- Email: [email protected]
Surgical treatment of late postgastrectomy syndromes
Surgical Treatment of Late Postgastrectomy
Syndromes
Howard A. Reber, MD, San Francisco, California Lawrence W. Way, MD, San Francisco, California
Most patients who undergo gastric surgery for peptic ulcer make an uneventful recovery and readjust well to their new anatomic and physiologic situation. A minority do not and develop one or more of the late postgastrectomy syndromes. A small fraction of these patients are so incapacitated that they require surgical reconstruction in an attempt to restore a more tolerable existence. Because the surgical management of severe postgastrectomy syndromes continues to represent a significant and unsolved problem, the extensive experience at the University of California Hospital, San Francisco, was analyzed. Because their diagnosis and therapy rarely overlap the syndromes discussed herein, we have not included in this review patients with marginal ulcer, severe’ anemia: osteomalacia, hypoglycemia, or inadvertent gastroileostomy. The following abbreviations are used throughout this paper: postgastrectomy syndrome, PCS; gastric resection with gastroduodenostomy, Bl; gastric resection with gastrojejunostomy, BII. Clinical Material The records of all patients who were operated on for correction of PCS between January 1963 and December 1973 at the University of California, San Francisco, were
From the Department of Surgery. University of California School of Medicme. and the Department of Surgery, Veterans Administration Hospital, San Francisco, California. Reprint requests should be addressed to Dr Howard A Reber. Department of Surgery (112). VA Hospital, 4150 Clement Street, San Francisco, California 94 12 1. Presented at the Fifteenth Annual Meeting of the Society for Surgery of the Alimentary Tract, San Francisco, California. May 21 and 22. 1974.
Vohne
129, January 1975
reviewed. Of 185 patients with a diagnosis of PGS, 48 underwent corrective surgery a total of 52 times. Because several patients had more than one PCS diagnosed, a total of 59 PGS were analyzed. All patients were followed up by telephone interview or office visit during November and December 1973. Surgical Procedures and Results Initial Gvstric Procedure. Of the forty-eight patients reviewed, nineteen were men and twentynine were women. The ages at the time of the ciriginal surgery were evenly distributed from twenty to fifty years, and about 10 per cent of the patients were over fifty at the time of the first operation. Thirty patients had duodenal ulcer, ten had gastric ulcer, and in eight patients no ulcer was identified. Thirty-seven patients (71 per cent), the largest group, were operated on for intractable pain. The next largest group consisted of patients with acute (eight) or chronic hemorrhage (ten); bleeding at some time, therefore, accounted for the original procedure in 35 per cent of the series. Six patients were operated on because of persistent gastric ulcer in order to rule out a malignant lesion (IQ per cent of series). TWO patients had a perforated ulcer, one acutely and one previously. One patient required the original operation because of _
obstruction. The original operative procedures are listed in Table 1. In one patient a new PCS arose after conversion of a BII to a BI which itself had been performed to correct a PGS. Two patients had 85 per cent BI gastrectomy, one patient had pkimary vagotomy and BI gastrectomy with isoperistaltic jejunal interposition, and one patient had 75 per
71
Reber and Way
TABLE
I
Initial Gastric Patients
-_
Syndromes ___~_
_
Procedure
with Fifty-Nine (Fifty-Two
in Forty-Eight Postgastrectomy
Corrective
Operation* I311gastrectomy Vagotomy and Vagotomy and Vagotomy and Vagotomy and
Number -__
BII antrectomy BI antrectomy pyloroplasty gastrojejunostomy
Conversion from BI I to BI Vagotomy and BI gastrectomy with isoperistaltic jejunal interposition Subtotal gastrectomy and BI with gastrojejunostomy *All
vagotomies
Operations)
20 17 7 1 1 1 1 1
in the series were truncal.
cent BI gastrectomy
and gastrojejunostomy to correct an early postoperative stoma1 obstruction at the gastroduodenostomy. In all of these primary operations there was a high incidence of postoperative complications (18 per cent) requiring reoperation. Seven of these were obstructions at the newly created stoma, and the remaining two were suture line leaks. For many patients the over-all result in the early postoperative months was unsatisfactory since 40 per cent of the entire group were not relieved of the major preoperative symptoms. Postgastrectomy Symptoms. Loss of weight was present in forty-two patients. Six patients (14 per cent) had lost less than 10 pounds after the initial operation, another six (14 per cent) had lost 10 to 20 pounds, ten (24 per cent) had lost 20 to 30 pounds, and twenty patients (48 per cent) had lost in excess of 30 pounds.
TABLE
II
Corrective Operations and Results in Sixteen Patients with Stoma1 Obstruction Procedure
Complete reconstruction Stoma1 revision
Total
Improved*
13 3
12 0
* 75 per cent. TABLE
III
Corrective Operations and Results in Fifteen Patients with the Dumping Syndrome Procedure
Conversion from BII to BI lsoperistaltic jejunal conversion Antiperistaltic jejunal conversion Roux-en-Y * 53 per cent.
