The results of Hill's operation
LD. Hill, S.J.M. Kraemer, C.E. Pope, RA. Kozarek (Seattle)
The Hill Repair is an operation designed to restore the function of the antireflux barrier. The normal gastroesophageal junction (GEJ) is a highly competent barrier against reflux of gastric contents into the esophagus. With the development of a hiatal hernia, the GEJ slides into the posterior mediastinum and the barrier is lost, resulting in reflux.
Gastroesophageal reflux with its complications of esophagitis, heartburn and pneumonitis is the most common abnormality of the upper gastrointestinal (GI) tract. Despite the frequency of gastroesophageal reflux (GER), the components of the antireflux barrier have been poorly understood until recently. For antireflux surgery to be effective, the components of the antireflux barrier must be clearly understood in order that they may be restored by an operation.
The antireflux barrier
The antireflux barrier consists of the gastroesophageal valve (GEV), the lower esophageal sphincter (LES), the diaphragm , posterior fixation of the GEJ and esophageal clearance.
With our ability to measure the lower esophageal sphincter pressure (LESP) , the GEV has been overlooked. The LES generates a pressure of around 15-18 mmHg in the resting state and can generate pressures up to 100 mmHg or more. However, it is a weak sphincter and cannot alone withstand the large pressures exerted against the GEJ with heavy lifting, straining and trauma.
The GEV can be readily viewed in the cadaver through a gastrostomy. In the living person it can be seen with the retroflexed endoscope. We have also viewed the GE valve in patients through a gastrostomy. When seen through a gastrostomy without an endoscope through the valve, it has essentially the same structural appearance as that seen through the retroflexed endoscope. The valve is created by the angle of entry of the esophagus into the stomach and is an intragastric musculo-mucosal fold that, in the normal subject, adheres to the endoscope through all phases of respiration. It opens with swallowing, belching and vomiting, and closes promptly.
In our series using cadavers with no premorbid evidence of hiatal hernia or esophageal disease, we have shown that a measurable gradient of approximately 15 cm of water exists across the GE junction . This gradient can be eliminated effectively by depressing the fundus of the stomach 45°. This maneuver causes the angle of His to become obtuse, which eliminates the GE flap valve, converts the ostium of the esophagus into a funnel and results in free reflux. Since there is no lower esophageal sphincter (LES) function in the cadaver, the presence of a gradient
Figure 1. .Grading system of the GEV developed from retroflexed endoscopy in 32 patients with and without GERD.
across the GE junction and the maneuver of depressing the cardia to eliminate the gradient, helps to confirm the importance of the valve in an in situ preparation.
The appearance of the valve through the retroflexed endoscope in 20 normal volunteers without reflux was studied in order to determine how the normal valve appears.
Thirty-two patients, with and without reflux, were examined with a retroflexed endoscope and the valves were graded by a gastroenterologist blinded to the clinical status of the patient. From this study, a grading system of the valve was developed (Fig. 1).
In a separate study of 33 patients seen in the gastric laboratory who had both standard acid reflux test and grading of the GEV, the results were shown in terms of prediction of the clinical status of the patient. The results are shown in Table 1. It is noteworthy that grading of the GEV predicted the clinical picture in 32 of 33
Table 1. .This study demonstrates the correlation with clinical status of patients seen in the GI laboratory. Grading of the GEV correlated more closely with clinical picture than measurement of the LESP
Measurement of LESP
Grading of esophageal value
Correlation with clinical picture of GERD
17 of 33
Grading of GEV 32 of 33
Endoscopist blinded to clinical picture
Figure 2. .Grade I valve is a normal valve. It is a musculomucosal fold that adheres to the retroflexed endoscope through all phases of respiration, opens for swallowing, but closes promptly.
patients; whereas measurement of the LESP correlated with the clinical picture in only 17. This study showed clearly that grading of the GEV more accurately portrays the clinical status than does measurement of the LESP.
The grade I valve is a normal valve (Fig. 2). This consists of a musculomucosal fold adherent to the endoscope through all phases of respiration. It opens for swallowing and belching, and closes promptly. The grade II gastroesophageal valve is only slightly less well defined and shorter. It opens but closes promptly (Fig. 3). The grade III valve opens frequently, remains open for varying periods of time, is poorly defined and often is associated with a hiatal hernia (Fig. 4). The grade IV valve shows no well-defined musculomucosal fold; the esophageal orifice is wide open and it is invariably accompanied by a hiatal hernia (Fig. 5).
