with epiphrenic diverticula are variable. Many are asymptomatic and others have only mild dysphagia that is readily managed with simple methods such as thorough mastication and adequate fluids at mealtime. The majority of our patients (63%) were in this category, and the diverticulum was an incidental finding on barium swallow for unrelated reasons. Still other patients, however, have worsening and frequently incapacitating symptoms, most often severe dysphagia, chest pain, food retention, regurgitation, and subsequent aspiration. These latter symptoms may become life-threatening, as repeated episodes of pneumonia may result in progressive destruction of lung parenchyma. In our series, the ratio of patients with absent or only minimal symptoms to those with incapacitating symptoms was 1.7 to 1. All patients with suspected epiphrenic diverticulum should have barium upper gastrointestinal roentgenographic examination. Barium swallow provides proof of diagnosis, serves as a baseline if the patient is asymptomatic, provides clues to any associated motility disorder, and may detect other lesions, such as cancer or stricture, that are causing symptoms. Those patients with incapacitating symptoms should have further evaluation with both esophagoscopy and esophageal manometry. Esophagoscopy allows careful evaluation of the esophageal mucosa for esophagitis and the rare presence of cancer. However, as our patient with leiomyoma demonstrates, endoscopy is not 100% sensitive. Esophagoscopy may also be of value in removing retained debris from the sac before operation in patients with severe retention and regurgitation. Manometry is mandatory to define associated motility disorders. These manometric findings may help determine the length of esophagomyotomy required to relieve functional obstruction. However, manometry may underestimate the extent of abnormal motility because of the difficulty in passing the probe into the stomach, as was the situation in nearly 40%of our surgical patients. If gastroesophageal reflux is suspected, a 24hour pH study can also be performed [ll, 191 to confirm reflux before proceeding with an antireflux procedure on clinical findings alone. If not confirmed, the symptoms thought to be related to reflux may be caused by other conditions, such as abnormal motility or regurgitation of diverticular contents. The decision to proceed with surgical repair can be difficult. The surgeon must balance the risk of the procedure against the potential benefit when selecting surgical candidates. Although only half of our asymptomatic or mildly symptomatic patients had long-term follow-up available for review, progressive symptoms did not develop in any of them. Thus, we believe that patients with minimal symptoms should be managed conservatively and followed up at regular intervals. If symptoms are
BENACCI ET AL EPIPHRENIC DIVERTICULUM
Ann Thorac Surg 1993;55:1109-14
incapacitating, a s was t h e situation in more than one-third of o u r patients, an operation should be advised, if the patient is otherwise i n good health. Neither size nor d e p e n d e n t location of t h e diverticulum correlated with symptoms in o u r patients. O u r current treatment of choice is diverticulectomy combined with a long esophagomyotomy. The sac is mobilized and t h e diverticulectomy performed longitudinally over a 50F dilator. W e prefer t o use a stapling device and to close t h e muscular wall over t h e diverticular s t u m p . The esophagomyotomy should be performed opposite t h e site of t h e diverticulectomy and should be carried onto t h e stomach for a few millimeters a n d extended cephalad through all regions of the esophagus documented t o have abnormal motility. If motility is normal, t h e esophagomyotomy should be carried to a level above the diverticulum, which is usually between t h e inferior pulmonary vein a n d t h e arch of the aorta. Some surgeons have suggested that all patients undergoing an esophagomyotomy should have a concomitant antireflux procedure [ll,20, 211. W e do not routinely a d d a n antireflux procedure i n t h e absence of preoperative gastroesophageal reflux or hiatal hernia. If either is present, a less obstructive antireflux procedure, ie, Belsey mark IV, should be performed. The 18.2% incidence of postoperative esophageal leak in o u r series is cause for concern. All leaks occurred in patients who had both diverticulectomy a n d esophagomyotomy, and t h e leak rate was greater i n patients with abnormal motility. Although no significant difference in leak rate was apparent with t h e different types of surgical procedures, t h e rate was higher for patients who had concomitant hiatal hernia repair (50.0%) than for patients without hiatal hernia repair (12.5%). Despite this, we still believe that all patients should have an esophagomyoto m y regardless of manometric findings and that antireflux procedures should be used selectively. O u r mortality of 9.1% is high and associated with both abnormal motility and esophageal leak. Others have reported similiar findings [22]. W e infer from these data that every effort should be m a d e to identify a n d correct associated esophageal conditions to minimize postoperative complications. These data also reinforce o u r conclusion that epiphrenic diverticula should be surgically managed only in patients who h a v e incapacitating symptoms. In summary, s y m p t o m s i n patients with epiphrenic diverticula are variable. Patients with absent o r minimal symptoms should b e managed conservatively as progression of s y m p t o m s is unlikely. Patients with incapacitating s y m p t o m s should u n d e r g o diverticulectomy a n d esophagomyotomy. Operative risks, however, are substantial a n d portend the difficulties in performing multiple concomitant procedures on t h e esophagus. Nonetheless, long-term results are acceptable a n d durable. W e hope that further understanding of the pathophysiology of epi-
1113
phrenic diverticula will allow better selection of patients, reduced morbidity, and improved long-term results.