72
Total 9 4 1 1
improved* -_5 2 1 0
Eighty-eight per cent of the patients began to have symptoms of PGS within the first year of surgery, although in many cases the severity continued to increase up to the time of the corrective operation. Nevertheless in five patients (10 per cent), symptoms began three years or more after the initial operation (two patients with bilious vomiting, two with malabsorption, and one with stoma1 obstruction). Thirty-five patients (74 per cent) had one reconstructive procedure to correct the PGS, seven patients (15 per cent) had two corrective procedures, and five patients (11 per cent) had three. Stoma1 Obstruction. Sixteen patients (28 per cent) had stoma1 obstruction, the largest single group of PGS. The diagnosis rested on a history of postprandial distention and fullness, usually relieved by vomiting of undigested food. In all cases there was radiographic and/or endoscopic evidence of stoma1 obstruction and gastric retention. The original operations in this group were BII gastrectomy (eight patients), vagotomy and BI antrectomy (six patients), pyloroplasty and vagotomy (one patient), and vagotomy and RI gastric resection, with isoperistaltic jejunal interposition (one patient). Obstructive symptoms developed in all but one of these fifteen patients (96 per cent) within six months of the original operation; the remaining patient did not manifest symptoms until three years later. However, only t,hree were reoperated on within a year of the original procedure. Eleven (69 per cent) had corrective surgery after one to five years, and two (13 per cent), not until eight years later. Three patients required another operation after the first attempt to relieve the stoma1 obstruction. In two of these cases stoma1 revision only had been performed. The corrective procedures themselves varied depending on the original operation but in genera1 could be classified as either stoma1 revision or total reconstruction. (Table II.) Twelve of the patients with stoma1 obstruction were rendered asymptomatic (two patients) or markedly improved (ten patients), an over-all SUCcess rate after reoperation of 75 per cent. Four patients (25 per cent) did not benefit. Of the entire group, two patients later required additional surgery for PGS; one had recurrent stoma1 obstruction and one had the dumping syndrome. Dumping. Fifteen patients had severe dumping syndrome manifested by classic cardiovascular and gastrointestinal symptoms within fifteen minutes after eating. Eight, were men and seven were
The American Journal of Surgery
Postgastrectomy
women, a reversal of the sex ratio for the entire PGS series in which women predominated in a ratio of 3:2. The original procedures in this group included vagotomy and BII antrectomy (ten patients) and subtotal BII gastric resection (five patients). Symptoms of dumping began within one year after surgery in all patients. Upper gastrointestinal x-ray films showed rapid gastric emptying in about one third the cases; the remainder had no radiologic abnormality. Although virtually all patients showed some improvement with dietary alteration, it was either temporary or insufficient to avoid reoperation. The reconstructive procedures included conversions of BII to BI in nine cases, conversions of BII to BI with isoperistaltic jejunal interposition of a 10 cm segment in four cases, conversion with antiperistaltic jejunal interposition in one case, and Roux-en-Y gastrojejunostomy in one case. (Table III.) Eight patients (53 per cent) were markedly improved. Two patients .with isoperistaltic interpositions, four patients with BII-BI conversions, and the patient with a Roux-en-Y were not improved. Bilious Vomiting. Bilious vomiting was diagnosed in eleven patients (three men and eight women) on the basis of abdominal distention and/ or pain initiated by eating and relieved by vomiting bile-stained material. In this group, eight patients had BII gastric resection and three had vagotomy and BII antrectomy. Four patients (36 per cent) had undergone reoperation after gastrectomy, two for stoma1 obstruction, one for anastomotic leak, and one for hemorrhage at an anastomotic suture line. Upper gastrointestinal x-ray films showed a dilated, poorly emptying afferent loop in four patients but in the remainder the loop either did not fill (three) or was normal (four). A positive secretin-cholecystokinin stimulation test result was obtained in the only patient in whom it was employed. In nine patients (82 per cent) symptoms developed within the first year of surgery, but two patients were asymptomatic until three years had elapsed. At corrective surgery, four patients had obvious anatomic obstruction in the afferent loop, but the majority (seven) did not. The following corrective procedures were employed (Table IV): conversion of a BII to a BI in seven patients, five of whom improved (71 per cent) and 2 of whom were no better (29 per cent); conversion of a BII to a BI with isoperistaltic jejunal (10 cm) interposition in two patients, both of whom had unsuccessful results. A Roux-en-Y gastrojejunostomy improved the only patient in whom it was performed; one patient was treated
volume 129, January 1975
Syndromes
by stoma1 revision which did not change the preoperative complaints. The over-all success rate for those with bilious vomiting included improvement or cure in 55 per cent and no change or a worse condition in 45 per cent.. Four patients required reoperation later for new or recurring symptoms: reflux alkaline gastritis developed in two, stoma1 obstruction in one, and the dumping syndrome in one. Malabsorption. There were five patients with malabsorption, four men and one woman. At. the original operation, four had vagotomy and BII antrectomy and one had BII gastric resection. None had immediate postoperative complications and all had complete relief of the principal preoperative symptoms. In three of the patients, severe diarrhea developed within six months after the original operation, but two had no diarrhea or weight loss until three years had elapsed. By the time corrective surgery was undertaken, one patient had lost 10 pounds but the remaining four had lost 35, 35, 50, and 65 pounds, respectively. Steatorrhea was documented in all five patients, with two exhibiting 20 to 30 per cent fecal fat loss and three, between 30 and 40 per cent. All five patients were given a course of tetracycline and three of the five had a clear-cut symptomatic response. In each case the improvement was either temporary or clinically insufficient. All of these patients had conversion of a BII to a BI and each was improved. Diarrhea was either eliminated or decreased to a manageable level and all patients gained weight to near or above normal levels. Diarrhea. In five patients (one man and four women) diarrhea was the principal reason for corrective operation. Four of the original procedures included truncal vagotomy, two with BII antrectomy, one with BI antrectomy, and one with gastrojejunostomy. There was also an 85 per cent BI gastrectomy with jejunal interposition. In four patients, diarrhea occurred within one year of surgery; in one it occurred two years postoperatively. Weight loss was 20 to 30 pounds in four patients
TABLE
IV
Corrective Operations and Results in Eleven Patients with Bilious Vomiting Procedure
Conversion from BII to BI Conversion with isoperistaltic jejunal interpositibn Roux-en-Y Stoma1 revision
_
Total
Improved*
7
5
2 1 1
0 1 0
* 55 per cent.
73
Reber
TABLE
and Way
V
Corrective Operations and Results in Five Patients with Diarrhea ~..____ Procedure
Conversion from BII to BI Conversion from BI to BII Takedown of jejunal interposition
Total
Improved*
3 1 1
3 1 0
* 80 per cent.
and in excess of 30 pounds in one patient. Each complained of loose, watery stools occurring from ten to twelve times daily. There had been no response to codeine, diphenoxylate, or dietary manipulation. Fecal fat excretion was less than 10 per cent in all of these patients. Corrective operations were conversions from a BI to a BII (one patient), and a BII to a BI (three patients). (Table V.) These four patients improved to the point where there were fewer than five stools daily and their weight level before diarrhea was regained. The fifth patient, treated by takedown of a jejunal interposition, did not improve. Thus, the over-all results in this group were satisfactory in 80 per cent. Reftux Alkaline Gastritis. Bile reflux with alkaline gastritis was responsible for reoperation in three patients, one man and two women, all with severe, continuous epigastric distress aggravated by meals and unrelieved by vomiting. Endoscopy revealed marked gastritis with hypertrophic gastric rugae and free reflux of bile into the gastric lumen. The original operations were vagotomy and BI antrectomy in two patients and conversion of a BII subtotal gastrectomy to a BI in one patient in whom bilious vomiting developed. In each case the symptoms appeared within one year after the original operation and reoperative surgery was undertaken within three years of this time. Cholestyramine was of no value in the one patient in whom it was used. Preoperative gastric analyses showed less than 1.5 mEq H+/hr basally. The corrective procedures were Roux-en-Y gastrojejunostomy in two cases and interposition of an isoperistaltic 15 cm jejunal segment between the stomach and duodenum in one case. All patients were markedly improved postoperatively. Miscellaneous. Two patients had symptoms and clinical findings that did not allow them to be placed confidently in any of the preceding PGS categories. They both had multiple gastrointestinal symptoms including pain, nausea, and emesis. Preoperative diagnosis was vague and the rationale of converting a BII to a BI in each of these patients was unclear. Neither was improved by the procedure.
