- Email: [email protected]
Paraesophageal hiatal hernia: Is an antireflux procedure necessary?
Paraesophageal Hiatal Hernia: Is an Antireflwx Procedure Necessary? Warren A. Williamson, MD, F. Henry Ellis, Jr, MD, PhD, John M. Streitz, Jr, MD, and David M. Shahian, MD Department of Thoracic and Cardiovascular Surgery, Lahey Clinic Medical Center, Burlington, and Division of Cardiothoracic Surgery, New England Deaconess Hospital and Harvard Medical School, Boston, Massachusetts
Between January 1970 and October 1992, 119 patients underwent 126 repairs of a paraesophageal hiatal hernia at the Lahey Clinic. Seven patients with a recurrent hernia required reoperation. Of the procedures, 19 (15%) included an antireflux procedure because of severe reflux symptoms and objective evidence of reflux demonstrated by grade 2 esophagitis on endoscopy, manometric evidence of a hypotensive lower esophageal sphincter pressure (110 mm Hg), positive results on 24-hour pH monitoring, or all three methods. Follow-up ranged from 6 months to 18 years with a median of 61.5 months, and the results of 115 operations were analyzed. Symptom-
atic results were good to excellent after 96 (83.5%) of these 115 operations. Thirteen symptomatic paraesophageal hernias recurred in 12 patients (one recurrence per 58 patient-years of follow-up). Severe reflux symptoms accompanied by endoscopic evidence of esophagitis developed in 2 patients who had not undergone an antireflux procedure at the time of repair of the hernia. We conclude that an antireflux procedure is rarely required in patients undergoing repair of a paraesophageal hiatal hernia and should be employed only when objective evidence of reflux is seen preoperatively. (Ann Thorac Surg 1993;56:447-52)
M
hernia at the Lahey Clinic in 119 patients in an effort to resolve this controversy.
isconceptions regarding the anatomy and pathophysiology of paraesophageal hiatal hernia have led to confusion and controversy regarding appropriate therapy. Paraesophageal hiatal hernia is both anatomically and pathophysiologically distinct from the more common sliding hiatal hernia. As depicted in Figure 1, the gastroesophageal junction is fixed posteriorly by the phrenoesophageal membrane. Patients with this anatomic abnormality rarely have a hypotensive lower esophageal sphincter (LES) with symptoms or objective evidence of gastroesophageal reflux (GER). On the other hand, a sliding hiatal hernia is characterized by laxity in fixation of the gastroesophageal junction, which allows cephalad mediastinal migration. All the same, most patients with a sliding hiatal hernia do not have a hypotensive LES with GER. When sliding and paraesophageal hernias coexist, they are called "mixed" hernia. In this condition, the LES can still be competent, however. Some authors [l]believe that all paraesophageal hernias begin as a sliding hernia and, therefore, require an antireflux procedure. Others [2] believe that most patients with paraesophageal hernia have a hypotensive LES with demonstrable GER on 24-hour pH monitoring and, therefore, should have an antireflux procedure. Because our experience is at variance with these views, we reviewed the results of surgical repair of paraesophageal hiatal Presented at the Twenty-ninth Annual Meeting of The Society of Thoracic Surgeons, San Antonio, TX, Jan 25-27, 1993. Address reprint requests to Dr Williamson, Department of Thoracic and Cardiovascular Surgery, Lahey Clinic Medical Center, 41 Mall Rd, Burlington, MA 01805.
0 1993 by The Society of Thoracic Surgeons
Material and Methods Between January 1970 and October 1992, 119 patients underwent 126 operations for repair of a paraesophageal hiatal hernia. Seven patients required reoperation for a recurrent hernia. Results in 43 of these patients have previously been reported [3]. Thirteen other patients, not included in this study, were operated on during this time for iatrogenic paraesophageal hernias, and they were included in an earlier study [4]. The 44 men and 75 women ranged in age from 37 to 95 years, with a median age of 64 years.
Clinical Feu t ures The most common symptom was pain, which was a major complaint of 88 patients (74%). Pain was usually epigastric or substernal in location and typically occurred postprandially. The next most common symptom was obstruction, evidenced by vomiting or dysphagia or both and occurring in 60 patients (50%). Heartburn was present in 39 patients (33%). Two or more symptoms occurred in 61 patients (51%). Anemia was noted in 31 patients (26%), and acute bleeding requiring transfusion occurred in 9 patients (8%). In 7 patients (6%), acute obstruction necessitated emergency surgical repair.
Diagnostic Features Contrast upper gastrointestinal radiography performed in 114 patients (96%) demonstrated the typical radiologic appearance of a paraesophageal hiatal hernia, with the fundus and often the body of the stomach protruding 0003-4975/93/$6.00
448
WILLIAMSON ET AL PARAESOPHAGEAL HIATAL HERNIA
Ann Thorac Surg 1993;56:447-52
PARAESOPHAGEAL HERNIA
ANT!R'PR
.ATERAL
Peritoneum
Phrenoesophageal I igament (dotted line)
'
ulcer was identified within the herniated stomach in 1 patient and was benign by endoscopic biopsy. Ninety-one patients (76%) were considered to have a pure paraesophageal hernia, whereas 28 patients (24%) were identified as having a "mixed" hernia on the basis of endoscopic or radiographic studies and subsequent operative findings. Esophageal manometry was performed preoperatively or intraoperatively in 34 patients (29%)and attempted in several others. Lower esophageal sphincter pressures ranged from 0 to 27 mm Hg, with a median of 16 mm Hg. Currently, 24-hour pH monitoring has been added to our diagnostic evaluation of these patients when possible. In 6 patients in whom the study was performed, 2 had frequent reflux and poor clearance and had scores of 218 and 57.7 that were considered positive. Both of these patients also had a hypotensive LES.
Surgical Features Reduction of the hernia, excision of the sac, and narrowing of the hiatus by crural reapproximation anterior to the esophagus, as described by Collis and associates [5, 61, was performed in all patients. Fixation of the stomach was accomplished with an anterior gastropexy in 15 patients (13%)and a Stamm gastrostomy in 69 patients (58%). Nineteen patients also underwent antireflux procedures. Fourteen underwent Nissen fundoplication; 4, a Hill posterior gastropexy; and 1, a Collis-Nissen procedure. Antireflux procedures were deemed necessary because of endoscopic evidence of grade 2 esophagitis, a hypotensive LES ( 4 0 mm Hg), or positive results on 24-hour pH monitoring. No difference was evident in the number of antireflux procedures performed in the mixed or pure paraesophageal hernias (25% and 15%, respectively; p = 0.16 using Miettinen's modification of the Fisher exact test). Eleven patients were found to have incarcerated hernias that could not be reduced by simple manual traction
Fig I. Hernias through the esophageal hiatus. (Reprinted with the permission of Lahey Clinic.)
through the hiatus into the posterior mediastinum. A large hernia with the entire stomach in the chest accompanied by counterclockwise rotation of the stomach was demonstrated radiologically (Fig 2) and confirmed surgically in 8 patients (6.7%). Portions of the colon or small intestine or both were identified in the hernial sac in 5 patients by contrast studies. Diagnostic endoscopic evaluation of the esophagus was performed in 67 patients (56%)with symptoms of heartburn, dysphagia, or bleeding. Evidence of esophagitis was found in 16 (24%) of these patients, but it was low grade in all. Ten had grade 2 esophagitis with evidence of linear ulceration, and 6 had mild grade 1 esophagitis (4 of whom had normal LES pressures). In addition, 2 patients had endoscopic evidence of Barrett esophagus. A gastric
Fig 2 . Typical radiologic appearance of a large paraesophageal hiatal hernia.