It is noteworthy that in 32 patients with or without a history of varying degrees of reflux, the GEV was graded by gastroenterologists blinded to the clinical status of the patient. No patient with grade I or II GEV showed reflux; whereas, all patients with grade III and IV valves showed reflux (Fig. 1).
In a separate study of 33 patients seen in the gastric laboratory who had both a standard acid reflux test and grading of the GEV, the results were shown in terms of prediction of the clinical status of the patient. The results are shown in Table 1. It is noteworthy that grading of the GEV predicted the clinical picture of 32 of the 33 patients whereas measurement of the LESP correlated with the clinical picture in only 17.
Figure 3. .Grade II GE valve, slightly less well-defined, but one which still opens for swallowing, closes promptly and does not allow reflux.
These studies show clearly that the GEV is an important component of the gastroesophageal barrier. The role of surgery, therefore, is to re-establish a normal grade I, 180° valve in a patient who has lost the valve and is therefore suffering from reflux.
In addition to recreating the valve, calibrating of the LES is important and can be done with intraoperative measurement of the sphincter pressure. The relationship of the LES to the valve is shown in Fig. 6. This computer generated view shows that the sphincter resides inside the valve and aids the valve in discriminating between gas, liquids and solids, and does the discriminatory work while the valve does the heavy work, in terms of preventing reflux. Increased intragastric pressure serves to close the valve against the lesser curve.
Posterior fixation of the GEJ is essential. This is lost when the patient develops a hiatal hernia and the GEJ ascends into the posterior mediastinum. The esophagus can no longer generate propulsive waves that are necessary for esophageal clearance since the esophagus no longer has a fulcrum or point of fixation from which to work. It should be recalled that the entire GI tract, including the hollow as well as solid viscera in humans and most vertebrate animals, is suspended by the dorsal mesentery to the posterior body wall. The esophagus is no exception to this rule. Extensive cadaver dissections demonstrate that the esophagus is primarily fixed posteriorly by a dense plate of fibroareolar tissue extending from the median arcuate ligament all the way to the aortic arch. The posterior attachment of the GE junction by the dorsal
Figure 4. .The grade III valve opens frequently, closes poorly, is poorly defined and may be associated with a hiatal hernia.
mesentery to the preaortic fascia is key to the integrity of the entire barrier to reflux. It has been demonstrated in the cadaver that, with division of the posterior attachment the GEJ slides into the chest and the effect of the GEV is lost. This is also demonstrated with the retroflexed endoscope in the human. As the GE junction ascends into the posterior mediastinum, the valve is lost and the sphincter is distracted. Reattachment of the GE junction is therefore important for restoration of esophageal function.
Closure of the enlarged diaphragmatic opening is important to prevent recurrence of hiatal hernia. The diaphragm should be closed loosely about the esophagus so that at least one finger can be placed alongside the esophagus with a nasogastric tube in the lumen. Fixation of the cardia to the rim of the diaphragm is also important to accentuate the valve and to close the opening into the posterior mediastinum to prevent herniation of the cardia into the posterior mediastinum. To summarize, the goals of surgery are:
- Restoration of the GEV.
- Calibration of the LES to the proper range.
- Posterior fixation of the GE junction to restore esophageal peristalsis and clearance.
- Reduction of the hiatal hernia.
- Closure of the diaphragm.
Figure 5. .In patients with a grade IV valve, there is no definable mucosal fold and the esophagus remains open the majority of the time viewed, and a hiatal hernia is invariably present.