74
One patient was thought to have jejunogastric intussusception because of abdominal pain and radiologic findings. Conversion of a BII to a BI relieved the pain for three years but it returned and another operation was subsequently performed elsewhere. She still has pain. One woman who had vagotomy and BII antrectomy for benign gastric ulcer three years later underwent a BII to BI conversion with interposition of an isoperistaltic 25 cm jejunal segment. Shortly thereafter constant pain developed in the left upper quadrant. X-ray films revealed an anastomotic ulcer, and basal acid output was 9 mEq/hr. The interposition was resected, a residual vagal trunk severed, and gastroduodenostomy performed. Although this cured the ulcer, she continues to have abdominal pain and diarrhea. Serum gastrin levels are within normal limits. Comments
The patient with late postgastrectomy complaints typically presents a perplexing clinical picture with multiple symptoms. The initial task of the physician is to sort out each of these symptoms carefully and to classify the patient whenever possible into one specific PGS category. Since treatment and prognosis for the syndromes may differ, the importance of adequate preoperative evaluation should be evident. In the sections that follow, each of the syndromes is discussed and the appropriate diagnostic approaches are outlined. Stoma1 Obstruction. Stoma1 obstruction may occur early or late. The early variety, associated with postoperative edema, is usually self-limited. Chronic obstruction may result from too narrow an anastomosis, unresolved edema, or recurrent ulceration of the stoma. Some authors [I ] consider stoma1 obstruction to be a rare cause of late postgastrectomy symptoms, but our experience is just the opposite. As an indication for reoperation, it was more common than dumping or bilious vomiting. Diagnosis was based on a history of pain, vomiting, or both after meals, but required endoscopic confirmation of stoma1 narrowing and x-ray evidence of delay in the passage of barium. cJejunal intussusception is an unusual cause for stoma1 obstruction in which diagnosis is more difficult because of the intermittent nature of’ the problem. Stoma1 obstruction is a symptom complex that has a clear anatomic basis and should be correctable in all cases. Our experience indicates t.hat total reconstruction of the anastomosis rather than attempts at stoma1 revision are more likely to be successful.
The American Journal of Surgery
Postgastrectomy
Too rapid emptying of Dumping Syndrome. gastric contents is generally accepted as the underlying cause of dumping, and the operations designed to correct it must slow gastric emptying to be successful. Three operations have been widely used: simple conversion from a BII to a BI anastomosis, conversion with interposition of a 10 cm isoperistaltic segment of jejunum, and interposition of a similar segment in an antiperistaltic fashion. The clinical experience with simple conversion has been variable. Woodward and Hastings [Z ] report satisfactory results in twelve of fifteen patients and consider the small diameter of the gastroduodenal stoma to be the most important factor. In the present series, nine patients were treated by simple conversion, five with improvement (56 .per cent); no patient was cured. This is in keeping with the experience of most authors [3,4] who find significant improvement in about half the patients. Isoperistaltic jejunal interposition produced improvement or cure in eleven of thirtyone patients in the series reported by Fenger, Kallehauge, and Gudmand-Hoyer [5] and nine of those eleven patients originally had a BII anastomosis. Only two of twelve patients with BI gastrectomy were improved. They concluded that the majority of patients who were improved had symptoms from partial afferent loop obstruction as well as dumping and that for pure dumping, isoperistaltic jejunal interposition was not effective. This reinforces our contention that accurate diagnosis is critical in predicting operative results in these patients and that when accurate diagnosis is impossible, a successful outcome is uncommon. Half our patients (two) who underwent this procedure had satisfactory results. Although there is reason to believe [6] that slower gastric emptying should result from isoperistaltic jejunal interposition compared with gastroduodenostomy, the results in our series indicate that the effect, if real, often must he inadequate. On the basis of our over-all success rate of 53 per cent, we would support a wider trial of reversed jejunal interposition, an operation that retards gastric emptying more than these other procedures [7,12]. Our patient treated with .this procedure was markedly improved. Recent observations that the incidence of dumping may be greater after selective vagotomy compared with truncal vagotomy suggest that we may encounter more patients in the near future who require surgical relief for this problem. Of twenty patients who had selective vagotomy at our institution, in the past four years, severe dumping has developed in two and one has
Syndromes
been scheduled for a reversed jejunal segment in the near future. Bilious Vomiting. There has been an unwarranted tendency to explain all bilious vomiting as representing examples of the afferent loop syndrome [8] w h’ic h involves partial obstruction of the afferent loop, accumulation of bile and pancreatic juice proximal to the stoma, and sudden bilious vomiting as pressure in the afferent loop overcomes resistance and bile floods the gastric remnant. Alexander- Williams [S ] has offered evidence that vomiting in some patients is due to topical effects of bile on the gastric mucosa. In these cases, reflux, not obstruction, is the primary abnormality. Also included among those who complain of vomiting are patients with stoma1 obstruction, efferent limb obstruction, and alkaline gastritis and a significant group in whom no specific cause can be identified. Often those in this last category will have had vomiting (without organic obstruction) as a prominent complaint even before the initial operation. They are surgically incurable but sometimes hard to identify. The first step is to search for bona fide signs of obstruction. If, for example, persistent narrowing and slowed passage of barium can be demonstrated at some point, the diagnosis and appropriate therapy are usually obvious. More often signs of obstruction are either absent or equivocal. We believe that such patients are best classified as having bilious vomiting, a more general term than afferent loop syndrome. In some cases, bile reflux is responsible, and in other, partial obstruction, but since it may be impossible to distinguish definitely between these possibilities, surgical treatment should attempt to correct them both. This can be accomplished with Roux-en-Y BII reconstruction or isoperistaltic BI jejunal interposition; conversion of a BII to a BI still allows reflux and is therefore less successful. In the present series, conversion improved five patients, two jejunal interpositions were unsuccessful, and one Roux-en-Y gastrojejunostomy was effective. It must be remembered that operations that keep bile away from the gastrointestinal anastomosis are potentially ulcerogenic and should be combined with vagotomy if gastric analysis reveals the presence of significant acid secretion. In fact, some [9] have even suggested that in the absence of acid, vagotomy should be performed since recovery of acid secretory capacity may occur after bile reflux is eliminated. One patient was referred to us with a marginal ulcer that developed after
Reber and Way