WILLIAMSON ET AL PARAESOPHAGEAL HIATAL HERNIA
Ann Thorac Surg 1993:56:447-52
but required sharp dissection in the posterior mediastinum to free the stomach. Seven emergency operations were performed for incarcerated and obstructed hernias, but none of these patients had a strangulated hernia or perforation of the stomach. Only 2 of these patients reported symptoms of GER preoperatively. They had no evidence of esophagitis on endoscopy and had anatomic repair without an antireflux procedure.
449
reported poor results, 12 were the result of recurrence and two were the result of severe GER. Emergency operation for obstruction did not adversely affect symptomatic results. Of the 7 patients operated on for acute obstruction, all had anatomic repairs without an antireflux procedure, and 6 reported excellent results without reflux symptoms. One patient died postoperatively of unrelated causes, as already mentioned.
Recurrence Results Morbidity and Mortality Two patients (1.7%)died in the postoperative period. One death occurred in a 67-year-old woman who required reoperation for a recurrent hernia on postoperative day 4; the cause of death was pulmonary embolism. The second death occurred in a 95-year-old man who experienced respiratory insufficiency resulting from intrapulmonary hemorrhage from a Swan-Ganz catheter after emergency repair of an incarcerated hernia with complete obstruction. Postoperative complications developed in 14 patients (11.8%):wound infections in 5 patients, ventral hernias in 5 patients, and respiratory complications secondary to retained pulmonary secretions in 3 patients. In 1 patient, esophageal leakage developed after reoperation for a recurrent paraesophageal hernia. The patient recovered after an esophageal diversion-exclusion procedure and later underwent reconstruction with interposition of a segment of colon.
Symptomatic Results Follow-up data were obtained by chart review, reexamination of the patient, responses to a questionnaire, or all three methods. Excluding the 2 patients who died postoperatively and the 8 patients lost to follow-up review, the results of 115 operations were analyzed, and patients were observed for 6 months to 18 years for a total of 751 patient-years. The median follow-up was 61.5 months. Fifteen patients who died of unrelated causes but whose follow-up data were available at the time of death are included. An asymptomatic patient was considered to have an excellent result. A good result implied that gastrointestinal symptoms rarely occurred, and a fair result indicated that the patient continued to have symptoms but to a lesser degree than before operation. A poor result indicated that the condition of the patient was unimproved or worsened by the operation or an anatomic recurrent hernia had developed. Evidence of GER was carefully sought by questions concerning heartburn, regurgitation, and anemia, the frequency of these symptoms, and the need for antacids or acid-suppression therapy. Findings by endoscopy or gastrointestinal radiography or both were obtained when possible. Of the 115 operatiok with complete follow-up data, the result was considered to be excellent in 62 procedures, good in 34, fair in 5, and poor in 14. Thus, results were good to excellent in 83.5% of these operations. Of the 14
Thirteen symptomatic paraesophageal hernias recurred in 12 patients, an incidence of 11%(one recurrence per 58 patient-years of follow-up). Recurrent hernias developed between 4 days and 12 years after the original operation, with a median of 2 years. Eight of these patients have undergone reoperation, 7 by us and 1at another hospital. Results in 5 patients are good to excellent 3 to 12 years later. One of the 8 patients had a second recurrent hernia, for which she underwent partial gastrectomy and Billroth I reconstruction at another hospital and still reports poor results with severe gastrointestinal symptoms. Four patients with recurrent hernias have refused or deferred operation. One of the postoperative deaths occurred from a pulmonary embolus after reoperation for a recurrent hernia. One serious complication, esophageal leakage, emphasizes the serious morbidity and mortality associated with reoperation for a recurrent hernia. Identifiable causes of recurrent hernia were found in 3 patients: inadequate closure of the hiatus in 2 patients and failure to remove the hernia sac in 1 patient. In 2 other patients, recurrent hernias developed 2 weeks and 4 days after operation. Gastric distention was implicated in these two recurrent hernias. Recurrent hernias developed in 11.8%and 13.3% of patients with fixation of the stomach by gastrostomy and gastropexy, respectively, whereas patients who had no fixation of the stomach had a recurrence rate of 9.1%. The differences were not significant ( p = 0.89 using Miettinen’s modification of the Fisher exact test).
Gastroesophageal Reflux Two patients who did not have an antireflux procedure at the time of hernia repair experienced severe reflux symptoms postoperatively. This represents 2% of the patients who had anatomic repair of the hernia without an antireflux procedure. One of these patients clearly should have undergone an antireflux operation when the hernia was repaired because he had symptoms of reflux as well as evidence of linear ulceration on endoscopy preoperatively. None of the 19 patients who had an antireflux procedure at the time of hernia repair had major reflux symptoms postoperatively. However, 21 patients (IS%), including 3 who had an antireflux procedure, reported occasional episodes of heartburn postoperatively, although only 4 required medication for control of symptoms.
