Duodenal Switch

BPD with a Duodenal Switch

The entire contents of this page are used with permission, and are copyright ©1998 by Douglas S. Hess, MD, FACS, et al, all rights reserved. Text, graphics, and HTML code are protected by US and International Copyright Laws, and may not be copied, reprinted, published, translated, hosted, or otherwise distributed by any means without explicit permission. Correspondence should be directed to Douglas S. Hess, M.D., 640 South Wintergarden Road, Bowling Green, Ohio 43402, USA. Telephone: 419-352-1452; Fax: 419-352-1244

Biliopancreatic Diversion with a Duodenal Switch

Douglas S. Hess, MD, FACS; Douglas W. Hess, MD

Background
Methods
Results
Conclusions
Introduction
Rationale for Duodenal Switch
Patient Selection And Preoperative Evaluation
Procedure
 
Measurements and Size
 
Time
 
Blood Loss
Follow-Up
 
Results
 
Super Morbid Obesity
 
Complications
  
Gastroduodenal Leaks
  
Distal Roux-En-Y Leaks
  
Deaths
  
Revisions
Laboratory
Diabetics
Anemia
Calcium
Advantages
Acknowledgements
References

Background: This paper is an evaluation of the use of the biliopancreatic bypass combined with the duodenal switch procedure forming a new hybrid procedure which is a combination of restriction and malabsorption.

Methods: Evaluation is of the first 440 patients with this operation who have had no previous bariatric surgery. The mean starting weight is 183 kg., with 41% of our patients considered super morbidly obese (BMI > 50).

Results: There is an average maximum weight loss of 80% excess weight which occurs at 24 month post-operative and continues at a 70% level for eight years. Our major complications are listed and total near 9% of all cases. We have two perioperative deaths, one from pulmonary embolism and one with acute pulmonary obstruction. There are 36 type II diabetics all of whom are no longer taking medication since the surgery. Seventeen revisions were performed to correct excess weight loss and low protein. There are no marginal ulcers , no dumping syndrome, no foreign material used, and it is a pyloric saving procedure which is functionally reversible.

Conclusions: This operation has vastly improved the lives of our seriously obese patients with many co-morbidities. All type II diabetics have essentially been cured of their disease. The procedure is tolerated well and the patients are all quite satisfied. There does not seem to be the late regain of weight in this method.

back to top

INTRODUCTION

In 1988 using a combination of Dr. Scopinaro's biliopancreatic bypass and Dr. DeMeester's duodenal switch procedures, we developed a hybrid which has the advantages of the biliopancreatic bypass without some of the associated problems. This operation is now used by us for all our bariatric patients, both in primary surgical procedures and reoperations.

The difficulty of establishing an operation that has both long and short term success is well known. Bariatric surgery is either restrictive or malabsorptive in nature, each with its own advantages, disadvantages and complications. While trying to find a procedure that would produce better long term results and fewer failures i.e., regain of weight, we began to look at Dr. Scopinaro's biliopancreatic bypass1. First we considered it only for reoperations on failed restrictive procedures. Our first cases were reoperations with the standard Scopinaro biliopancreatic bypass, and a few without a distal gastrectomy. However, the dense adhesions found in the upper gastric area of previously stapled patients caused difficulty in placing the anastomosis and we were troubled with marginal ulcer formation. We considered a method to anastomose the ileum to the duodenum away from the site of the previous surgery. While searching the literature on the subject of duodenogastric reflux, we found Dr. DeMeester's article about the duodenal switch procedure2 and adapted it to our use.

The first patient to have a biliopancreatic bypass with a duodenal switch procedure was a male patient who had had a transverse gastroplasty nine years earlier, in 1979. He was six feet, four inches tall, and at the time of his surgery in 1979 , he weighed 166 kg (365 lbs.). That surgery eventually failed due to staple disruption, and his weight had increased to 206 kg (454 lbs) by the time of his reoperation in March, 1988. He also had developed chronic heart failure, shortness of breath, and was no longer able to perform many daily activities. Presently, he is approximately nine and one half years post-operative (converted to a duodenal switch with a biliopancreatic bypass), weighs 125 kg (275 lbs.), and is free from problems associated with his obesity. Due to the success of our reoperations, we decided to use this procedure for our primary operation for both the morbidly and super morbidly obese patients.

Our first duodenal switch primary procedure was performed on a morbidly obese male patient (BMI 46) in May, 1988. At his most recent weighing on October 29, 1997, his excess weight loss was 88%, BMI was 26, and his percent of ideal weight was 112. The first super morbidly obese patient (BMI 58.7) was a female operated in June, 1988. At her last weighing on June 10, 1997, her excess weight loss was 78%, BMI was 30, and her percent of ideal weight was 137. Both of these patients are more than nine years post-operative and are doing well. This operation is now our procedure of choice for surgical correction of morbid and super morbid obesity.