conversion from a BIT gastrectomy to a BI with jejunal interposition. If preoperative studies of acid secretion had been performed, the predisposition towards marginal ulcer could have been predicted and interposition either avoided or accompanied by vagotomy. Malabsorption and Diarrhea. Whenever weight loss or diarrhea is prominent after gastric operations, one of the first steps should be an investigation of the efficiency of intestinal absorption. Measurement of daily fecal fat excretion should be performed while the patient is given a diet containing 100 gm of fat per day. BII gastrectomy alters the anatomy, so that mixing between ingested food and digestive enzymes is reduced and the average asymptomatic patient can be shown postoperat ively to increase fat excretion from the normal of 5 per cent to about 10 per cent. This amount of malabsorption by itself is insignificant, but it probably reduces the patient’s threshold for serious nutritional difficulty if defects develop in the micellar or mucosal uptake phases of absorption. Postgastrectomy patients with malnutrition due to malabsorption will almost always excrete over 20 per cent of dietary fat. Once malabsorption has been established, it is necessary to look for bacterial overgrowth from stasis in the afferent loop, pancreatic insufficiency, and sometimes intestinal mucosal dysfunction. The former patients may be diagnosed by a response to tetracycline therapy, a positive CL” cholate breath test, or direct sampling of afferent loop contents for bacteriologic study. Those with pancreatic insufficiency may respond to pancreatic enzyme replacement with increased efficiency of absorption. When pancreatic insufficiency seems to account for the malabsorption, it will most often represent an exaggeratioil of the mechanism outlined above, with bile and enzymes arriving at the stoma only after the food has already passed into the efferent limb. These patients can be effectively treated surgically. A smaller number have pancreatic insufficiency on the basis of intrinsic pancreatic disease. This group can usually be identified from the rest because the parenchymal destruction frequently is accompanied by endocrine insufficiency and visible calcifications on abdominal x-ray films. In the few patients with neither bacterial overgrowth nor pancreatic insufficiency of a marked degree, additional studies, such as d-xylose absorption tests may indicate nontropical and jejunal biopsy, sprue, lactase deficiency, or some other absorptive abnormality. Overall, the majority of patients with malabsorption will become candidates for surgical
reconstruction since both afferent loop stasis and inefficient mixing can be corrected. In the present series, all five patients with malabsorption were benefited by conversion operations. Other authors [10,12 1 have also reported high rates of success with this or similar operations. The common feature of tile successful procedures is elimination of the afferent loop: postoperatively, food passing through the previous blind loop wipes the bacteria along with it, stasis is no longer present, and food once again enters the intestine at the same point as hile and pancreatic enzymes. In the absence of malabsorption, patients with uncontrollable diarrhea usually have postvagotomy diarrhea. If codeine or dietary manipulation fails and the problem is severe, reconstruction may be considered. Although four of five patients in this series (80 per cent) had good results, the reason for the improvement is difficult to explain. There were three conversions from RI1 to BI and one from BI to BII. The reversal of a segment of midjejunum, which is reported to improve this condition, may deserve wider use, but we have had no experience with it. At, present the pathophysiologic aspects of postvagotomy diarrhea are poorly understood. It is noteworthy that parietal cell vagotomy without pyloroplasty seems free of postoperative diarrhea, suggesting that uncontrolled gastric emptying is an important contributing factor. Reflux Gastritis. With widespread use of gastroscopy, reflux alkaline gastritis has been identified as a distinct postgastrectomy syndrome. The patients present with pain, an endoscopic picture of severe gastritis and edematous rugae, low acid secretion, and perhaps bilious vomiting, In the treatment of this disorder, excellent results have been reported after Roux-en-Y gastr<).jejunostomy or interposition operations (6,.Y,Z.‘$]. Our three patients had uniformly excellent results from either of these procedures. The common feature here is diversion of bile and pancreatic secretions away from the gastric stoma so that reflux c,annot occur. Again, vagotomy should be added if it has not already heen performed.
Conclusions
With some justification, an over-all the results of surgical correction of tomy syndromes has been pessimistic. is gradually becoming apparent how ous these patients are despite the
appraisal of postgastrecHowever, it heterogenesimilarity of
Postgastrectomy
their subjective complaints. As a result, it is possible to more accurately tailor the surgical reconstruction to suit the problem and in some groups to achieve a high degree of success. At present, most patients with stoma1 obstruction, malabsorption, and alkaline gastritis can be helped by further surgery. Roux-en-Y gastrojejunostomy or isoperistaltic jejunal interposition with BI is recommended for bilious vomiting and alkaline gastritis; conversion from BII to BI for malabsorption; and total stoma1 reconstruction for stoma1 obstruction, Reconstructions for dumping or bilious vomiting have been less successful, but different approaches are being tested and some seem quite promising. We believe that the dumping syndrome should be treated by a 10 cm antiperistaltic jejunal interposition with BI. There will probably always be a few patients who do not fit clearly into any of the presently recognized syndromes. Improvement from any reconstructive procedures in this group is usually doomed to failure. Summary
An analysis of the experience at the University of California Hospital, San Francisco, with the diagnosis and treatment of fifty-nine postgastrectomy syndromes shows that stoma1 obstruction (sixteen patients), the most common syndrome, was best treated by total reconstruction rather than stoma1 revision. Disappointment with the results of other procedures for the dumping syndrome (50 per cent improvement) has convinced us of the need to adopt the use of reversed jejunal interposition for surgical treatment of this condition. Because a precise etiologic diagnosis of bilious vomiting is often elusive, the preferred procedure is isoperistaltic jejunal interposition, since it eliminates the afferent loop and prevents bile from entering the stomach. All five patients with malabsorption were improved by conversion from BII to BI. Four of five patients with diarrhea were im-
Valumm 129. Januarv
1275
Syndromes
proved by various procedures (not including a reversed segment of intestine in the midjejunum). Three patients with reflux alkaline gastritis were improved by Roux-en-Y gastrojejunostomy or isoperistaltic jejunal interposition. Either is effective. Thus, in our experience if an unquestionable diagnosis of stoma1 obstruction, malabsorption, or reflux alkaline gastritis can be established, there is ample justification for an optimistic outlook regarding surgical therapy.
References 1. Hirschowitz BI: Classification of the post-gastrectomy syndromes based on observation of 580 patients. Ala J Med SC; a: 50, 1971. 2. Woodward ER, Hastings N: Surgical treatment of the postgastrectomy dumping syndrome. Surg Gynecol Obstet 111: 429, 1960. 3. Andreasson M: Surgical treatment of severe dumping syndrome: conversion of Billroth il into Billroth I. Acta Chir Stand SuppI p 222. 1961. 4. Borg I, Borgstrom SG, Haeger K: The value of the 811-81conversion operation in the treatment of the postgastrectomy syndrome. Acta Chir Stand 134: 655, 1968. 5. Fenger HJ, Kallehauge HE, Gudmand-Hoyer E: Pathophysiological studies of the effect of isoperistaltic jejunal interposition on the treatment of the incapacitating dumping syndrome. Stand J Gastroenterol 7: 283, 1972. 6. Alexander-Williams J: Gastric reconstructive surgery. Ann R Co/l Surg En@ 52: 1, 1973. 7. Herrington JL, Sawyers JL: A new operation for the dumping syndrome and post-vagotomy diarrhea. Ann Surg 175: 790, 1972. 8. Woodward ER: The pathophysiology df the afferent loop syndrome. Surg C/in N&th Am 46: 411, 1966. 9. Van Heerden JA, Priestly JT, Farrow GM, Philips SF: Postoperative alkaline reflux gastritis: surgical implications. Am J Surg 118: 427, 1969. 10. Alexander-Williams J: The place of surgery in the treatment of postgastrectomy syndromes. Proc R Sot Med 57: 585, 1964. 11. Fineberg C, Templeton JY, Wirts CW, Goldstein F: The correction of post gastrectomy malabsorption by jejunal interposition. Surg Clin North Am 53: 581, 1973. 12. Jordan GL: Surgical management of postgastrectomy problems. ArchSurg 102: 251, 1971. 13. Joseph WL, Rivera RA, O’Kieffe DA, Geelhoed GW, McCune WS: Management of postoperative alkaline reflux gastritis. Ann Surg 177: 655, 1973.