The conventional open technique is accomplished through an upper abdominal midline incision. The abdomen is thoroughly explored. The pylorus, in particular, is examined carefully for any evidence of pyloric stenosis which might impede gastric emptying. The triangular ligament of the left lobe of the liver is divided so that the left lobe can be retracted to the patient's right. This exposes the esophageal hiatus with its covering phrenoesophageal membrane. An upper hand retractor with two blades is placed to facilitate exposure of the upper abdomen. The phrenoesophageal membrane is then divided on the diaphragm (Fig. 7), keeping as much of the fibroareolar tissue that makes up the phrenoesophageal bundles as possible with the GE junction. It is these bundles that normally hold the GE junction in place in the diaphragm and these bundles will be used to anchor the GE junction to the preaortic fascia. The lesser omentum is divided and the esophageal hiatus is exposed. The esophagus is gently diverted to the patient's left and the attachment of the cardia to the diaphragm is divided. The repair can be accomplished without dividing the short gastric vessels. Only rarely do we need to divide a short gastric vessel. This dissection must be done with care so as not to damage the spleen. Capsular tears of the spleen may be repaired by cautery or by suturing and applying Avitene. Division of the phrenogastric and superior portions of the gastrosplenic ligament mobilizes the
Figure 6. .The relationship of the LES to the valve is shown. This computer-generated view shows the relationship of the sphincter and the GEV. The sphincter resides inside the valve and aids the valve in discriminating between gas, liquid and solids, and aids in the prevention of reflux. The arrows are pressure vectors demonstrating that increased intragastric pressure closes the valve.
upper part of the gastric fundus. The fundus can then be rotated so that the posterior part of the stomach can be visualized. This allows the GE junction to be retracted down and the hiatal hernia reduced. The bundles of tissue that constitute the anterior and posterior attachments of the GE junction to the diaphragm, the anterior and posterior phrenoesophageal bundles, can then be displayed. By retracting these caudally, an intra-abdominal segment of esophagus becomes visible. The anterior and posterior vagus nerves are visualized and kept in view so as not to be damaged.
Attention is again turned to the pylorus. If the pylorus is scarred because the patient has had a duodenal ulcer or if there is a pyloric diaphragm obstructing the outlet of the stomach, a pyloromyotomy or pyloroplasty is planned. These findings should be interpreted in light of the patient's history. If the patient has a history of duodenal ulcer, not only should a pyloroplasty or pyloromyotomy be planned, but a vagotomy to decrease the gastric acid should be added. These findings should be anticipated by careful preoperative workup. Preoperative endoscopy should rule out pyloric stenosis or duodenal ulcer. Only if the duodenum is markedly scarred or if there is active ulceration should pyloroplasty and vagotomy be performed. We generally employ a highly selective vagotomy and Jaboulay pyloroplasty. If the pylorus is simply scarred and there is no active ulceration, the pylorus may be dilated
Figure 7. .The phrenoesophageal membrane is divided on the diaphragm in order to retain the fibroareolar tissue on the stomach to be used in the repair.
and a pyloromyotomy performed. It is imperative to relieve any gastric outlet obstruction to obtain a good result from an antireflux procedure. On the other hand, to add a vagotomy to a routine hiatal hernia repair is unwise. In our experience, this has led to complications of vagotomy without benefit to the patient.
Retracting the stomach to the patient's left exposes the preaortic fascia. The aorta and celiac axis are easily felt. The median arcuate ligament (MAL) lies immediately above the celiac trunk. It can be exposed by careful blunt dissection at this point over the midpoint of the aorta. The celiac artery usually arises cephalad to the median arcuate ligament. When the free edge of the MAL has been exposed, the celiac artery can be compressed into the aorta and the fibroareolar tissue overlying the artery can be carefully divided. An instrument such as a Goodell dilator is then passed beneath the median arcuate ligament. If the instrument is in the correct plane, it should simply float beneath the preaortic fascia. If the instrument meets an obstruction, there may be a branch of the celiac artery in the midline. The branch may be damaged if force is used on insertion.
Dissection of the celiac axis has been the deterrent to performing this operation in the opinion of other surgeons. If it is difficult to locate the MAL and if the surgeon is not familiar with vascular surgery and is uncomfortable dissecting out the celiac axis, a safer alternative procedure is recommended.
By retracting the esophagus to the patient's left, the surgeon can expose the
Figure 8. .The fibroareolar tissue in the esophageal hiatus is divided and a finger is passed down posterior to the preaortic fascia. Sutures may then be placed in the preaortic fascia without dissecting out the median arcuate ligament. The preaortic fascia is lifted off the aorta with a Babcock of stay suture.
esophageal hiatus. The fibroareolar tissue overlying the aorta and the esophageal hiatus can be simply divided by sharp dissection, thereby exposing the aorta. A finger is then passed gently beneath the preaortic fascia, down to the celiac artery and the preaortic fascia can be lifted off the aorta. The fascia can be grasped with a Babcock clamp and sutures simply placed through the preaortic fascia. This is a much simpler and safer approach than dissecting out the celiac artery. This technique was described by Vansant (Fig. 8) and is used quite frequently. In passing the finger behind the fascia, care must be taken not to damage short branches that pass from the aorta to the crura. By staying in the midline, these branches are avoided. We find that this approach is preferable and we now rarely dissect out the median arcuate ligament.