Comment The data we have presented indicate that surgical repair of a paraesophageal hiatal hernia is safe and should be
450
WILLIAMSON ET AL PARAESOPHAGEAL HIATAL HERNIA
performed electively in all symptomatic patients who are suitable candidates for operation. Mortality rates associated with elective repair range from 0% to 1%[7-lo]. The two deaths in our experience followed an emergency operation for obstruction and a reoperation for a recurrent hernia. This is consistent with the experience of other authors [ 11-14]. Emergency operation for obstruction, strangulation, or perforation is associated with an operative mortality rate of between 16% and 50%, which is well documented in the literature [ll-141. That none of our patients had a strangulated hernia or perforation of the stomach may be explained by our early operative intervention in all patients who had obstruction on presentation and by our emphasis on elective surgical repair of these hernias. The paraesophageal hiatal hernia, like the inguinofemoral hernia, is an anatomic defect that can result in serious complications. Furthermore, nonsurgical treatment or only observation of such patients can lead to progression of symptoms, as reported in 45% of patients by Treacy and Jamieson [15], or to death secondary to obstruction, as reported in 27% of patients by Skinner and Belsey (161. For this reason, these hernias merit repair in all symptomatic patients who are suitable operative candidates, and a case can be made for surgical correction in the asymptomatic patient as well [17]. Disagreement arises, however, over the incidence of GER and the perceived need for an antireflux procedure in all patients with a paraesophageal hiatal hernia. Our observations and experience have led us to an algorithm that has been useful in the management of these patients. One third of our patients complained of reflux symptoms preoperatively. Our approach has been to evaluate these patients for objective evidence of reflux by endoscopy, manometry, 24-hour pH monitoring, or all three procedures. Of these 39 patients, 19 had objective evidence of GER and underwent an antireflux procedure. In the absence of objective findings, we have not performed an antireflux procedure. As long as the posterior attachments of the esophagus at the hiatus are not disturbed, it is unlikely, in our experience, that reflux symptoms will occur postoperatively after anatomic repair of a paraesophageal hiatal hernia. In fact, if an antireflux procedure is performed in all patients with a paraesophageal hiatal hernia, these attachments may be disturbed and result in GER. Ozdemir and colleagues [8] noted that symptoms of GER and endoscopic evidence of severe esophagitis developed postoperatively in 7.1% of their patients who underwent a routine antireflux procedure at the time of repair of a paraesophageal hernia even though they had no reflux symptoms preoperatively. Of 51 patients in the series of Pearson and associates [l], 9 (18%) experienced reflux symptoms after repair of a paraesophageal hernia when an antireflux procedure was routinely performed. In 4 of the 9 patients, the symptoms were severe. Although only a minority of our patients underwent preoperative esophageal manometry, they were patients who had reflux symptoms and thus were more likely to demonstrate a hypotensive LES. Manometric studies per-
Ann Thorac Surg 1993;56:447-52
formed in these 34 patients (29%)support the notion that the LES functions normally in most patients with paraesophageal hiatal hernia. Of the 34 patients, 26 had a normal LES. This experience is at variance with that of Walther and colleagues [2], who reported a 60% incidence of LES incompetence in patients with a paraesophageal hernia. Their experience, however, was based on 24-hour pH studies. In 6 of our patients who had 24-hour pH monitoring, results were abnormal in only 2. These patients were also found to have a hypotensive LES on manometry. It is unclear what role 24-hour pH monitoring plays in evaluating patients with paraesophageal hiatal hernia because many of these patients have intermittent gastric obstruction caused by the hernia. In the presence of gastric outlet obstruction, a normal LES may become incompetent at times and may result in reflux symptoms and abnormal reflux by 24-hour pH studies. After reduction of the hernia and relief of the obstruction, it would be expected that these symptoms would resolve. In our series, heartburn resolved completely in 21 of our patients after anatomic repair alone, thus corroborating this fact. This, again, emphasizes the importance of physiologic assessment of the competency of the LES in these patients. The incidence of symptomatic recurrence was surprisingly high in our series, with 13 recurrences among 12 patients. In five of these recurrences, identifiable causes included inadequate closure of the hiatus, failure to remove the hernia sac, and acute postoperative gastric distention. To avoid a recurrent hernia, it is important to close the hiatus to permit only one fingerbreadth alongside the esophagus and to attempt to remove most of the hernia sac when possible. Although we prefer gastrostomy to gastropexy, no statistical evidence exists that fixation of the stomach by either method made a difference in the overall recurrence rate. Nevertheless, we continue to perform gastrostomy to avoid postoperative distention, which may lead to recurrence. The true incidence of anatomic recurrent hernia may well be underestimated in this series because only 36 patients had postoperative upper gastrointestinal radiography or endoscopy. Because these studies were carried out in patients with symptoms, an asymptomatic anatomic recurrent hernia could easily have been overlooked and is clearly one of the shortcomings of a retrospective review. Despite these shortcomings, this review represents a large clinical experience with surgical repair of paraesophageal hiatal hernia. It supports the fact that anatomic repair with removal of the hernia sac and crural reapproximation is an appropriate operation for most patients with this condition. An antireflux procedure should be reserved for patients with objective evidence of GER.
We thank Gerald J. Heatley, MS, Sias Surgical Research Unit, Lahey Clinic Medical Center, for statistical evaluation of the data.
Ann Thorac Surg 1993;56447-52
References 1. Pearson FG, Cooper JD, Ilves R, Todd TRJ, Jamieson WRE. Massive hiatal hernia with incarceration: a report of 53 cases. Ann Thorac Surg 1983;35:45-51. 2. Walther B, DeMeester TR, Lafontaine E, Courtney JV, Little AG, Skinner DB. Effect of paraesophageal hernia on sphincter function and its implication on surgical therapy. Am J Surg 1984;1471114. 3. Ellis FH Jr, Crozier RE, Shea JA. Paraesophageal hiatus hernia. Arch Surg 1986;121:41&20. 4. Streitz JM Jr, Ellis FH Jr. Iatrogenic paraesophageal hiatus hernia. Ann Thorac Surg 1990;50:446-9. 5. Collis JL, Kelly TD, Wiley AM. Anatomy of the aura of the diaphragm and the surgery of hiatus hernia. Thorax 1954;9:17589. 6. Collis JL. Surgical control of reflux in hiatus hernia. Am J Surg 1986;115:465-71. 7. Allen MS, Trastek VF, Dechamps C, Pairolero PC. Intrathoracic stomach: presentation and results of operation. J Thorac Cardiovasc Surg 1993;105:253-9. 8. Ozdemir IA, Burke WA, Ikins PM. Paraesophageal hernia: a life-threatening disease. Ann Thorac Surg 1973;16:547-54. 9. Menguy R. Surgical management of large paraesophageal hernia with complete intrathoracic stomach. World J Surg 1988;12:415-22.
WILLIAMSON ET AL PARAESOPHAGEAL HIATAL HERNIA
451
10. Landreneau RJ, Johnson JA, Marshall JB, Hazelrigg SR, Boley TM, Curtiss JJ. Clinical spectrum of paraesophageal herniation. Dig Dis Sci 1992;3753744. 11. Wichterman K, Geha AS, Cahow CE, Baue AE. Giant paraesophageal hiatus hernia with intrathoracic stomach and colon: the case for early repair. Surgery 1979;86:497-506. 12. Hill LD. Incarcerated paraesophageal hernia: a surgical emergency. Am J Surg 1973;126:286-91. 13. Babb RR, Peck OC, Jamplis RW. Gastric volvulus and obstruction in paraesophageal hiatus hernia: a surgical emergency. Am J Dig Dis 1972;17:119-28. 14. Shocket E, Neber J, Drosd RE. The acutely obstructed incarcerated paraesophageal hiatal hernia. Am J Surg 1964;108: 80510. 15. Treacy PJ, Jamieson GG. An approach to the management of para-oesophageal hiatus hernia. Aust NZ J Surg 1987;57 813-7. 16. Skinner DB, Belsey RH. Surgical management of esophageal reflux and hiatus hernia: long-term results with 1,030 patients. J Thorac Cardiovasc Surg 1967;53:3&54. 17. Harriss DR, Graham TR, Galea M, Salama FD. Paraoesophageal hiatal hernias: when to operate. J R Coll Surg Edinb 1992;37:97-8.