This paper will be limited to our first 440 patients, primary procedures only. Reoperations are not included.

back to top

RATIONALE FOR DUODENAL SWITCH

Examining the experimental work on dogs of Dr. DeMeester et al. as described in the Figure 1, it is evident that a small segment of proximal duodenum protects against marginal ulceration. The surgical procedures for all groups are basically the same, except that the dogs in groups C and D had a short segment of proximal duodenum left in place before the jejunostomy. The dogs in groups A and B had a high incidence of ulcers, perforation and weight loss. Groups C and D had significantly fewer ulcers and almost no perforations, however, the weight loss did not change. Using this data and DeMeester's report of patient cases (i.e. "duodenal switch") we felt the combination of the biliopancreatic bypass and the duodenal switch to be a logical progression of this procedure.

Figure 1. Results of four diversion procedures with their effect on ulcer incidence, perforation and weight loss.

To achieve the gastric restriction required by Dr. Scopinaro's biliopancreatic bypass, a vertical gastrectomy is performed, which removes a large portion of the fundus and also reduces acid formation, helping to prevent marginal ulceration (Figure 2).

Figure 2. Biliopancreatic diversion with duodenal switch procedure.

back to top

PATIENT SELECTION AND PREOPERATIVE EVALUATION

Patients selected for surgery in this cohort were between the ages of 16 and 62, the average age being 39 years. All patients were 45.5 kg (100 lbs.) or more above their ideal weight using the 1983 Metropolitan height and weight tables. The presence of co-morbidities is an additional indication for surgery. Patients with combinations of co-morbidities with a lower excess weight are occasionally candidates for surgery. However, most patients are more than twice their ideal weight at the time of surgery (Table 1).

Table 1. Preoperative data: morbid obese patients

 

Average values (n = 440)

Sex (M/F)

95/345

Age (years)

40

Weight (kg)

138.3 (304.3 lb.)

Height (cm)

166.1 (65.4 in.)

Starting BMI

50

Percent ideal weight

222

Excess weight (kg)

76 (167.2 lb.)

All patients view a video explaining the surgical procedure and possible complications. They are later seen by the physician for an extended interview and second explanation of the procedure, including expectations. They have the usual preoperative complete blood count, chemistry profiles, cardiac evaluation, and x-rays. They are then seen by an internist for a medical evaluation. The patients are encouraged to attend one of the monthly support meetings before surgery, which is also attended by a physician. The meetings allow the perspective patients to acquire additional information and talk to patients who have had this surgery.

back to top

PROCEDURE

Under general and epidural anesthesia a midline incision is used in the upper abdomen from the xyphoid to the umbilicus. A Gomez retractor is put in place, and the abdomen is explored. We start with an appendectomy, and complete measurement of the small bowel along the antimesenteric border with a mild stretch between two Babcock clamps, from the cecum to the ligament of Treitz. A silk suture is placed in the margin of the bowel at the estimated proximal common channel and the proximal end of the alimentary limb. These measurements may be changed after complete measurement of the bowel. The length of the alimentary limb is calculated by multiplying the bowel length by 40% to determine the distance from the cecum to the ligament of Treitz. Using 25 cm. increments, the final length of the alimentary canal is the nearest to the 40% area. Generally, the alimentary canal will be 250 cm., 275 cm., or 300 cm. Occasionally 225 cm. or 325 cm. will be used in patients with an unusually short or long small bowel. Since the epidural anesthesia may cause spasm of the small bowel, it is not used until after the measurements are completed.

The position of the table is changed to reverse trendelenberg as our attention is focused on the upper abdomen. The left triangler and Falciform ligaments of the liver are incised to allow retraction of the liver away from the stomach and is held with gauze and a liver retractor. We open into the lesser sack by transecting a couple of short gastrics at the mid greater curvature of the stomach. With a tripolar cautery (i.e., a bipolar cautery with a knife for cutting cauterized tissue) we transect all the short gastrics from the pylorus to the esophagus, freeing up the entire greater curvature portion of the stomach. Blood vessels are transected along the greater curvature side of the duodenum for about five centimeters distal to the pylorus, and a penrose drain is passed around the duodenum at this point. The Salem tube is removed and a 40 F dilator, to be used as a sizer, is passed into the stomach, along the lesser curvature, and through the pylorus into the first part of the duodenum. The greater curvature side of the stomach is held upward and to the left by an assistant, using Babcock clamps. We size the new stomach by placing the ILA-100 stapler along the dilator, starting from the distal stomach a few centimeters proximal to the pylorus, directed upward, separated from the dilator by a space of one to two finger breadths. This stapling is repeated three or four times as needed to remove the entire greater curvature side of the stomach. Bleeding points along the staple row are oversewn with figure-of-eight silk sutures. Then the staple row is inverted by a continuous suture using (U.S. Surgical O-Biosin) a serosal to serosal suture. The dilator is removed and the duodenum is transected as far distal to the pylorus as possible (generally 4 to 5 centimeters) with an ILA 52 and the Salem tube is reinserted.