The crura of the esophageal hiatus are loosely approximated behind the esophagus with nonabsorbable sutures. The crura are closed so that a finger can be placed alongside the esophagus, making certain the closure is not too tight.
The stomach is then rotated to expose the anterior and posterior phrenoesophageal bundles. The bundles are grasped with Babcock clamps well above the left gastric artery, taking care not to traumatize the vagal nerves. Strong, nonabsorbable sutures are used for the repair. Sutures are taken through the anterior and posterior phrenoesophageal bundles. These are then passed through the preaortic fascia which is lifted well off the aorta with a Babcock clamp. Usually five sutures are placed in the anterior and posterior phrenoesophageal bundles and carried through the preaortic
Figure 9. .The hiatus is closed loosely about the esophagus and sutures are commenced in the anterior and posterior phrenoesophageal bundles and carried through the preaortic fascia. The median arcuate ligament is not dissected out. Four such sutures are placed and the top suture is tied with a single throw and a knot to allow for pressure measurements and alteration of the sutures according to the pressure obtained.
fascia (Fig. 9). These sutures are placed with the vagus nerves in full view, in order not to damage them. A single knot is then placed in the top three sutures which are then clamped with long hemostats. Measurement of the barrier pressure is then obtained by passing the side hole of the modified nasogastric tube attached to a monitor through the GE junction. If the pressure is above 40 mmHg, the sutures are loosened. If it is below 25 mmHg, the sutures are tightened, depending on the problem at hand. After the proper pressure is obtained, all five sutures are tied and a final pressure measurement is taken. The barrier is usually 3-4 cm long. Additional cardiodiaphragmatic sutures are taken. The final appearance of the repair is shown in Fig. 10. In addition to restoring the sphincter, the GEV is accentuated and can be readily palpated through the wall of the stomach. The valve measures from 3-4 cm along the lesser curve and is important in the prevention of reflux. In patients who have had previous operations with scarring and destruction of the GEJ, the valve may be destroyed or inadequate. In these cases, a gastrostomy is performed, and the valve is secured with sutures in the anterior and posterior edges of the valve, thereby lengthening the valve to 3-4 cm. Attempts to calibrate the cardia with a bougie are unsatisfactory. It is impossible to determine whether the wrap around the bougie is too tight or too loose.
Figure 10. .The final appearance of the repair is shown. All 4-5 sutures are tied, anchoring the GE junction to the preaortic fascia. If the procedure is done open, the GEV can be palpated. If it is done closed, endoscopy is done to check the valve. Cardiodiaphragmatic sutures are placed in to prevent herniation back into the posterior mediastinum.
In 1977 the authors' group first reported the use of a simplified method of measuring the pressure in the antireflux barrier during operation to give the surgeon an objective determination of the pressure that has been created. This measurement is obtained by simply modifying the nasogastric tube which is routinely used in these patients. The tip of the smaller silastic sump portion of the tube is sealed at the end and a 1 mm side hole is cut 18 cm from the tip of the tube (Fig. 11). This small silastic tube is attached to a strain gauge and to a manometer that produces a digital reading. If this manometer is not available, the pressure tube can simply be attached to the arterial line that the anesthesiologist has available. The tube is constantly perfused at a slow rate (0.7 ml/min). This apparatus is identical to the one that has been used in the gastric lab for over a decade and has been thoroughly standardized and used in over 19,000 patients at the institution. The side hole is passed across the GEJ at operation and a baseline pressure is obtained prior to repair. Most often there is no pressure whatever in the GE junction. As the side hole passes through the junction, both a
Figure 11. .Modified nasogastric tube with a side hole which can be passed back and forth across the lower esophageal sphincter during surgery to determine the appropriate barrier pressure.
tracing and a digital readout are obtained. After the repair, if the pressure is over 40 mmHg, the sutures which have been placed are loosened. If the pressure is less than 20 mmHg, the sutures are tightened. This process is continued until a pressure of somewhere between 25 and 35 mmHg is obtained. The side hole must be pulled at a steady rate through the barrier. If it is pulled through too rapidly, a peak pressure will be missed.