77
Syndromes
Howard A. Reber, MD, San Francisco, California Lawrence W. Way, MD, San Francisco, California
Most patients who undergo gastric surgery for peptic ulcer make an uneventful recovery and readjust well to their new anatomic and physiologic situation. A minority do not and develop one or more of the late postgastrectomy syndromes. A small fraction of these patients are so incapacitated that they require surgical reconstruction in an attempt to restore a more tolerable existence. Because the surgical management of severe postgastrectomy syndromes continues to represent a significant and unsolved problem, the extensive experience at the University of California Hospital, San Francisco, was analyzed. Because their diagnosis and therapy rarely overlap the syndromes discussed herein, we have not included in this review patients with marginal ulcer, severe’ anemia: osteomalacia, hypoglycemia, or inadvertent gastroileostomy. The following abbreviations are used throughout this paper: postgastrectomy syndrome, PCS; gastric resection with gastroduodenostomy, Bl; gastric resection with gastrojejunostomy, BII. Clinical Material The records of all patients who were operated on for correction of PCS between January 1963 and December 1973 at the University of California, San Francisco, were
From the Department of Surgery. University of California School of Medicme. and the Department of Surgery, Veterans Administration Hospital, San Francisco, California. Reprint requests should be addressed to Dr Howard A Reber. Department of Surgery (112). VA Hospital, 4150 Clement Street, San Francisco, California 94 12 1. Presented at the Fifteenth Annual Meeting of the Society for Surgery of the Alimentary Tract, San Francisco, California. May 21 and 22. 1974.
Vohne
129, January 1975
reviewed. Of 185 patients with a diagnosis of PGS, 48 underwent corrective surgery a total of 52 times. Because several patients had more than one PCS diagnosed, a total of 59 PGS were analyzed. All patients were followed up by telephone interview or office visit during November and December 1973. Surgical Procedures and Results Initial Gvstric Procedure. Of the forty-eight patients reviewed, nineteen were men and twentynine were women. The ages at the time of the ciriginal surgery were evenly distributed from twenty to fifty years, and about 10 per cent of the patients were over fifty at the time of the first operation. Thirty patients had duodenal ulcer, ten had gastric ulcer, and in eight patients no ulcer was identified. Thirty-seven patients (71 per cent), the largest group, were operated on for intractable pain. The next largest group consisted of patients with acute (eight) or chronic hemorrhage (ten); bleeding at some time, therefore, accounted for the original procedure in 35 per cent of the series. Six patients were operated on because of persistent gastric ulcer in order to rule out a malignant lesion (IQ per cent of series). TWO patients had a perforated ulcer, one acutely and one previously. One patient required the original operation because of _
obstruction. The original operative procedures are listed in Table 1. In one patient a new PCS arose after conversion of a BII to a BI which itself had been performed to correct a PGS. Two patients had 85 per cent BI gastrectomy, one patient had pkimary vagotomy and BI gastrectomy with isoperistaltic jejunal interposition, and one patient had 75 per
71
Reber and Way
TABLE
I
Initial Gastric Patients
-_
Syndromes ___~_
_
Procedure
with Fifty-Nine (Fifty-Two
in Forty-Eight Postgastrectomy
Corrective
Operation* I311gastrectomy Vagotomy and Vagotomy and Vagotomy and Vagotomy and
Number -__
BII antrectomy BI antrectomy pyloroplasty gastrojejunostomy
Conversion from BI I to BI Vagotomy and BI gastrectomy with isoperistaltic jejunal interposition Subtotal gastrectomy and BI with gastrojejunostomy *All
vagotomies
Operations)
20 17 7 1 1 1 1 1
in the series were truncal.
cent BI gastrectomy
and gastrojejunostomy to correct an early postoperative stoma1 obstruction at the gastroduodenostomy. In all of these primary operations there was a high incidence of postoperative complications (18 per cent) requiring reoperation. Seven of these were obstructions at the newly created stoma, and the remaining two were suture line leaks. For many patients the over-all result in the early postoperative months was unsatisfactory since 40 per cent of the entire group were not relieved of the major preoperative symptoms. Postgastrectomy Symptoms. Loss of weight was present in forty-two patients. Six patients (14 per cent) had lost less than 10 pounds after the initial operation, another six (14 per cent) had lost 10 to 20 pounds, ten (24 per cent) had lost 20 to 30 pounds, and twenty patients (48 per cent) had lost in excess of 30 pounds.
TABLE
II
Corrective Operations and Results in Sixteen Patients with Stoma1 Obstruction Procedure
Complete reconstruction Stoma1 revision
Total
Improved*
13 3
12 0
* 75 per cent. TABLE
III
Corrective Operations and Results in Fifteen Patients with the Dumping Syndrome Procedure
Conversion from BII to BI lsoperistaltic jejunal conversion Antiperistaltic jejunal conversion Roux-en-Y * 53 per cent.