We measure the volume of the stomach in all cases at this time by filling the stomach with saline and methylene blue dye. With one hand we press against the esophagus to stop reflux of the saline, we do not dissect around the esophagus. After the stomach is distended we withdraw the saline and this amount is considered to be the volume of the stomach, as well as our check for possible leaks.

A cholecystectomy is performed in all patients who still have a gallbladder in the routine manner. The small bowel which was previously marked for the alimentary limb is transected with an ILA 52, and the mesentery is divided with the tripolar cautery until the distal ileum can be mobilized up to the pyloric area. The distal ileum is taken retrocolic to the right of center up to the duodenum which is still attached to the pylorus and anastomosed to this duodenum with a Valtrac anastomosing ring end to end using 1.5 mm gap and a 25 mm diameter Valtrac. The mesentery is sutured to the post peritoneal wall to prevent any internal hernias. The proximal ileum is taken distally to the previously marked area on the distal ileum and an end to side anastomosis is performed with a Valtrac anastomosing ring (1.5 mm gap and 25 mm diameter) and the mesentery is closed with a running suture. The abdomen is closed in the usual way and we generally place a drain in the gallbladder bed area.

There are several concurrent procedures performed with this surgery as illustrated in Table 2.

Table 2. Concurrent procedures, by type and number performed

Type

n

Cholecystectomy

330

Appendectomy

278

hernia repair

25

tubal ligation

11

liver biopsy

7

incidental splenectomy

4

hiatus hernia repair

2

oopherectomy

2

back to top

MEASUREMENTS and SIZE

After resection of the greater curvature we would prefer to have the volume of the stomach to be somewhere around 150 cc (generally, no less than 100 cc or more than 175cc). We size the stomach with a #40 dilator along the lesser curvature and we stay 1 1/2 to 2 finger breaths away from the dilator towards the greater curvature side when we apply the ILA-100 stapler. In general, this will give us a stomach that will fall in that range. The serosa to serosa suture reduces the size of the stomach which is a factor in early weight loss as time passes this stomach volume will enlarge.

We believe that the total length of the alimentary limb (from the cecum to the stomach) should be approximately 40% of the total length and that the common channel (the distal portion of the alimentary limb just beyond the anastomosis of the biliary limb ) should be somewhere around 10% of the total small bowel length. We make the common channel 50 cm, 75 cm or 100 cm, whichever seems the most proper. The length of our alimentary limb is always calculated in 25cm increments: 250 cm, 275cm, 300 cm, etc. Table 3 illustrates the common channel, the alimentary limb and approximately the number of cases we used in that group. Also, it indicates the average length of the bowel, patient's weight, and stomach volume, and this illustrates that in general, the measurements are close to 40% alimentary canal and 10% common channel. These are only guidelines and we evaluate every patient individually at the time of surgery as to their age, their weight, and to the number of co-morbidities in making our decision on the above sizes. The proper volume of the stomach, and the lengths of the roux-en-y limbs and bilio limbs have been debated. Table 3 is of the first 382 cases of primary BPD with a duodenal switch and it gives an impression of the variety of measurements in lengths with similar bowel lengths in percentages. We believe that the percentage should be considered when deciding how long the limbs should be, then as long as consistency is maintained, the method of measurement is insignificant.