In the repair of recurrent hernias, there is no doubt that intraoperative pressure measurement could avoid some of the disastrous complications of the Nissen procedure which occur when the wrap is either too loose or too tight. Further efforts to simplify intraoperative manometrics and to make the technique more readily available are underway. The present technique is safe and simple and requires only a few minutes to obtain valuable information. It is in the authors' opinion that in a major antireflux operation, so dependent upon the construction of an adequate barrier, intraoperative assessment of the barrier should become a standard part of any technique that is used.
It is important to point out that the Hill repair is not a fundoplication. The phrenoesophageal bundles are imbricated together and there is no wrap of stomach around the lower esophagus. Very often this operation is erroneously described as partial fundoplication or a wrap. There is no intention upon the part of the authors of this procedure to do a blind wrap around of the stomach, but rather a careful
calibration of the antireflux barrier, restoration of the GEV and posterior fixation of the GEJ. There is no wrap to slip. The basic differences between the gastroesophageal restoration repair and the Nissen repair are as follows:
- The Hill procedure depends upon augmentation of the intrinsic pressure and its special features. By placing tension on the collarsling musculature, the repair automatically restores the GEV which has been shown to be important in the prevention of reflux. The Nissen repair, on the other hand, depends upon extrinsic pressure of a wrap around the lower esophagus with indirect pressure on the lower esophagus.
- The Hill procedure anchors the GEJ posteriorly to its normal primary attachment, the preaortic fascia. The Nissen repair is allowed to float freely and the GEJ is not anchored. The unanchored esophagus has no fulcrum from which to operate and almost always develops a dysmotility since the esophagus cannot generate propulsive waves.
- In the Hill procedure, no sutures into the esophagus are used because the esophagus has no serosa and no strength. The Nissen procedure employs esophageal sutures to hold the wrap in place. The weakness of these sutures accounts for the frequency of the slipped Nissen. If these sutures are taken deeply, there is a risk of fistula formation from the esophagus.
- In the Hill procedure, intraoperative pressure measurement is used to calibrate the barrier created at operation, giving an objective assessment of the competence of the reflux barrier. This should be used in all repairs, whether it is the Hill, the Belsey or the Nissen. In the Nissen procedure, the surgeon either relies on a bougie or a finger placed up into the esophageal lumen. A number of patients have been seen in whom a large bougie was used only to find that as soon as the bougie was removed the wrap, which was made too tight, simply closed down or the wrap that was made loosely remains open after removal of the bougie.
Several reviews of our own series, and those reported by others, have shown good results for follow-up extending to 8 years. To determine how these early results stand the test of time, the authors conducted a long-term follow-up study. In this recent report of patients undergoing the Hill procedure with reconstruction of the GE junction and restoration of the flap valve for primary GE reflux, 447 patients were reviewed from the time of their initial operation to follow-up assessment. These assessments were conducted after 5 to 10 years and 15 to 20 years. All operations were performed during a 5-year period between 1968 and 1973 by the same surgeon. The mean follow-up was 17.5 years.
All patients (240 men and 207 women; mean age 54.3 years; range 18 to 79 years) had suffered from primary symptoms of GE reflux of heartburn and regurgitation prior to surgery. Twenty-nine of these patients also had significant dysphagia.
Patients rated the results of their operations on a scale ranging from excellent to
poor. The proportion of patients subjectively rating their operative results as good to excellent, increased from 82% during the 1977-1978 evaluation period to 88.5% during the 1987-1988 evaluation period. Of the total number of patients assessed, 95% rated their results as excellent to good with only 5.5% rating their outcome as fair. Eight patients in this series rated their surgical outcome as poor. Only three patients from the original operative series of 447 patients were found to have required reoperation for recurrent symptoms during the follow-up period. Therefore, the anatomical recurrence rate is less than 1%. It was gratifying to find that there were none of the severe complications such as esophageal fistula, thoracic incarceration, obstruction and severe gas bloat, that have been reported with other types of antireflux repairs.