References 1. Habein HC, Kirklin JW, Clagett OT, Moersch HJ. Surgical treatment of lower esophageal pulsion diverticula. AMA Arch Surg 1956;72:1018-24.
2. Debas HT, Payne WS, Cameron AJ, Carlson HC. Physiopathology of lower esophageal diverticulum and its implications for treatment. Surg Gynecol Obstet 1980;151:593-600. 3. Bruggeman LL, Seaman WB. Epiphrenic diverticula: an analysis of 80 cases. Am J Roentgenol 1973;119:26&76. 4. Conrad C, Nissen F. Giant epiphrenic diverticula. Eur J Radio1 1982;2:48-9. 5. Ravich WWJ. Esophageal dysphagia. In: Groher ME, ed. Dysphagia: diagnosis and management, 2nd ed. Stoneham: Butterworth-Heineman, 1992:85-6. 6. Montesani C, DAmato A, Citone G, Procacciante F, Narilli P, Ribotta G. Surgical treatment of epiphrenic diverticula of the esophagus: is diverticulectomy always necessary? A report of two cases. ltal J Surg 1985;15:6%73. 7. Allen TH, Clagett OT. Changing concepts in the surgical treatment of pulsion diverticula of the lower esophagus. J Thorac Cardiovasc Surg 1965;50455-62. 8. Habein HC, Moersch HJ, Kirklin JW. Diverticula of the lower part of the esophagus: a clinical study of 140 nonsurgical cases. Arch Intern Med 1956;97:768-77. 9. Harrington SW. The surgical treatment of pulsion diverticula of the thoracic esophagus. Ann Surg 1949;129:606-18. 10. Altorki NK, Sunagawa M, Skinner DB. Thoracic esophageal diverticula: why is operation necessary? J Thorac Cardiovasc Surg 1993;105:260-4. 11. Evander A, Little AG, Ferguson MK, Skinner DB. Diverticula of the mid- and lower esophagus: pathogenesis and surgical management. World J Surg 1986;10:820-8. 12. Duda M, S e q 2, Vojaeek K, Rotek V, Rehulka M. Etiopathogenesis and classification of esouhageal diverticula. Int Sura "
-1985;70:291-5.
I
"
13. Trastek VF. Pavne WS. EsoDhaeeal diverticula. In: Shields 14. 15. 16.
17.
18.
TW, ed. Geneial thoracic &gky. Philadelphia: Lea and Febiger, 1989:989-1001. Mondiere JT. Notes sur quelques maladies de I'oesophage. Arch Gen Med Paris 1833;3:28-65. Vinson PP. Diverticula of the thoracic portion of the esophagus: report of 42 cases. Arch Otolaryngol 1934;19:508-13. Bontempo I, Corazziari E, Mineo TC, Tosoni M, Anzini F, Ricci C. Esophageal motor activity in patients with esophageal diverticula. In: DeMeester TR, Skinner DB, eds. Esophageal disorders, pathophysiology and therapy. New York: Raven Press, 1985:427-9. Dodds WJ, Stef JJ, Hogan WJ, Hoke SE, Stewart ET, Arndorfer RC. Dstribution of esophageal peristaltic pressure in normal subjects and patients with esophageal diverticulum. Gastroenterology 1975;69:58&90. Kaye MD. Oesophageal motor dysfunction in patients with diverticula of the mid-thoracic oesophagus. Thorax 1974;
29:66&72. 19. Viard H. Sala JJ, Favre JP, Bernard A, Cougard P. Le
traitement chirurgical des diverticules de pulsion de I'oesophage. J Chir (Paris) 1987;124:658-62. 20. Belsey R. Functional disease of the esophagus. J Thorac Cardiovasc Surg 1966;52:1W8. 21. Little AG, Soriano A, Ferguson MK, Winans CS, Skinner DB. Surgical treatment of achalasia: results with esophagomyotoma and Belsey repair. Ann Thorac Surg 1988;45:489-94. 22. Fekete F, Vonns C. Surgical management of esophageal thoracic diverticula. Hepatogastroenterology 1992;39:97-9.