Table 3. Measurements of the first 382 primary BPD with duodenal switch patients

Common channel/Alimentary limb

n

Avg. Measurements (kg/cm/cc)

 

AL/CCH %

50cm/225cm

6

weight
bowel length
stomach volume

139 kg (116-175)
545 cm (516-589)
133 cc (110-160)

41/9

50cm/250cm

52

weight
bowel length
stomach volume

129 (90-185)
620 (519-760)
172 (75-305)

40/8

50cm/275cm

32

weight
bowel length
stomach volume

132 (102-208)
697 (640-814)
179 (100-250)

39/7

50cm/300cm

19

weight
bowel length
stomach volume

139 (98-194)
800 (706-900)
193 (125-255)

37/6

75cm/250cm

26

weight
bowel length
stomach volume

130 (90-164)
612 (473-713)
164 (120-235)

41/12

75cm/275cm

122

weight
bowel length
stomach volume

140 (92-266)
708 (577-858)
164 (100-250)

39/10

75cm/300cm

83

weight
bowel length
stomach volume

148 (92-232)
795 (497-985)
164 (100-240)

39/9

100cm/300cm

35

weight
bowel length
stomach volume

129 (89-225)
804 (635-997)
193 (125-250)

37/12

75and100cm/325cm

3

weight
bowel length
stomach volume

165 (132-186)
877 (856-888)
192 (125-250)

36/11

100cm/350cm

4

weight
bowel length
stomach volume

133 (90-168)
860 (826-900)
189 (135-245)

40/12

CCH, common channel; AL, alimentary limit

Table 4 illustrates that the difference in bowel length from one person to the other may be more than 100%. If the same lengths are used in all patients then some will be too short and some too long. The exact proper size is elusive and is yet to be discovered.

Table 4. Total small bowel length

 

n

Max length

Min length

Avg length

Female

342

997 cm

473 cm

707 cm

Male

98

1065 cm

497 cm

776 cm

Both

440

1065 cm

473 cm

722 cm

Figure 3 is a graph of weight loss according to the sizes used in Table 3.

Figure 3. Percent excess weight loss with time.

The information presented at the bariatric surgery meeting in June 19973 (and by personal contact ) in Chicago by Dr. Scopinaro indicated that by increasing the length of the alimentary limb the absorption of protein increases substantially, whereas increasing the size of the common channel alone does not necessarily increase the amount of protein absorption, but it does reduce fat malabsorption. It is his feeling that most of the digestive juices and bile are reabsorbed before they reach the common channel and that an increase in the alimentary canal would be a method of increasing protein absorption without reducing the selective malabsorption of fat and weight loss capabilities. However, as you increase the length of the alimentary limb you also increase the absorption of carbohydrates. We feel that this is probably correct and in general it supports our method of using 40% for the guideline of the length of the alimentary canal. It appears that the alimentary canal can be made longer, the common channel made relatively short, and a selective malabsorption of fat will occur while maintaining adequate absorption of protein.

back to top

TIME

Time in any surgical procedure is of importance. In the first 114 cases that we performed our average operative time was 199 minutes. After we had performed 213 cases our average time was 187 minutes. In the last 115 operations our operative time was 158 minutes, which is our current average.

These improvements on operative time are partially due to newer techniques. We now measure the small bowel at the beginning of the operation and remove the appendix. We use a tripolar cautery to take down all the short gastrics along the greater curvature of the stomach. We rarely tie any short gastrics. This has reduced our blood loss and operative time. We use this same method for dividing the mesentary in the small bowel. There are two anastomosis, both of which are made with the Valtrac ring anastomoser. This has significantly decreased our operative time and reduced possible complications. All of these factors, technical and mechanical improvements, have made this operation more efficient.

back to top

BLOOD LOSS

Our blood loss has not been excessive and in most cases we have the patient donate an autologous unit of blood two weeks prior to surgery which we give back at the time of surgery. Generally no further blood is needed. We have given 342 units of autologous blood in our first 440 patients. There are some patients who were either not able to give autologous blood or required extra blood and we have used a total 26 units of banked blood of our 440 primary surgeries, excluding two patients who had interabdominal bleeding post-operatively and required surgery to control the bleeding. In most cases our blood loss is about 300 cc.

back to top

FOLLOW-UP

In bariatric surgery, patient follow-up and weight loss results over several years are very important. Without this follow-up we would have little idea about the success or failure of the surgery we are performing. Most of our patients are on six month follow-up visits after the second year of surgery, therefore, we have a six month period in which our follow-up is current. In this series of 440 primary biliopancreatic diversions with the duodenal switch there are six patients that we have not been able to contact for follow-up. This gives us in this cohort a total percent of follow of 98.60% We feel this gives credence to the value of our statistical results.

back to top

RESULTS

The results are shown in Figure 4 by weight loss (kg.), percent excess weight loss, and percent total body weight loss. This is a graph of the average of the first 440 primary duodenal switch patients. The maximum weight loss occurs around 24 months and at about 80% excess weight loss and continues at a satisfactory level out to eight years.

Figure 4. Weight loss in 440 primary bariatric surgery patients.

This cohort is over nine years old and Table 5 illustrates the number of cases which were performed each year from which this graph is taken. This graph shows the average weight loss but we also evaluate our success according to individual patients.