72
Total 9 4 1 1
improved* -_5 2 1 0
Eighty-eight per cent of the patients began to have symptoms of PGS within the first year of surgery, although in many cases the severity continued to increase up to the time of the corrective operation. Nevertheless in five patients (10 per cent), symptoms began three years or more after the initial operation (two patients with bilious vomiting, two with malabsorption, and one with stoma1 obstruction). Thirty-five patients (74 per cent) had one reconstructive procedure to correct the PGS, seven patients (15 per cent) had two corrective procedures, and five patients (11 per cent) had three. Stoma1 Obstruction. Sixteen patients (28 per cent) had stoma1 obstruction, the largest single group of PGS. The diagnosis rested on a history of postprandial distention and fullness, usually relieved by vomiting of undigested food. In all cases there was radiographic and/or endoscopic evidence of stoma1 obstruction and gastric retention. The original operations in this group were BII gastrectomy (eight patients), vagotomy and BI antrectomy (six patients), pyloroplasty and vagotomy (one patient), and vagotomy and RI gastric resection, with isoperistaltic jejunal interposition (one patient). Obstructive symptoms developed in all but one of these fifteen patients (96 per cent) within six months of the original operation; the remaining patient did not manifest symptoms until three years later. However, only t,hree were reoperated on within a year of the original procedure. Eleven (69 per cent) had corrective surgery after one to five years, and two (13 per cent), not until eight years later. Three patients required another operation after the first attempt to relieve the stoma1 obstruction. In two of these cases stoma1 revision only had been performed. The corrective procedures themselves varied depending on the original operation but in genera1 could be classified as either stoma1 revision or total reconstruction. (Table II.) Twelve of the patients with stoma1 obstruction were rendered asymptomatic (two patients) or markedly improved (ten patients), an over-all SUCcess rate after reoperation of 75 per cent. Four patients (25 per cent) did not benefit. Of the entire group, two patients later required additional surgery for PGS; one had recurrent stoma1 obstruction and one had the dumping syndrome. Dumping. Fifteen patients had severe dumping syndrome manifested by classic cardiovascular and gastrointestinal symptoms within fifteen minutes after eating. Eight, were men and seven were
The American Journal of Surgery
Postgastrectomy
women, a reversal of the sex ratio for the entire PGS series in which women predominated in a ratio of 3:2. The original procedures in this group included vagotomy and BII antrectomy (ten patients) and subtotal BII gastric resection (five patients). Symptoms of dumping began within one year after surgery in all patients. Upper gastrointestinal x-ray films showed rapid gastric emptying in about one third the cases; the remainder had no radiologic abnormality. Although virtually all patients showed some improvement with dietary alteration, it was either temporary or insufficient to avoid reoperation. The reconstructive procedures included conversions of BII to BI in nine cases, conversions of BII to BI with isoperistaltic jejunal interposition of a 10 cm segment in four cases, conversion with antiperistaltic jejunal interposition in one case, and Roux-en-Y gastrojejunostomy in one case. (Table III.) Eight patients (53 per cent) were markedly improved. Two patients .with isoperistaltic interpositions, four patients with BII-BI conversions, and the patient with a Roux-en-Y were not improved. Bilious Vomiting. Bilious vomiting was diagnosed in eleven patients (three men and eight women) on the basis of abdominal distention and/ or pain initiated by eating and relieved by vomiting bile-stained material. In this group, eight patients had BII gastric resection and three had vagotomy and BII antrectomy. Four patients (36 per cent) had undergone reoperation after gastrectomy, two for stoma1 obstruction, one for anastomotic leak, and one for hemorrhage at an anastomotic suture line. Upper gastrointestinal x-ray films showed a dilated, poorly emptying afferent loop in four patients but in the remainder the loop either did not fill (three) or was normal (four). A positive secretin-cholecystokinin stimulation test result was obtained in the only patient in whom it was employed. In nine patients (82 per cent) symptoms developed within the first year of surgery, but two patients were asymptomatic until three years had elapsed. At corrective surgery, four patients had obvious anatomic obstruction in the afferent loop, but the majority (seven) did not. The following corrective procedures were employed (Table IV): conversion of a BII to a BI in seven patients, five of whom improved (71 per cent) and 2 of whom were no better (29 per cent); conversion of a BII to a BI with isoperistaltic jejunal (10 cm) interposition in two patients, both of whom had unsuccessful results. A Roux-en-Y gastrojejunostomy improved the only patient in whom it was performed; one patient was treated
volume 129, January 1975
Syndromes
by stoma1 revision which did not change the preoperative complaints. The over-all success rate for those with bilious vomiting included improvement or cure in 55 per cent and no change or a worse condition in 45 per cent.. Four patients required reoperation later for new or recurring symptoms: reflux alkaline gastritis developed in two, stoma1 obstruction in one, and the dumping syndrome in one. Malabsorption. There were five patients with malabsorption, four men and one woman. At. the original operation, four had vagotomy and BII antrectomy and one had BII gastric resection. None had immediate postoperative complications and all had complete relief of the principal preoperative symptoms. In three of the patients, severe diarrhea developed within six months after the original operation, but two had no diarrhea or weight loss until three years had elapsed. By the time corrective surgery was undertaken, one patient had lost 10 pounds but the remaining four had lost 35, 35, 50, and 65 pounds, respectively. Steatorrhea was documented in all five patients, with two exhibiting 20 to 30 per cent fecal fat loss and three, between 30 and 40 per cent. All five patients were given a course of tetracycline and three of the five had a clear-cut symptomatic response. In each case the improvement was either temporary or clinically insufficient. All of these patients had conversion of a BII to a BI and each was improved. Diarrhea was either eliminated or decreased to a manageable level and all patients gained weight to near or above normal levels. Diarrhea. In five patients (one man and four women) diarrhea was the principal reason for corrective operation. Four of the original procedures included truncal vagotomy, two with BII antrectomy, one with BI antrectomy, and one with gastrojejunostomy. There was also an 85 per cent BI gastrectomy with jejunal interposition. In four patients, diarrhea occurred within one year of surgery; in one it occurred two years postoperatively. Weight loss was 20 to 30 pounds in four patients
TABLE
IV
Corrective Operations and Results in Eleven Patients with Bilious Vomiting Procedure
Conversion from BII to BI Conversion with isoperistaltic jejunal interpositibn Roux-en-Y Stoma1 revision
_
Total
Improved*
7
5
2 1 1
0 1 0
* 55 per cent.
73
Reber
TABLE
and Way
V
Corrective Operations and Results in Five Patients with Diarrhea ~..____ Procedure
Conversion from BII to BI Conversion from BI to BII Takedown of jejunal interposition
Total
Improved*
3 1 1
3 1 0
* 80 per cent.
and in excess of 30 pounds in one patient. Each complained of loose, watery stools occurring from ten to twelve times daily. There had been no response to codeine, diphenoxylate, or dietary manipulation. Fecal fat excretion was less than 10 per cent in all of these patients. Corrective operations were conversions from a BI to a BII (one patient), and a BII to a BI (three patients). (Table V.) These four patients improved to the point where there were fewer than five stools daily and their weight level before diarrhea was regained. The fifth patient, treated by takedown of a jejunal interposition, did not improve. Thus, the over-all results in this group were satisfactory in 80 per cent. Reftux Alkaline Gastritis. Bile reflux with alkaline gastritis was responsible for reoperation in three patients, one man and two women, all with severe, continuous epigastric distress aggravated by meals and unrelieved by vomiting. Endoscopy revealed marked gastritis with hypertrophic gastric rugae and free reflux of bile into the gastric lumen. The original operations were vagotomy and BI antrectomy in two patients and conversion of a BII subtotal gastrectomy to a BI in one patient in whom bilious vomiting developed. In each case the symptoms appeared within one year after the original operation and reoperative surgery was undertaken within three years of this time. Cholestyramine was of no value in the one patient in whom it was used. Preoperative gastric analyses showed less than 1.5 mEq H+/hr basally. The corrective procedures were Roux-en-Y gastrojejunostomy in two cases and interposition of an isoperistaltic 15 cm jejunal segment between the stomach and duodenum in one case. All patients were markedly improved postoperatively. Miscellaneous. Two patients had symptoms and clinical findings that did not allow them to be placed confidently in any of the preceding PGS categories. They both had multiple gastrointestinal symptoms including pain, nausea, and emesis. Preoperative diagnosis was vague and the rationale of converting a BII to a BI in each of these patients was unclear. Neither was improved by the procedure.