DISCUSSION DR F. GRIFFITH PEARSON (Toronto, Ont, Canada): Doctor Williamson and his colleagues have given a very clear presentation of their experience with large, incarcerated hiatal hernias at the Lahey Clinic. They note, at the commencement of their report, the confusion that still exists about terminology. The terms paraesophageal hernia, intrathoracic stomach hernia, and giant hiatal hernia are all used. Doctor Williamson and colleagues conclude that a majority, three quarters or more, are true paraesophageal hernias with no potential for gastroesophageal reflux, and this conclusion modifies their approach to management. I d o not believe their data support this judgment. Indeed, I think some of the data suggest otherwise. Our opinion is that almost all of these giant hernias are the end result of an ordinary sliding hernia. As herniation progresses, the lesser curve is tethered, not by the gastroesophageal junction, but by the gastrohepatic omentum on the short lesser curve of the stomach. As the greater curve rolls up, it lies alongside the esophagus and, when extensive, inevitably undergoes organoaxial volvulus. Why is the clarification of the nature of the anatomy important? It is important because this evaluation influences treatment and the decision whether or not to add an antireflux repair. I will compare our observations with those just reported. The Lahey Clinic group studied 119 patients. We reported on 53 patients in 1983. The symptoms in both groups are close to identical: Pain was experienced in 74% of their patients and 74% of ours; obstruction, in 50% of their patients and 90% of ours; anemia, in 26% of their group and 40% of ours; and reflux at the time of presentation, 33% in the present report and 30% in our experience. I believe that these are similar patients we are talking about. I think our patients may have had more advanced herniation in that 48% of them had organoaxial volvulus compared with 7% in the Lahey Clinic report. Fifty-six percent of the Lahey Clinic patients underwent endoscopy. There is no record of the position of the esophagogastric junction. Since the advent of flexible endoscopy, we critically record the level of the esophagogastric junction. It is a clearly definable location.
It is true that these patients often have no esophagitis once a giant hernia has developed, and only 24% of the Lahey Clinic patients had esophagitis versus 30% of ours, again a similar figure. We examined 65% of our patients by manometry. In almost all, the lower esophageal sphincter was respiratory negative, which indicates an intrathoracic location and a sliding-type hiatal hernia. Our patients were managed by a transthoracic operation, and in three quarters of them, we identified evidence of shortening and added a gastroplasty. Now let us look at results. In our 53 patients, 91% had good to excellent results, none had a poor result, and no patient required reoperation. Doctor Williamson and associates report that 83.5% had good to excellent results, and 10% of patients had anatomic recurrence of hernia. Seven reoperations were for this problem. We had no anatomic recurrences. I contend that you cannot identify the status of esophagitis through a transabdominal approach. You cannot identify the presence or absence of a sliding hernia without recording the position of the junction at endoscopy; all patients require that assessment. The reason for the high recurrence rate just reported is that Dr Williamson and his colleagues were dealing with a group of patients in whom esophageal shortening was not recognized, resulting in failure of a standard transabdominal Nissen repair. DR DAVID B. SKINNER (New York, NY): First, the Collis repair that was used in this study was introduced by Dr Collis. He, Nissen, and Belsey were the early pioneers of antireflux surgery. In fact, the repair that Dr Williamson and associates have done is an antireflux repair. The diaphragm rises vertically from the retroperitoneum before it flattens out on the dome. Putting the esophagus in a tunnel like this makes it into a retroperitoneal tunnel comparable with buttressing the esophagus as we do with fundoplications. So I do not agree with Dr Williamson and associates. They insist that an antireflux repair is not necessary, but they did it on all of their patients. Second, the pure paraesophageal hernia is very rare. I agree with Dr Hiebert about that. Almost all of these are combined,
452
WILLIAMSON ET AL PARAESOPHAGEAL HIATAL HERNIA
sliding, and paraesophageal hernias, and it is hard to predict ahead of time which ones will have reflux if you do not do an antireflux repair. The abdominal versus thoracic discussion continues. I am aware of at least 3 patients whose esophagus was avulsed while it was being reduced through an abdominal incision. All these patients had a posterior as well as an anterior hiatal hernia sac, and in our experience, resection of both the anterior and posterior sacs is the key to avoiding an anatomic recurrence. Early identification and mobilization of the esophagus with the vagus nerves through a thoracotomy and then identification and removal of both Sacs leads to a safe anatomic repair. I congratulate Dr Williamson and associates on their data but disagree with their conclusion. DR CLEMENT A. HIEBERT (Windham, ME): At the risk of ”piling on,” I have two comments to make. The first issue is whether the preoperative absence of reflux really means anything. We know the bloated pouch of intrathoracic stomach compresses the esophagus and may, at least in some instances, account for the absence of reflux symptoms. The surgeon really has no guarantee that the sphincter will behave responsibly when it is displaced and anchored near the hiatus. Second, why is there a fuss about six additional sutures? Three quarters of an antireflux operation has already been done: the hernia has been reduced and anchored with a gastrostomy, and the escape route has been narrowed. The trouble is that there is an occasional recurrence of reflux and sometimes a serious throttling of stomach. Considering these things and knowing how difficult reoperations can be, why not do a complete correction the first time around?
Ann Thorac Surg 1993;56447-52
DR WILLIAMSON: We disagree with Dr Pearson’s comments and believe that the physiologic integrity of the lower esophageal sphincter and not the location of the gastroesophageal junction determines whether an antireflux procedure is performed. In 66 of our patients whose gastroesophageal junction was endoscopically identified, 27 had mixed and 39, pure paraesophageal hernias. These two groups were not significantly different with respect to postoperative heartburn, antireflux procedures, or recurrence. In 8 patients with gastroesophageal junctions well above 30 cm from the incisors, the lower esophageal sphincter pressures were normal. Doctor Hiebert, we do not advocate a routine antireflux procedure because of associated morbidity. In the series reported by Dr Pearson, 9 patients (18%)had postoperative heartburn; it was moderate to severe in 4, and 2 required reoperation. Seven patients had postoperative dysphagia, severe in 2. We dispute his contention that there were no poor results when an antireflux procedure was performed. In the Mayo Clinic experience when an antireflux procedure was routinely performed, 4 patients had esophageal leakage and 7 had dysphagia. Even in the best of hands, an antireflux procedure has serious morbidity. Doctor Skinner, we do not consider anterior crural repair an antireflux procedure. Doctor Collis’ antireflux operation involved a thoracoabdominal incision, extensive mobilization of the esophagus, and creation of a retroperitoneal tunnel by closing the hiatus anteriorly. We did not mobilize the esophagus but preserved its posterior attachments and closed the defect anteriorly. An antireflux procedure does not appear to prevent recurrence. The 5-year recurrence-free survival (Kaplan-Meier analysis) was no different whether a gastrostomy, gastropexy, or antireflux procedure was used.