Table 5. Primary bariatric patients with BPD and a duodenal switch

Year

n

1988

3

1989

28

1990

40

1991

23

1992

45

1993

43

1994

71

1995

76

1996

72

1997

79*

* Only 39 cases from 1997 were included in this paper.

We calculate each patient individually. Table 6 consists of the grade determinants that are used for individual results and includes 361 patients. We did not include any of the patients whose last recorded weight was taken less than 9 months after surgery (these results are based on the last weight recorded, not an average).

Table 6. Composite grade and results for 361 primary cases.

Grade

n

Total (%)

Excellent

288

79.8

Good

47

13

Fair

25

6.9

Poor

1

0.3

Failure

0

0

Total Satisfactory

 

99.7

Total Unsatisfactory

 

0.3

Grade determinants (all minimum values): Excellent, 80% EWL;
Good, 60% EWL; Fair, 40% EWL; Poor, 20% EWL; Failure <2% EWL.
Satisfactory = excellent, good, fair; unsatisfactory = poor, failure.

Table 6 illustrates the number of patients in the different categories and their results with follow-up to nine years. There are 335 patients in the "good" and "excellent" categories for a total of 93%. The other group falls into the "fair" category which is still satisfactory. Most of these are considered successful, as they have had a stable weight for a long time, but did not lose excess weight of more than 60%. In the "fair" group there are 25 patients whose excess weight loss is from 40.2 to 59.8 percent with an average of 50.2% excess weight loss. When comparing those to the determinates in Table 3, our total satisfactory results is 99% with 1% unsatisfactory. Table 7 illustrates the total patients counted in different years follow-up by number of patients, number of satisfactory cases, and number in good and excellent categories at that year. Of the long term results, there is a only small amount of regain and we have never had a patient regain weight up to their original weight. Our one patient in the "poor" category has a weight loss of 35% of excess weight.

Table 7. Follow-up results of the BPD with duodenal switch patients, grouped according to time since surgery

Years since surgery

n

Satisfactory

Good-Excellent

8

11

11

9

7

29

29

25

6

51

51

45

5

92

92

80

4

132

132

117

3

187

186

167

2

264

263

240

1

345

344

320

Of the 440 patient cohort, 301 patients have a minimum of one year follow-up and some have a follow-up time to as much as nine years. The average maximum weight loss is at 23 months with an average maximum loss of 63.2 kg. and an average maximum excess weight loss of 85%. Table 1 gives the demographics on the group, total number of 440 patients with normal distribution of male/female and age. Our average weight is rising, and is now 137 kilograms with a body mass index of 50 and 222% excess weight.

As illustrated in Table 5, the success of the BPD with the duodenal switch procedure has contributed to the increasing number of cases we perform. This method of surgery has been the most successful for patient weight loss that we have used so far to this date.

back to top

SUPER MORBID OBESITY

By defining the super morbid obese patient to be 225% of ideal weight, or having a BMI of 50 or more, then 41% of our patients fall into that category4. Our average patient is now 138 kg (304 lbs) and the average starting weight has recently been rising. Figure 5 shows the weight loss curve of the super morbid obese patient (BMI>50), shows weight loss comparable to the morbid obese group (BMI<50) and it to continues to have weight loss in the 70% excess and maintains good weight loss throughout the eight year period.

Figure 5. Weight loss shown by % EWL in morbid obese and super morbid obese patients.

The statistical data of our 181 super morbidly obese patients is listed in Table 8, and the composite results of the super morbidly obese patients with at least nine months post-surgery follow-up are seen in Table 8. No unsatisfactory results were seen and 97.3% of the patients fell into the good and excellent categories.

Table 8. Preoperative data: super morbid obese patients

 

Average values (n=181)

Sex (M/F)

40/141

Age (years)

39

Weight (kg)

159.9 (351.8 lb.)

Height (cm)

165.8 (65.3 in.)

Starting BMI

58

Percent ideal wt.

257

Excess weight (kg)

97.4 (214.2 lb.)

The super morbid obese patient has a much more difficult time successfully losing weight. We feel that a restrictive procedure alone will probably not work for most super morbidly obese patients. The biliopancreatic bypass with a duodenal switch works quite satisfactorily and could be used for all super morbid obese patients in our hands. These are the patients who have the most serious co-morbidities, high degree of sleep apnea, cardiac problems, diabetics and so forth. These are the people whose lives are threatened daily by their severe obesity and if possible they need to have one successful operation that will bring them down to a near-normal weight.

back to top

COMPLICATIONS

There are several complications that can occur in, but are not limited to, bariatric surgery. Table 9 list these complications in our group. In addition, there is the frequent problem of atalectasis (20%) in the obese post-operative patient.