74
One patient was thought to have jejunogastric intussusception because of abdominal pain and radiologic findings. Conversion of a BII to a BI relieved the pain for three years but it returned and another operation was subsequently performed elsewhere. She still has pain. One woman who had vagotomy and BII antrectomy for benign gastric ulcer three years later underwent a BII to BI conversion with interposition of an isoperistaltic 25 cm jejunal segment. Shortly thereafter constant pain developed in the left upper quadrant. X-ray films revealed an anastomotic ulcer, and basal acid output was 9 mEq/hr. The interposition was resected, a residual vagal trunk severed, and gastroduodenostomy performed. Although this cured the ulcer, she continues to have abdominal pain and diarrhea. Serum gastrin levels are within normal limits. Comments
The patient with late postgastrectomy complaints typically presents a perplexing clinical picture with multiple symptoms. The initial task of the physician is to sort out each of these symptoms carefully and to classify the patient whenever possible into one specific PGS category. Since treatment and prognosis for the syndromes may differ, the importance of adequate preoperative evaluation should be evident. In the sections that follow, each of the syndromes is discussed and the appropriate diagnostic approaches are outlined. Stoma1 Obstruction. Stoma1 obstruction may occur early or late. The early variety, associated with postoperative edema, is usually self-limited. Chronic obstruction may result from too narrow an anastomosis, unresolved edema, or recurrent ulceration of the stoma. Some authors [I ] consider stoma1 obstruction to be a rare cause of late postgastrectomy symptoms, but our experience is just the opposite. As an indication for reoperation, it was more common than dumping or bilious vomiting. Diagnosis was based on a history of pain, vomiting, or both after meals, but required endoscopic confirmation of stoma1 narrowing and x-ray evidence of delay in the passage of barium. cJejunal intussusception is an unusual cause for stoma1 obstruction in which diagnosis is more difficult because of the intermittent nature of’ the problem. Stoma1 obstruction is a symptom complex that has a clear anatomic basis and should be correctable in all cases. Our experience indicates t.hat total reconstruction of the anastomosis rather than attempts at stoma1 revision are more likely to be successful.
The American Journal of Surgery
Postgastrectomy
Too rapid emptying of Dumping Syndrome. gastric contents is generally accepted as the underlying cause of dumping, and the operations designed to correct it must slow gastric emptying to be successful. Three operations have been widely used: simple conversion from a BII to a BI anastomosis, conversion with interposition of a 10 cm isoperistaltic segment of jejunum, and interposition of a similar segment in an antiperistaltic fashion. The clinical experience with simple conversion has been variable. Woodward and Hastings [Z ] report satisfactory results in twelve of fifteen patients and consider the small diameter of the gastroduodenal stoma to be the most important factor. In the present series, nine patients were treated by simple conversion, five with improvement (56 .per cent); no patient was cured. This is in keeping with the experience of most authors [3,4] who find significant improvement in about half the patients. Isoperistaltic jejunal interposition produced improvement or cure in eleven of thirtyone patients in the series reported by Fenger, Kallehauge, and Gudmand-Hoyer [5] and nine of those eleven patients originally had a BII anastomosis. Only two of twelve patients with BI gastrectomy were improved. They concluded that the majority of patients who were improved had symptoms from partial afferent loop obstruction as well as dumping and that for pure dumping, isoperistaltic jejunal interposition was not effective. This reinforces our contention that accurate diagnosis is critical in predicting operative results in these patients and that when accurate diagnosis is impossible, a successful outcome is uncommon. Half our patients (two) who underwent this procedure had satisfactory results. Although there is reason to believe [6] that slower gastric emptying should result from isoperistaltic jejunal interposition compared with gastroduodenostomy, the results in our series indicate that the effect, if real, often must he inadequate. On the basis of our over-all success rate of 53 per cent, we would support a wider trial of reversed jejunal interposition, an operation that retards gastric emptying more than these other procedures [7,12]. Our patient treated with .this procedure was markedly improved. Recent observations that the incidence of dumping may be greater after selective vagotomy compared with truncal vagotomy suggest that we may encounter more patients in the near future who require surgical relief for this problem. Of twenty patients who had selective vagotomy at our institution, in the past four years, severe dumping has developed in two and one has
Syndromes
been scheduled for a reversed jejunal segment in the near future. Bilious Vomiting. There has been an unwarranted tendency to explain all bilious vomiting as representing examples of the afferent loop syndrome [8] w h’ic h involves partial obstruction of the afferent loop, accumulation of bile and pancreatic juice proximal to the stoma, and sudden bilious vomiting as pressure in the afferent loop overcomes resistance and bile floods the gastric remnant. Alexander- Williams [S ] has offered evidence that vomiting in some patients is due to topical effects of bile on the gastric mucosa. In these cases, reflux, not obstruction, is the primary abnormality. Also included among those who complain of vomiting are patients with stoma1 obstruction, efferent limb obstruction, and alkaline gastritis and a significant group in whom no specific cause can be identified. Often those in this last category will have had vomiting (without organic obstruction) as a prominent complaint even before the initial operation. They are surgically incurable but sometimes hard to identify. The first step is to search for bona fide signs of obstruction. If, for example, persistent narrowing and slowed passage of barium can be demonstrated at some point, the diagnosis and appropriate therapy are usually obvious. More often signs of obstruction are either absent or equivocal. We believe that such patients are best classified as having bilious vomiting, a more general term than afferent loop syndrome. In some cases, bile reflux is responsible, and in other, partial obstruction, but since it may be impossible to distinguish definitely between these possibilities, surgical treatment should attempt to correct them both. This can be accomplished with Roux-en-Y BII reconstruction or isoperistaltic BI jejunal interposition; conversion of a BII to a BI still allows reflux and is therefore less successful. In the present series, conversion improved five patients, two jejunal interpositions were unsuccessful, and one Roux-en-Y gastrojejunostomy was effective. It must be remembered that operations that keep bile away from the gastrointestinal anastomosis are potentially ulcerogenic and should be combined with vagotomy if gastric analysis reveals the presence of significant acid secretion. In fact, some [9] have even suggested that in the absence of acid, vagotomy should be performed since recovery of acid secretory capacity may occur after bile reflux is eliminated. One patient was referred to us with a marginal ulcer that developed after
Reber and Way