Between January 1970 and October 1992, 119 patients underwent 126 repairs of a paraesophageal hiatal hernia at the Lahey Clinic. Seven patients with a recurrent hernia required reoperation. Of the procedures, 19 (15%) included an antireflux procedure because of severe reflux symptoms and objective evidence of reflux demonstrated by grade 2 esophagitis on endoscopy, manometric evidence of a hypotensive lower esophageal sphincter pressure (110 mm Hg), positive results on 24-hour pH monitoring, or all three methods. Follow-up ranged from 6 months to 18 years with a median of 61.5 months, and the results of 115 operations were analyzed. Symptom-
atic results were good to excellent after 96 (83.5%) of these 115 operations. Thirteen symptomatic paraesophageal hernias recurred in 12 patients (one recurrence per 58 patient-years of follow-up). Severe reflux symptoms accompanied by endoscopic evidence of esophagitis developed in 2 patients who had not undergone an antireflux procedure at the time of repair of the hernia. We conclude that an antireflux procedure is rarely required in patients undergoing repair of a paraesophageal hiatal hernia and should be employed only when objective evidence of reflux is seen preoperatively. (Ann Thorac Surg 1993;56:447-52)
M
hernia at the Lahey Clinic in 119 patients in an effort to resolve this controversy.
isconceptions regarding the anatomy and pathophysiology of paraesophageal hiatal hernia have led to confusion and controversy regarding appropriate therapy. Paraesophageal hiatal hernia is both anatomically and pathophysiologically distinct from the more common sliding hiatal hernia. As depicted in Figure 1, the gastroesophageal junction is fixed posteriorly by the phrenoesophageal membrane. Patients with this anatomic abnormality rarely have a hypotensive lower esophageal sphincter (LES) with symptoms or objective evidence of gastroesophageal reflux (GER). On the other hand, a sliding hiatal hernia is characterized by laxity in fixation of the gastroesophageal junction, which allows cephalad mediastinal migration. All the same, most patients with a sliding hiatal hernia do not have a hypotensive LES with GER. When sliding and paraesophageal hernias coexist, they are called "mixed" hernia. In this condition, the LES can still be competent, however. Some authors [l]believe that all paraesophageal hernias begin as a sliding hernia and, therefore, require an antireflux procedure. Others [2] believe that most patients with paraesophageal hernia have a hypotensive LES with demonstrable GER on 24-hour pH monitoring and, therefore, should have an antireflux procedure. Because our experience is at variance with these views, we reviewed the results of surgical repair of paraesophageal hiatal Presented at the Twenty-ninth Annual Meeting of The Society of Thoracic Surgeons, San Antonio, TX, Jan 25-27, 1993. Address reprint requests to Dr Williamson, Department of Thoracic and Cardiovascular Surgery, Lahey Clinic Medical Center, 41 Mall Rd, Burlington, MA 01805.
0 1993 by The Society of Thoracic Surgeons
Material and Methods Between January 1970 and October 1992, 119 patients underwent 126 operations for repair of a paraesophageal hiatal hernia. Seven patients required reoperation for a recurrent hernia. Results in 43 of these patients have previously been reported [3]. Thirteen other patients, not included in this study, were operated on during this time for iatrogenic paraesophageal hernias, and they were included in an earlier study [4]. The 44 men and 75 women ranged in age from 37 to 95 years, with a median age of 64 years.
Clinical Feu t ures The most common symptom was pain, which was a major complaint of 88 patients (74%). Pain was usually epigastric or substernal in location and typically occurred postprandially. The next most common symptom was obstruction, evidenced by vomiting or dysphagia or both and occurring in 60 patients (50%). Heartburn was present in 39 patients (33%). Two or more symptoms occurred in 61 patients (51%). Anemia was noted in 31 patients (26%), and acute bleeding requiring transfusion occurred in 9 patients (8%). In 7 patients (6%), acute obstruction necessitated emergency surgical repair.
Diagnostic Features Contrast upper gastrointestinal radiography performed in 114 patients (96%) demonstrated the typical radiologic appearance of a paraesophageal hiatal hernia, with the fundus and often the body of the stomach protruding 0003-4975/93/$6.00
448
WILLIAMSON ET AL PARAESOPHAGEAL HIATAL HERNIA
Ann Thorac Surg 1993;56:447-52
PARAESOPHAGEAL HERNIA
ANT!R'PR
.ATERAL
Peritoneum
Phrenoesophageal I igament (dotted line)
'
ulcer was identified within the herniated stomach in 1 patient and was benign by endoscopic biopsy. Ninety-one patients (76%) were considered to have a pure paraesophageal hernia, whereas 28 patients (24%) were identified as having a "mixed" hernia on the basis of endoscopic or radiographic studies and subsequent operative findings. Esophageal manometry was performed preoperatively or intraoperatively in 34 patients (29%)and attempted in several others. Lower esophageal sphincter pressures ranged from 0 to 27 mm Hg, with a median of 16 mm Hg. Currently, 24-hour pH monitoring has been added to our diagnostic evaluation of these patients when possible. In 6 patients in whom the study was performed, 2 had frequent reflux and poor clearance and had scores of 218 and 57.7 that were considered positive. Both of these patients also had a hypotensive LES.
Surgical Features Reduction of the hernia, excision of the sac, and narrowing of the hiatus by crural reapproximation anterior to the esophagus, as described by Collis and associates [5, 61, was performed in all patients. Fixation of the stomach was accomplished with an anterior gastropexy in 15 patients (13%)and a Stamm gastrostomy in 69 patients (58%). Nineteen patients also underwent antireflux procedures. Fourteen underwent Nissen fundoplication; 4, a Hill posterior gastropexy; and 1, a Collis-Nissen procedure. Antireflux procedures were deemed necessary because of endoscopic evidence of grade 2 esophagitis, a hypotensive LES ( 4 0 mm Hg), or positive results on 24-hour pH monitoring. No difference was evident in the number of antireflux procedures performed in the mixed or pure paraesophageal hernias (25% and 15%, respectively; p = 0.16 using Miettinen's modification of the Fisher exact test). Eleven patients were found to have incarcerated hernias that could not be reduced by simple manual traction
Fig I. Hernias through the esophageal hiatus. (Reprinted with the permission of Lahey Clinic.)
through the hiatus into the posterior mediastinum. A large hernia with the entire stomach in the chest accompanied by counterclockwise rotation of the stomach was demonstrated radiologically (Fig 2) and confirmed surgically in 8 patients (6.7%). Portions of the colon or small intestine or both were identified in the hernial sac in 5 patients by contrast studies. Diagnostic endoscopic evaluation of the esophagus was performed in 67 patients (56%)with symptoms of heartburn, dysphagia, or bleeding. Evidence of esophagitis was found in 16 (24%) of these patients, but it was low grade in all. Ten had grade 2 esophagitis with evidence of linear ulceration, and 6 had mild grade 1 esophagitis (4 of whom had normal LES pressures). In addition, 2 patients had endoscopic evidence of Barrett esophagus. A gastric
Fig 2 . Typical radiologic appearance of a large paraesophageal hiatal hernia.