Table 9. Complications of BPD with duodenal switch

Complication

n

Total (%)

Medical complications (perioperative)

 

 

  Deep vein thromboflibitis

3

0.75

  Non-fatal pulmonary embous

2

0.5

  Pneumonia

2

0.5

  ARDS

1

0.25

Surgical Complications (perioperative)

 

 

  Splenectomy (incidental)

3

0.75

  Gastric leak and fistula*

8

1.7

  Duodenal leak

6

1.5

  Distal Roux-en-Y leak

1

0.25

  Post-op bleeding (surgery req.)

2

0.5

  Abscess (not related to leaks)

1

0.25

Late Complications

 

 

  Duodenal stomal obstruction

3

0.75

  Small bowel obstruction

8

2

*For explanation, see text

Gastric leaks are the most significant complication of bariatric surgery. This is the perforation of some portion of the stomach or staple line which occurs after the bariatric surgical procedure. The incidence of gastric leaks has been reported from one half percent to five percent in different series of bariatric surgery. In our series of 440 cases, in the first 252 operations we had a total of eight leaks, or 3.1 percent. These occurred on both two row or four row staple lines with simple oversuturing of the staples.

After viewing a video by Dr. Aniceto Baltasar5 at an annual meeting of the American Society for Bariatric Surgery, we modified our procedure for the last 188 cases. We now use the ILA-100 in all cases which forms two rows of staples. We control bleeding edges through the staple line with a "figure-8" suture, and then a complete serosa to serosa closure over the staple line inverts the mucosa from the esophagus down to the pylorus with O-Biosin (a synthetic glycomer by United States Surgical). Of these 188 patients we have had one leak. This leak was in a patient who was very heavy and almost always required C-pap which we think may have increased the intragastric pressure, and did develop a leak on the upper portion of the stomach two day after discharge home on her tenth-post operative day.

It is our feeling that the serosal to serosal inversion of the mucosal portion of the staple line is very important to prevent leaks in these cases. One leak out of the 188 cases brings our leak percentage down to 0.5 percent which is a marked improvement from our 3.1 percent leaks before complete inversion of the staple row. In our hands leaks can be reduced to a minimum even in complex surgery as described above by carefully inverting the staple row with a serosal to serosal stitch.

back to top

GASTRODUODENAL LEAKS

The anastomosis of the ileum to the duodenum near the pylorus is high up on the right side in a difficult area to work. This anastomosis is particular to the duodenal switch procedure. In the first 119 cases we did a hand-sewn anastomosis with an end to side anastomosis double layer with Vicral and silk. Two patients developed leaks from this duodenal-ileo anastomosis (1.6%). They were both small and closed simply by keeping the patient NPO and on IV's for about ten days. For the past four years we have been using Davis and Geck's Valtrac anastomosis ring (a biofragmentable product) for this anastomosis. We have performed over 600 Valtrac ring anastomoses in this manner. Of these there were six anastomatic leaks (1%). Three were treated conservatively with IV fluids and restricting intake by mouth and they closed spontaneously. Three required a surgical procedure to close. One had a leak and an erosion into a small artery producing massive GI bleeding requiring emergency surgery. The other two had what appeared to be a leak larger than expected and we felt this would not close by itself.

In contrast to gastric leaks, if there are proper drains in the area of the pylorus or if the leak is small and you have no other problems it is probably better not to operate these leaks at the duodenum since they generally close with conservative therapy. In our 440 cases described above we have never had any type of leak on the duodenal stump, this does not seem to be a problem.

back to top

DISTAL ROUX-EN-Y LEAKS

Leaks in the distal portion of the roux-en-y are significantly different, and particularly when you have a long limb roux-en-y. We have had one out of all our roux-en-y cases (over 550 or 0.1%). Anastomotic leaks in this area are very critical. There are two major problems with this particular area. First, they are difficult to diagnose. Careful attention must be paid if there are elevated white counts, temperature and a tender abdomen. Secondly, they are quite dangerous and can lead to generalized peritonitis and sepsis. The one case we had had negative x-rays but continued to be septic and we explored the patient on the basis of his sepsis and found a breakdown of the stapled anastomosis at the distal roux-en-y. We exteriorized this distal breakdown and treated the patient with IV fluids and TPN. The controlled fistula was successfully closed later.

back to top

DEATHS

We had two perioperative deaths in our series of 440 patients, both had autopsies performed. One died on the fourteenth post-operative day with a massive pulmonary embolus thought to be 24 hours old. The second patient died from a respiratory arrest with bilateral bronchial obstruction. Both were quite large with sleep apnea and respiratory insufency (Table 10). We have had three late deaths which are in the table below indicating the time post surgery and cause of death (Table 10).