conversion from a BIT gastrectomy to a BI with jejunal interposition. If preoperative studies of acid secretion had been performed, the predisposition towards marginal ulcer could have been predicted and interposition either avoided or accompanied by vagotomy. Malabsorption and Diarrhea. Whenever weight loss or diarrhea is prominent after gastric operations, one of the first steps should be an investigation of the efficiency of intestinal absorption. Measurement of daily fecal fat excretion should be performed while the patient is given a diet containing 100 gm of fat per day. BII gastrectomy alters the anatomy, so that mixing between ingested food and digestive enzymes is reduced and the average asymptomatic patient can be shown postoperat ively to increase fat excretion from the normal of 5 per cent to about 10 per cent. This amount of malabsorption by itself is insignificant, but it probably reduces the patient’s threshold for serious nutritional difficulty if defects develop in the micellar or mucosal uptake phases of absorption. Postgastrectomy patients with malnutrition due to malabsorption will almost always excrete over 20 per cent of dietary fat. Once malabsorption has been established, it is necessary to look for bacterial overgrowth from stasis in the afferent loop, pancreatic insufficiency, and sometimes intestinal mucosal dysfunction. The former patients may be diagnosed by a response to tetracycline therapy, a positive CL” cholate breath test, or direct sampling of afferent loop contents for bacteriologic study. Those with pancreatic insufficiency may respond to pancreatic enzyme replacement with increased efficiency of absorption. When pancreatic insufficiency seems to account for the malabsorption, it will most often represent an exaggeratioil of the mechanism outlined above, with bile and enzymes arriving at the stoma only after the food has already passed into the efferent limb. These patients can be effectively treated surgically. A smaller number have pancreatic insufficiency on the basis of intrinsic pancreatic disease. This group can usually be identified from the rest because the parenchymal destruction frequently is accompanied by endocrine insufficiency and visible calcifications on abdominal x-ray films. In the few patients with neither bacterial overgrowth nor pancreatic insufficiency of a marked degree, additional studies, such as d-xylose absorption tests may indicate nontropical and jejunal biopsy, sprue, lactase deficiency, or some other absorptive abnormality. Overall, the majority of patients with malabsorption will become candidates for surgical
reconstruction since both afferent loop stasis and inefficient mixing can be corrected. In the present series, all five patients with malabsorption were benefited by conversion operations. Other authors [10,12 1 have also reported high rates of success with this or similar operations. The common feature of tile successful procedures is elimination of the afferent loop: postoperatively, food passing through the previous blind loop wipes the bacteria along with it, stasis is no longer present, and food once again enters the intestine at the same point as hile and pancreatic enzymes. In the absence of malabsorption, patients with uncontrollable diarrhea usually have postvagotomy diarrhea. If codeine or dietary manipulation fails and the problem is severe, reconstruction may be considered. Although four of five patients in this series (80 per cent) had good results, the reason for the improvement is difficult to explain. There were three conversions from RI1 to BI and one from BI to BII. The reversal of a segment of midjejunum, which is reported to improve this condition, may deserve wider use, but we have had no experience with it. At, present the pathophysiologic aspects of postvagotomy diarrhea are poorly understood. It is noteworthy that parietal cell vagotomy without pyloroplasty seems free of postoperative diarrhea, suggesting that uncontrolled gastric emptying is an important contributing factor. Reflux Gastritis. With widespread use of gastroscopy, reflux alkaline gastritis has been identified as a distinct postgastrectomy syndrome. The patients present with pain, an endoscopic picture of severe gastritis and edematous rugae, low acid secretion, and perhaps bilious vomiting, In the treatment of this disorder, excellent results have been reported after Roux-en-Y gastr<).jejunostomy or interposition operations (6,.Y,Z.‘$]. Our three patients had uniformly excellent results from either of these procedures. The common feature here is diversion of bile and pancreatic secretions away from the gastric stoma so that reflux c,annot occur. Again, vagotomy should be added if it has not already heen performed.
Conclusions
With some justification, an over-all the results of surgical correction of tomy syndromes has been pessimistic. is gradually becoming apparent how ous these patients are despite the
appraisal of postgastrecHowever, it heterogenesimilarity of
Postgastrectomy
their subjective complaints. As a result, it is possible to more accurately tailor the surgical reconstruction to suit the problem and in some groups to achieve a high degree of success. At present, most patients with stoma1 obstruction, malabsorption, and alkaline gastritis can be helped by further surgery. Roux-en-Y gastrojejunostomy or isoperistaltic jejunal interposition with BI is recommended for bilious vomiting and alkaline gastritis; conversion from BII to BI for malabsorption; and total stoma1 reconstruction for stoma1 obstruction, Reconstructions for dumping or bilious vomiting have been less successful, but different approaches are being tested and some seem quite promising. We believe that the dumping syndrome should be treated by a 10 cm antiperistaltic jejunal interposition with BI. There will probably always be a few patients who do not fit clearly into any of the presently recognized syndromes. Improvement from any reconstructive procedures in this group is usually doomed to failure. Summary
An analysis of the experience at the University of California Hospital, San Francisco, with the diagnosis and treatment of fifty-nine postgastrectomy syndromes shows that stoma1 obstruction (sixteen patients), the most common syndrome, was best treated by total reconstruction rather than stoma1 revision. Disappointment with the results of other procedures for the dumping syndrome (50 per cent improvement) has convinced us of the need to adopt the use of reversed jejunal interposition for surgical treatment of this condition. Because a precise etiologic diagnosis of bilious vomiting is often elusive, the preferred procedure is isoperistaltic jejunal interposition, since it eliminates the afferent loop and prevents bile from entering the stomach. All five patients with malabsorption were improved by conversion from BII to BI. Four of five patients with diarrhea were im-
Valumm 129. Januarv
1275
Syndromes
proved by various procedures (not including a reversed segment of intestine in the midjejunum). Three patients with reflux alkaline gastritis were improved by Roux-en-Y gastrojejunostomy or isoperistaltic jejunal interposition. Either is effective. Thus, in our experience if an unquestionable diagnosis of stoma1 obstruction, malabsorption, or reflux alkaline gastritis can be established, there is ample justification for an optimistic outlook regarding surgical therapy.
References 1. Hirschowitz BI: Classification of the post-gastrectomy syndromes based on observation of 580 patients. Ala J Med SC; a: 50, 1971. 2. Woodward ER, Hastings N: Surgical treatment of the postgastrectomy dumping syndrome. Surg Gynecol Obstet 111: 429, 1960. 3. Andreasson M: Surgical treatment of severe dumping syndrome: conversion of Billroth il into Billroth I. Acta Chir Stand SuppI p 222. 1961. 4. Borg I, Borgstrom SG, Haeger K: The value of the 811-81conversion operation in the treatment of the postgastrectomy syndrome. Acta Chir Stand 134: 655, 1968. 5. Fenger HJ, Kallehauge HE, Gudmand-Hoyer E: Pathophysiological studies of the effect of isoperistaltic jejunal interposition on the treatment of the incapacitating dumping syndrome. Stand J Gastroenterol 7: 283, 1972. 6. Alexander-Williams J: Gastric reconstructive surgery. Ann R Co/l Surg En@ 52: 1, 1973. 7. Herrington JL, Sawyers JL: A new operation for the dumping syndrome and post-vagotomy diarrhea. Ann Surg 175: 790, 1972. 8. Woodward ER: The pathophysiology df the afferent loop syndrome. Surg C/in N&th Am 46: 411, 1966. 9. Van Heerden JA, Priestly JT, Farrow GM, Philips SF: Postoperative alkaline reflux gastritis: surgical implications. Am J Surg 118: 427, 1969. 10. Alexander-Williams J: The place of surgery in the treatment of postgastrectomy syndromes. Proc R Sot Med 57: 585, 1964. 11. Fineberg C, Templeton JY, Wirts CW, Goldstein F: The correction of post gastrectomy malabsorption by jejunal interposition. Surg Clin North Am 53: 581, 1973. 12. Jordan GL: Surgical management of postgastrectomy problems. ArchSurg 102: 251, 1971. 13. Joseph WL, Rivera RA, O’Kieffe DA, Geelhoed GW, McCune WS: Management of postoperative alkaline reflux gastritis. Ann Surg 177: 655, 1973.
77