WILLIAMSON ET AL PARAESOPHAGEAL HIATAL HERNIA
Ann Thorac Surg 1993:56:447-52
but required sharp dissection in the posterior mediastinum to free the stomach. Seven emergency operations were performed for incarcerated and obstructed hernias, but none of these patients had a strangulated hernia or perforation of the stomach. Only 2 of these patients reported symptoms of GER preoperatively. They had no evidence of esophagitis on endoscopy and had anatomic repair without an antireflux procedure.
449
reported poor results, 12 were the result of recurrence and two were the result of severe GER. Emergency operation for obstruction did not adversely affect symptomatic results. Of the 7 patients operated on for acute obstruction, all had anatomic repairs without an antireflux procedure, and 6 reported excellent results without reflux symptoms. One patient died postoperatively of unrelated causes, as already mentioned.
Recurrence Results Morbidity and Mortality Two patients (1.7%)died in the postoperative period. One death occurred in a 67-year-old woman who required reoperation for a recurrent hernia on postoperative day 4; the cause of death was pulmonary embolism. The second death occurred in a 95-year-old man who experienced respiratory insufficiency resulting from intrapulmonary hemorrhage from a Swan-Ganz catheter after emergency repair of an incarcerated hernia with complete obstruction. Postoperative complications developed in 14 patients (11.8%):wound infections in 5 patients, ventral hernias in 5 patients, and respiratory complications secondary to retained pulmonary secretions in 3 patients. In 1 patient, esophageal leakage developed after reoperation for a recurrent paraesophageal hernia. The patient recovered after an esophageal diversion-exclusion procedure and later underwent reconstruction with interposition of a segment of colon.
Symptomatic Results Follow-up data were obtained by chart review, reexamination of the patient, responses to a questionnaire, or all three methods. Excluding the 2 patients who died postoperatively and the 8 patients lost to follow-up review, the results of 115 operations were analyzed, and patients were observed for 6 months to 18 years for a total of 751 patient-years. The median follow-up was 61.5 months. Fifteen patients who died of unrelated causes but whose follow-up data were available at the time of death are included. An asymptomatic patient was considered to have an excellent result. A good result implied that gastrointestinal symptoms rarely occurred, and a fair result indicated that the patient continued to have symptoms but to a lesser degree than before operation. A poor result indicated that the condition of the patient was unimproved or worsened by the operation or an anatomic recurrent hernia had developed. Evidence of GER was carefully sought by questions concerning heartburn, regurgitation, and anemia, the frequency of these symptoms, and the need for antacids or acid-suppression therapy. Findings by endoscopy or gastrointestinal radiography or both were obtained when possible. Of the 115 operatiok with complete follow-up data, the result was considered to be excellent in 62 procedures, good in 34, fair in 5, and poor in 14. Thus, results were good to excellent in 83.5% of these operations. Of the 14
Thirteen symptomatic paraesophageal hernias recurred in 12 patients, an incidence of 11%(one recurrence per 58 patient-years of follow-up). Recurrent hernias developed between 4 days and 12 years after the original operation, with a median of 2 years. Eight of these patients have undergone reoperation, 7 by us and 1at another hospital. Results in 5 patients are good to excellent 3 to 12 years later. One of the 8 patients had a second recurrent hernia, for which she underwent partial gastrectomy and Billroth I reconstruction at another hospital and still reports poor results with severe gastrointestinal symptoms. Four patients with recurrent hernias have refused or deferred operation. One of the postoperative deaths occurred from a pulmonary embolus after reoperation for a recurrent hernia. One serious complication, esophageal leakage, emphasizes the serious morbidity and mortality associated with reoperation for a recurrent hernia. Identifiable causes of recurrent hernia were found in 3 patients: inadequate closure of the hiatus in 2 patients and failure to remove the hernia sac in 1 patient. In 2 other patients, recurrent hernias developed 2 weeks and 4 days after operation. Gastric distention was implicated in these two recurrent hernias. Recurrent hernias developed in 11.8%and 13.3% of patients with fixation of the stomach by gastrostomy and gastropexy, respectively, whereas patients who had no fixation of the stomach had a recurrence rate of 9.1%. The differences were not significant ( p = 0.89 using Miettinen’s modification of the Fisher exact test).
Gastroesophageal Reflux Two patients who did not have an antireflux procedure at the time of hernia repair experienced severe reflux symptoms postoperatively. This represents 2% of the patients who had anatomic repair of the hernia without an antireflux procedure. One of these patients clearly should have undergone an antireflux operation when the hernia was repaired because he had symptoms of reflux as well as evidence of linear ulceration on endoscopy preoperatively. None of the 19 patients who had an antireflux procedure at the time of hernia repair had major reflux symptoms postoperatively. However, 21 patients (IS%), including 3 who had an antireflux procedure, reported occasional episodes of heartburn postoperatively, although only 4 required medication for control of symptoms.