This study clearly shows that a carefully done Hill repair with reconstruction of the GE junction and restoration of the flap valve has proved to be safe and reliable, providing a definitive antireflux repair which stands the test of time. The addition of intraoperative manometry has resulted in an even more precise operation, thus even better long-term results are predicted for the future. A complete follow-up, multicenter study is in progress.
For an operation to be of value, it should be reproducible with good results in other hands. The Hill repair has been performed and reported by a number of surgeons around the world: Csendes and Larrain in Chile, Hermreck and Coates in the United States, along with Thomas, Vansant, Warshaw, Ottinger, Mercer of Canada and Russell of Australia, to name a few. Csendes and Larrain achieved a 93% good results and no radiologic recurrence in 29 patients followed up to 16 months. Mercer has performed the Hill procedure in 110 patients with 95% good results. Vansant et al. have reported treating up to 400 patients with around 90% good results. A multi-institution study is underway which will include around 2,500 patients and the analysis thus far appears to yield a 90% good result over the long term without the complications reported with the Nissen procedure.
Summary and Conclusion
In summary, the Hill procedure includes reconstruction of the normal gastroesophageal junction and restoration of the gastroesophageal valve. This restores the esophagus to its normal point of attachment or fulcrum, allowing it to generate forceful peristaltic waves to propel food aborally into the stomach. In short, the esophageal motility is restored. The sphincter is calibrated, and the pressure measured so that a range of pressure is created that is high enough to prevent reflux but not so high as to create dysphagia. The gastroesophageal valve is restored and the diaphragm is closed loosely about the esophagus. With careful selection of patients and careful performance of the procedure, good results can be obtained over the long term.
23. Reid BJ, Haggitt RC, Rubin CE, Rabinovitch PS. Barrett's esophagus. Correlation between flow cytometry and histology in detection of patients at risk for adenocarcinoma. Gastroenterology 1987:93:1-11.
Larrain A, Pope CE. Respiratory complications of gastroesophageal reflux. In: The Esophagus: Medical and Surgical Management. Hill LD, Kozarek RA, McCallum RW, Mercer CD (eds). Philadelphia: WB Saunders, 1988, pp 70-77. It has become recognized that respiratory complications account for a large portion of the gastroesophageal reflux problem This is a very important review of that complication of reflux.
Low DE, Anderson RP, Ilves R et al. Fifteen- to twenty-year results after the Hill antireflux operation. J Thorac Cardiovasc Surg 1989;98:444-450. This study represents the longest follow-up of antireflux surgery to be published There was a mean follow-up of 17.5 years and the results are excellent considering the fact that these patients were among the early patients done with this technique and were low on the learning curve and done prior to the availability of new technology, including measurement of sphincter pressures.
Low DE, Mercer CD, James EC, Hill LD. Post-Nissen syndrome Surg Gynecol Obstet 1988;l67:l-5. This review represents the largest series of post-Nissen problems from any institution in the world. It represents an excellent survey of the serious problems that are being seen all too frequently with the Nissen procedure.
Mercer CD, Hill LD. Surgical management of peptic esophageal stricture. J Thorac Cardiovasc Surg 1986; 91:371-378. This important article demonstrates clearly that a simplified antireflux procedure will correct most patients with peptic esophageal stricture. Resection is reserved only for those patients who have a destroyed esophagus.
Reid BJ, Haggitt RC, Rubin CE Barrett's esophagus and esophageal adenocarcinoma. In: The Esophagus: Medical and Surgical Management. Hill LD, Kozarek RA, McCallum RW, Mercer CD (eds). Philadelphia: WB Saunders, 1988;157-166. The Reid, Haggitt, and Rubin group has the largest collection of biopsies in Barrett's esophagus in the world From this experience they are able to show that there is a direct relationship between Barrett's esophagus and adenocarcinoma.
Thor K, Hill LD, Mercer CD, Kozarek RA. Reappraisal of the flap valve mechanism: A study of a new valvuloplasty procedure in cadavers. Acta Chir Scand 1987; 153:25-28. This paper represents publication of the cadaver work done to illustrate the importance of the gastroesophageal valve which is the main component of the antireflux barrier. The valve, when visualized with the retroflexed endoscope, gives a better assessment of the clinical picture than measurement of the sphincter pressure.