Table 10. Perioperative and late deaths

Early deaths (2)*

 

 

Patient

BMI

Cause of death

(1)

74

pulmonary bronchial obstruction

(2)

58

massive pulmonary embolous

Late deaths (4)

 

 

 

Patient

wt. (kg.)

Mo. Post-op

Cause of Death

(1)

266

7

myocardial infarction

(2)

201

15

CVA

(3)

180

26

septic shock (infected panniculus)

(4)

145

10

fatty liver, liver failure, renal failure, ARDS, multiple organ failure

*520 cases for a death rate of 0.4%

back to top

REVISIONS

We have had a total of 17 revisions. All of these were performed along with other surgeries such as incisional hernias, panniculectomies, and hysterectomies (Table 11).

Table 11. Surgeries performed in the postoperative period

n

Operation performed

117

Panniculectomies

108

Insicional hernias

9

Inguinal hernias

16

Distal roux-en-y revisions *

11

Abdominal hysterectomies

7

Bilateral thigh reduction

5

Upper arm reduction

* Revisions are also listed. We have performed 117 panniculectomies on the first 440 primary switch patients in the post operative period. All of these patients had their panniculectomy performed together with one of the hernias listed. In addition to the hernias, all the other procedures listed were performed during these panniculectomies.

The revisions are for low protein, excessive weight loss, excessive diarrhea or in some cases less weight lost than the patient expected. Table 12 shows our results.

Table 12. Revisions

n

Reason

Procedure

8

Low protein and excess weight loss

Lengthen bowel

2

Excess diarrhea

Lengthen bowel

7

Poor weight loss

Shorten common channel

Eight patients were revised primarily because of low protein and excessive weight lost, two patients had excessive diarrhea, low protein and excessive weight loss. These were treated by lengthening the bowel. Seven patients had good weight loss but they felt they wanted to lose more and they were still 50 pounds or so above ideal weight and the common channel was shortened. Lengthening the bowel works quite well and stops the excessive weight loss and low protein. However, shortening the common channel a year or so after their primary surgery may not give satisfactory extra weight loss. The additional weight loss is generally not much more than around 20 pounds. We have had one reversal in our primary switch patients and this person was several states away and became concerned about not being able to have follow-up surgeons available to her easily in her area.

Three months post her original surgery the small bowel portion was reversed. She has been doing well and now a year or so later is wishing she had not had the small bowel portion changed.

back to top

LABORATORY

The average laboratory values are taken from the first 100 consecutive patients of this cohort covering a period of five years. Intermittent studies of CBC, chemistry profiles and vitamin studies are included. Figure 6 shows the averages of the lipids at the preoperative period compared to the levels three months, six months and one year after surgery. This graph indicates an average level of lipids in the normal range at the start of surgery. There is a lowering of the cholesterol and triglycerides, and at the same time the high density cholesterol (HDL) remains the same, improving the average ratio of low density cholesterol to high density (LDL/HDC) from 3 to 1.5.

Figure 6. Lipid levels in the first 100 consecutive BPD with duodenal switch patients.

Patient #444 is an unusual case in that her main pathology was her high lipid levels. She had been hospitalized several times with pancreatitis due to her elevated lipids. On the last admission prior to surgery her cholesterol was 880 and her triglycerides were 5000. This patient was 97 pounds above ideal weight and at the time of surgery her cholesterol was 440 and her triglycerides were 2000. The high lipid abnormality was the an important co-morbidity as an indication for surgery. This patient had a cholecystectomy several years before and since that surgery she had been troubled with post-prandial diarrhea, which was a concern to us. For this reason we increased the length of her common channel to one hundred centimeters and the total length of the alimentary channel to three hundred centimeters. This increase of bowel lengths has reduced her total weight loss (75% of excess weight loss in 12 months and has stabilized for the last few years at 50% of excess weight loss), but she has no difficulty with diarrhea and her post-prandial diarrhea has been corrected. Figure 7 shows an eight year follow-up of this patient's lipids. All lipids have fallen to normal levels and remained normal to this date. The high density cholesterol level remains constant, which has markedly improved her LDL to HDL ratio.

Figure 7. Lipid levels for patient #444.