Comment The data we have presented indicate that surgical repair of a paraesophageal hiatal hernia is safe and should be
450
WILLIAMSON ET AL PARAESOPHAGEAL HIATAL HERNIA
performed electively in all symptomatic patients who are suitable candidates for operation. Mortality rates associated with elective repair range from 0% to 1%[7-lo]. The two deaths in our experience followed an emergency operation for obstruction and a reoperation for a recurrent hernia. This is consistent with the experience of other authors [ 11-14]. Emergency operation for obstruction, strangulation, or perforation is associated with an operative mortality rate of between 16% and 50%, which is well documented in the literature [ll-141. That none of our patients had a strangulated hernia or perforation of the stomach may be explained by our early operative intervention in all patients who had obstruction on presentation and by our emphasis on elective surgical repair of these hernias. The paraesophageal hiatal hernia, like the inguinofemoral hernia, is an anatomic defect that can result in serious complications. Furthermore, nonsurgical treatment or only observation of such patients can lead to progression of symptoms, as reported in 45% of patients by Treacy and Jamieson [15], or to death secondary to obstruction, as reported in 27% of patients by Skinner and Belsey (161. For this reason, these hernias merit repair in all symptomatic patients who are suitable operative candidates, and a case can be made for surgical correction in the asymptomatic patient as well [17]. Disagreement arises, however, over the incidence of GER and the perceived need for an antireflux procedure in all patients with a paraesophageal hiatal hernia. Our observations and experience have led us to an algorithm that has been useful in the management of these patients. One third of our patients complained of reflux symptoms preoperatively. Our approach has been to evaluate these patients for objective evidence of reflux by endoscopy, manometry, 24-hour pH monitoring, or all three procedures. Of these 39 patients, 19 had objective evidence of GER and underwent an antireflux procedure. In the absence of objective findings, we have not performed an antireflux procedure. As long as the posterior attachments of the esophagus at the hiatus are not disturbed, it is unlikely, in our experience, that reflux symptoms will occur postoperatively after anatomic repair of a paraesophageal hiatal hernia. In fact, if an antireflux procedure is performed in all patients with a paraesophageal hiatal hernia, these attachments may be disturbed and result in GER. Ozdemir and colleagues [8] noted that symptoms of GER and endoscopic evidence of severe esophagitis developed postoperatively in 7.1% of their patients who underwent a routine antireflux procedure at the time of repair of a paraesophageal hernia even though they had no reflux symptoms preoperatively. Of 51 patients in the series of Pearson and associates [l], 9 (18%) experienced reflux symptoms after repair of a paraesophageal hernia when an antireflux procedure was routinely performed. In 4 of the 9 patients, the symptoms were severe. Although only a minority of our patients underwent preoperative esophageal manometry, they were patients who had reflux symptoms and thus were more likely to demonstrate a hypotensive LES. Manometric studies per-
Ann Thorac Surg 1993;56:447-52
formed in these 34 patients (29%)support the notion that the LES functions normally in most patients with paraesophageal hiatal hernia. Of the 34 patients, 26 had a normal LES. This experience is at variance with that of Walther and colleagues [2], who reported a 60% incidence of LES incompetence in patients with a paraesophageal hernia. Their experience, however, was based on 24-hour pH studies. In 6 of our patients who had 24-hour pH monitoring, results were abnormal in only 2. These patients were also found to have a hypotensive LES on manometry. It is unclear what role 24-hour pH monitoring plays in evaluating patients with paraesophageal hiatal hernia because many of these patients have intermittent gastric obstruction caused by the hernia. In the presence of gastric outlet obstruction, a normal LES may become incompetent at times and may result in reflux symptoms and abnormal reflux by 24-hour pH studies. After reduction of the hernia and relief of the obstruction, it would be expected that these symptoms would resolve. In our series, heartburn resolved completely in 21 of our patients after anatomic repair alone, thus corroborating this fact. This, again, emphasizes the importance of physiologic assessment of the competency of the LES in these patients. The incidence of symptomatic recurrence was surprisingly high in our series, with 13 recurrences among 12 patients. In five of these recurrences, identifiable causes included inadequate closure of the hiatus, failure to remove the hernia sac, and acute postoperative gastric distention. To avoid a recurrent hernia, it is important to close the hiatus to permit only one fingerbreadth alongside the esophagus and to attempt to remove most of the hernia sac when possible. Although we prefer gastrostomy to gastropexy, no statistical evidence exists that fixation of the stomach by either method made a difference in the overall recurrence rate. Nevertheless, we continue to perform gastrostomy to avoid postoperative distention, which may lead to recurrence. The true incidence of anatomic recurrent hernia may well be underestimated in this series because only 36 patients had postoperative upper gastrointestinal radiography or endoscopy. Because these studies were carried out in patients with symptoms, an asymptomatic anatomic recurrent hernia could easily have been overlooked and is clearly one of the shortcomings of a retrospective review. Despite these shortcomings, this review represents a large clinical experience with surgical repair of paraesophageal hiatal hernia. It supports the fact that anatomic repair with removal of the hernia sac and crural reapproximation is an appropriate operation for most patients with this condition. An antireflux procedure should be reserved for patients with objective evidence of GER.
We thank Gerald J. Heatley, MS, Sias Surgical Research Unit, Lahey Clinic Medical Center, for statistical evaluation of the data.
Ann Thorac Surg 1993;56447-52
References 1. Pearson FG, Cooper JD, Ilves R, Todd TRJ, Jamieson WRE. Massive hiatal hernia with incarceration: a report of 53 cases. Ann Thorac Surg 1983;35:45-51. 2. Walther B, DeMeester TR, Lafontaine E, Courtney JV, Little AG, Skinner DB. Effect of paraesophageal hernia on sphincter function and its implication on surgical therapy. Am J Surg 1984;1471114. 3. Ellis FH Jr, Crozier RE, Shea JA. Paraesophageal hiatus hernia. Arch Surg 1986;121:41&20. 4. Streitz JM Jr, Ellis FH Jr. Iatrogenic paraesophageal hiatus hernia. Ann Thorac Surg 1990;50:446-9. 5. Collis JL, Kelly TD, Wiley AM. Anatomy of the aura of the diaphragm and the surgery of hiatus hernia. Thorax 1954;9:17589. 6. Collis JL. Surgical control of reflux in hiatus hernia. Am J Surg 1986;115:465-71. 7. Allen MS, Trastek VF, Dechamps C, Pairolero PC. Intrathoracic stomach: presentation and results of operation. J Thorac Cardiovasc Surg 1993;105:253-9. 8. Ozdemir IA, Burke WA, Ikins PM. Paraesophageal hernia: a life-threatening disease. Ann Thorac Surg 1973;16:547-54. 9. Menguy R. Surgical management of large paraesophageal hernia with complete intrathoracic stomach. World J Surg 1988;12:415-22.
WILLIAMSON ET AL PARAESOPHAGEAL HIATAL HERNIA
451
10. Landreneau RJ, Johnson JA, Marshall JB, Hazelrigg SR, Boley TM, Curtiss JJ. Clinical spectrum of paraesophageal herniation. Dig Dis Sci 1992;3753744. 11. Wichterman K, Geha AS, Cahow CE, Baue AE. Giant paraesophageal hiatus hernia with intrathoracic stomach and colon: the case for early repair. Surgery 1979;86:497-506. 12. Hill LD. Incarcerated paraesophageal hernia: a surgical emergency. Am J Surg 1973;126:286-91. 13. Babb RR, Peck OC, Jamplis RW. Gastric volvulus and obstruction in paraesophageal hiatus hernia: a surgical emergency. Am J Dig Dis 1972;17:119-28. 14. Shocket E, Neber J, Drosd RE. The acutely obstructed incarcerated paraesophageal hiatal hernia. Am J Surg 1964;108: 80510. 15. Treacy PJ, Jamieson GG. An approach to the management of para-oesophageal hiatus hernia. Aust NZ J Surg 1987;57 813-7. 16. Skinner DB, Belsey RH. Surgical management of esophageal reflux and hiatus hernia: long-term results with 1,030 patients. J Thorac Cardiovasc Surg 1967;53:3&54. 17. Harriss DR, Graham TR, Galea M, Salama FD. Paraoesophageal hiatal hernias: when to operate. J R Coll Surg Edinb 1992;37:97-8.