1114
BENACCI ET AL EFIPHRENIC DIVERTICULUM
Ann Thorac Surg 1993;55:1109-14
DISCUSSION DR JOSEPH I. MILLER, JR (Atlanta, GA): I want to thank the authors for the opportunity to review the manuscript before presentation. As usual, Dr Benacci and Dr Deschamps are to be congratulated on their presentation. We have all come to expect that from the group at the Mayo Clinic. I have several comments. We would agree that all of their patients have been adequately worked up with appropriate manometrics, barium swallow, and endoscopy; however the complication rate is quite high, with six leaks, or 18% of the total group, with three deaths having resulted. This would actually be six leaks in 30 patients who underwent the repair excluding the 3 who had esophagogastrectomy. Also noted is that only 22 of their 33 patients actually had a myotomy, although I have the feeling that the Mayo Clinic must have changed recently, in that I noticed in their conclusions myotomy is recommended in all patients. Also, only 6 patients had a hiatal hernia repair. Having been at the Mayo Clinic 1 am well aware of their tendency not to do a concomitant hiatal hernia repair when a myotomy is done even in their patients with achalasia. At Emory University, Dr Mansour and I have had a smaller experience with only approximately 22 patients operated on with epiphrenic diverticulum; however, we have performed a diverticulectomy plus a modified Heller myotomy and modified Belsey repair in all patients in the last 17 years. We have had no leaks and no operative mortality in this group of patients. It is our impression that all patients should undergo a myotomy with their combined diverticulectomy, and in addition we would recommend performance of a modified Belsey at the time of repair. My questions to Dr Benacci are, do you still think that a concomitant hiatal hernia repair is contraindicated in this group of patients, in the majority of the patients? Perhaps the cause of your increased leak rate is an inadequate myotomy without complete relief of the increased pressure zone in the distal esophagus. This could be caused by inadequate mobilization at the area of the gastroesophageal junction. DR MILTON V. DAVIS (Kaufman, TX): This is a very beautiful report, and I agree completely with what Dr Miller said except for one thing. I would do today a different kind of reflux-preventing operation. This report is flawed, but not because they have reported their results honestly, which we appreciate. The fact is that the epiphrenic diverticulum is like the part of the iceberg that sticks out of the water. It is not the problem, and it should not be treated as the problem. It is the result of the problem. Another point 1 want to make is that there are terrific problems when surgeons allow the gastroenterology service to select these patients. Now, if that is the only place you are going to get your patients, you get them wherever you can, and I do not fault that. But you said you evaluated them by gastrointestinal studies, by scoping, and by manometry. What about a history? A history is what you need to evaluate these patients; it is what you need to decide on their treatment. I have one last point. The word stricture is kicked around in our literature time and time again. Our guest speaker from Switzerland, Dr Naef, reported before this Society when Hawley Seiler invited him first to be a guest speaker about the failure of adequate diagnosing of esophageal strictures. I will not go into it any farther, but that is my point.
DR KAMAL A. MANSOUR (Atlanta, GA): This is a nice and really important report for the practicing general thoracic surgeon. One is, however, struck with the high 24% incidence of esophageal leak postoperatively. I have a statement and two questions for you. The statement is, although these diverticula present more frequently on the right side, it is our practice, as it is yours also, to perform the diverticulectomy, the distal myotomy, and the Belsey mark IV through a left thoracic approach. Failure to adhere to this so-called triple-treat principle led one of the thoracic surgeons in our community to perform only diverticulectomy through a right thoracotomy, which was followed by massive and persistent esophageal leak for 7 weeks postoperatively before it was repaired through a left thoracic approach and by adding the myotomy and the Belsey mark IV. Now, I have two questions. Did you find any relationship between the incidence of leak postoperatively in patients who underwent diverticulectomy alone as compared with those who had myotomy in addition to excision of their diverticulum? Second I noticed that you worked up your patients with barium swallow, endoscopy, and manometry. How about 24-hour pH studies; are you going to d o them in the future? DR THOMAS M. EGAN (Chapel Hill, NC): I enjoyed your report. 1 think Dr Mansour asked my question, but occasionally problems provide you with opportunities, and you have an opportunity here to analyze your esophageal leak rate. Specifically, was there a relationship between esophageal leak and the absence of a myotomy, or were there any other factors that you identified that contributed to your esophageal leak rate? DR BENACCI: Doctor Miller, our patients spanned 16 years, and the operative procedure undoubtedly evolved over this time. We certainly d o not believe that hiatal hernia repair is contraindicated, but we do not perform routine concomitant hiatal hernia repair. Those patients who have a sliding hiatal hernia documented by upper gastrointestinal studies or those who have a compatible history of gastroesophageal reflux supported by 24hour pH monitoring should have a fundoplication. Because an associated motility disorder is invariably present, we favor an incomplete wrap, such as a Belsey repair. We also had no evidence that the cause of leak was an inadequate esophagomyotomy. Quite the contrary, the highest leak rates occurred in patients who had diverticulectomy, esophagomyotomy, and hiatal hernia repair. The esophagomyotomy should extend onto the stomach for 0.5 to 1 cm. Doctor Davis, most patients with epiphrenic diverticulum are referred through our gastroenterologists. We agree entirely that history is extremely important. If the patient is asymptomatic, the odds are great that the diverticulum will not be discovered. We also believe that only symptomatic patients should have their diverticulum repaired. As you stated, the diverticulum is not the cause of the problem, and symptomatic patients need further investigation as an underlying motility disorder is often present. Doctor Mansour, diverticulectomy alone was only done in 7 patients; leaks did not develop in any of them. All of our leaks occurred in patients who had both diverticulectomy and esophagomyotomy. Our current practice is to perform a 24-hour pH study only in those patients who have a history suggestive of pyrosis or who have a diaphragmatic hernia.