DISCUSSION DR F. GRIFFITH PEARSON (Toronto, Ont, Canada): Doctor Williamson and his colleagues have given a very clear presentation of their experience with large, incarcerated hiatal hernias at the Lahey Clinic. They note, at the commencement of their report, the confusion that still exists about terminology. The terms paraesophageal hernia, intrathoracic stomach hernia, and giant hiatal hernia are all used. Doctor Williamson and colleagues conclude that a majority, three quarters or more, are true paraesophageal hernias with no potential for gastroesophageal reflux, and this conclusion modifies their approach to management. I d o not believe their data support this judgment. Indeed, I think some of the data suggest otherwise. Our opinion is that almost all of these giant hernias are the end result of an ordinary sliding hernia. As herniation progresses, the lesser curve is tethered, not by the gastroesophageal junction, but by the gastrohepatic omentum on the short lesser curve of the stomach. As the greater curve rolls up, it lies alongside the esophagus and, when extensive, inevitably undergoes organoaxial volvulus. Why is the clarification of the nature of the anatomy important? It is important because this evaluation influences treatment and the decision whether or not to add an antireflux repair. I will compare our observations with those just reported. The Lahey Clinic group studied 119 patients. We reported on 53 patients in 1983. The symptoms in both groups are close to identical: Pain was experienced in 74% of their patients and 74% of ours; obstruction, in 50% of their patients and 90% of ours; anemia, in 26% of their group and 40% of ours; and reflux at the time of presentation, 33% in the present report and 30% in our experience. I believe that these are similar patients we are talking about. I think our patients may have had more advanced herniation in that 48% of them had organoaxial volvulus compared with 7% in the Lahey Clinic report. Fifty-six percent of the Lahey Clinic patients underwent endoscopy. There is no record of the position of the esophagogastric junction. Since the advent of flexible endoscopy, we critically record the level of the esophagogastric junction. It is a clearly definable location.
It is true that these patients often have no esophagitis once a giant hernia has developed, and only 24% of the Lahey Clinic patients had esophagitis versus 30% of ours, again a similar figure. We examined 65% of our patients by manometry. In almost all, the lower esophageal sphincter was respiratory negative, which indicates an intrathoracic location and a sliding-type hiatal hernia. Our patients were managed by a transthoracic operation, and in three quarters of them, we identified evidence of shortening and added a gastroplasty. Now let us look at results. In our 53 patients, 91% had good to excellent results, none had a poor result, and no patient required reoperation. Doctor Williamson and associates report that 83.5% had good to excellent results, and 10% of patients had anatomic recurrence of hernia. Seven reoperations were for this problem. We had no anatomic recurrences. I contend that you cannot identify the status of esophagitis through a transabdominal approach. You cannot identify the presence or absence of a sliding hernia without recording the position of the junction at endoscopy; all patients require that assessment. The reason for the high recurrence rate just reported is that Dr Williamson and his colleagues were dealing with a group of patients in whom esophageal shortening was not recognized, resulting in failure of a standard transabdominal Nissen repair. DR DAVID B. SKINNER (New York, NY): First, the Collis repair that was used in this study was introduced by Dr Collis. He, Nissen, and Belsey were the early pioneers of antireflux surgery. In fact, the repair that Dr Williamson and associates have done is an antireflux repair. The diaphragm rises vertically from the retroperitoneum before it flattens out on the dome. Putting the esophagus in a tunnel like this makes it into a retroperitoneal tunnel comparable with buttressing the esophagus as we do with fundoplications. So I do not agree with Dr Williamson and associates. They insist that an antireflux repair is not necessary, but they did it on all of their patients. Second, the pure paraesophageal hernia is very rare. I agree with Dr Hiebert about that. Almost all of these are combined,
452
WILLIAMSON ET AL PARAESOPHAGEAL HIATAL HERNIA
sliding, and paraesophageal hernias, and it is hard to predict ahead of time which ones will have reflux if you do not do an antireflux repair. The abdominal versus thoracic discussion continues. I am aware of at least 3 patients whose esophagus was avulsed while it was being reduced through an abdominal incision. All these patients had a posterior as well as an anterior hiatal hernia sac, and in our experience, resection of both the anterior and posterior sacs is the key to avoiding an anatomic recurrence. Early identification and mobilization of the esophagus with the vagus nerves through a thoracotomy and then identification and removal of both Sacs leads to a safe anatomic repair. I congratulate Dr Williamson and associates on their data but disagree with their conclusion. DR CLEMENT A. HIEBERT (Windham, ME): At the risk of ”piling on,” I have two comments to make. The first issue is whether the preoperative absence of reflux really means anything. We know the bloated pouch of intrathoracic stomach compresses the esophagus and may, at least in some instances, account for the absence of reflux symptoms. The surgeon really has no guarantee that the sphincter will behave responsibly when it is displaced and anchored near the hiatus. Second, why is there a fuss about six additional sutures? Three quarters of an antireflux operation has already been done: the hernia has been reduced and anchored with a gastrostomy, and the escape route has been narrowed. The trouble is that there is an occasional recurrence of reflux and sometimes a serious throttling of stomach. Considering these things and knowing how difficult reoperations can be, why not do a complete correction the first time around?
Ann Thorac Surg 1993;56447-52
DR WILLIAMSON: We disagree with Dr Pearson’s comments and believe that the physiologic integrity of the lower esophageal sphincter and not the location of the gastroesophageal junction determines whether an antireflux procedure is performed. In 66 of our patients whose gastroesophageal junction was endoscopically identified, 27 had mixed and 39, pure paraesophageal hernias. These two groups were not significantly different with respect to postoperative heartburn, antireflux procedures, or recurrence. In 8 patients with gastroesophageal junctions well above 30 cm from the incisors, the lower esophageal sphincter pressures were normal. Doctor Hiebert, we do not advocate a routine antireflux procedure because of associated morbidity. In the series reported by Dr Pearson, 9 patients (18%)had postoperative heartburn; it was moderate to severe in 4, and 2 required reoperation. Seven patients had postoperative dysphagia, severe in 2. We dispute his contention that there were no poor results when an antireflux procedure was performed. In the Mayo Clinic experience when an antireflux procedure was routinely performed, 4 patients had esophageal leakage and 7 had dysphagia. Even in the best of hands, an antireflux procedure has serious morbidity. Doctor Skinner, we do not consider anterior crural repair an antireflux procedure. Doctor Collis’ antireflux operation involved a thoracoabdominal incision, extensive mobilization of the esophagus, and creation of a retroperitoneal tunnel by closing the hiatus anteriorly. We did not mobilize the esophagus but preserved its posterior attachments and closed the defect anteriorly. An antireflux procedure does not appear to prevent recurrence. The 5-year recurrence-free survival (Kaplan-Meier analysis) was no different whether a gastrostomy, gastropexy, or antireflux procedure was used.