One interesting finding is that in taking our patients' history prior to surgery, we find that of the patients that have had a cholecystectomy in the past, between ten to fifteen percent of these patients are troubled with post-prandial diarrhea since having their gallbladder removed. Our first encounter with this in our obese patients was the one who had a very high cholesterol, which caused us some concern for fear that the post-pranial diarrhea may become more severe after surgery. In this cohort of 440 patients we have had 21 patients who complained of post-pranial diarrhea since their gallbladder was removed. Some of them were very severe. After the BPD with the duodenal switch, 18 patients (85%) said the post-prandial diarrhea was much improved and some claimed to be normal since surgery. Three felt that it was unchanged. Because of our results with our surgery on this particular patient problem we no longer feel that post-cholecystectomy, post-prandial diarrhea is a relative contraindication to this surgery. In fact, obese patients who have the problem of post-cholecystectomy, diarrhea would probably be markedly improved by a biliopancreatic bypass with a duodenal switch procedure. Post-cholecystectomy, post-pranial diarrhea could be considered as an additional indication for this surgery.

back to top

DIABETICS

Bariatric surgery has been known to improve or cure diabetes mellitus6. With the biliopancreatic bypass and the duodenal switch we have operated on 36 diabetics, all Type II, of which 18 of them are non-insulin dependent and 18 were insulin dependent. One patient was taking as high as 500 units of insulin a day, but generally they were taking insulin in the range of 40 to 50 units per day. The non-insulin dependent patients would leave the hospital after surgery taking no medication and have continued taking no medication since their surgery. The insulin dependent diabetics would occasionally take a small amount of insulin or a hypoglycemic agent for a short time, but never more than two months following surgery. All of the above patients, after a few months and up to seven years following surgery, are taking no medication of any type for their diabetes. All of them have normal blood sugars as indicated in the graph showing the pre-operative average blood sugar and the post-operative average blood sugar on all these patients up to 5 years (Figure 8). The present glycosylated hemoglobin average for this group is 5.0% (normal reference range is 4.2%-5.9%). We can say without hesitation for the obese Type II diabetic, this surgery will cure their diabetes.

Figure 8. Glucose levels of diabetics.

back to top

ANEMIA

It is well known that bypassing part of the lower stomach or part of the duodenum and proximal jejunum, raises the probability of developing anemia secondary to poor absorption of iron. Dr. Brolin at the 1997 meeting in Chicago7 reported a study of gastric bypasses with 48% anemia. An anemic was defined as any patient whose hemoglobin or hematacrit came back lower than the laboratory's normal range. Using that criteria we would expect to have similar results in our laboratory studies. However, our averages of hemoglobin of our first 100 patients at random intervals of three months, six months and one year up to five years showed values of 12 to 13 grams. There were 259 counts of which 20 (7.7%) were below 10.5 grams and of these there were 12 (4%) below 10.0 grams with the lowest value at 6.4 grams. We consider anemia requiring treatment to be a hemoglobin count of 10.5 or lower. When we evaluate our entire series of 440 patients we have 40 patients (9%) requiring iron in some form or another. Of these patients who had low hemoglobin, seven had excessive uterine bleeding and required abdominal hysterectomies, and ten patients had relatively severe anemia requiring intramuscular iron to correct. All anemias were correctable with the proper iron or surgical therapy.

back to top

CALCIUM

Serum calcium is very important in patients with a malabsorption procedure. The graph of the lab work of our first one hundred consecutive patients (Figure 9) shows that the calcium level stays relatively flat and slightly below the normal level (659 counts in 5 years, range 7 to 10.8). The alkaline phosphatase is elevated indicating increased calcium activity, possibly due to low calcium and vitamin D intake. The vitamin D-25 level is within the normal range, although it is on the low side. The vitamin D-1,25 level increases, as the 1,25 form of vitamin D is the active form in the mobilization of calcium from the bone to the serum. This graph is an indicator of the importance of adequate vitamin D and calcium supplementation.

Figure 9. Calcium metabolism in the first 100 consecutive BPD with duodenal switch patients.

Dr. Mark Jaroch showed in his poster presentation at the American College of Surgeons meeting at Chicago in 1997 that 31% of pre-operative patients had levels of 25 hydroxy vitamin D 30 or above (range 16 to 55) compared to 59% in the post-operative period at the end of one year8. These results indicate that if the patients take vitamin D and calcium they can maintain the proper levels and in some cases increase their calcium and vitamin D to levels higher than those before surgery. Alkaline phosphatase is a good indicator of adequate intake of calcium and vitamin D.

back to top

ADVANTAGES

There are several advantages to the biliopancreatic bypass with a duodenal switch. There is no isolated stomach, no foreign body or band required. There is preservation of the pylorus, no dumping syndrome, no marginal ulcers, and good weight loss.

This operation is both a restrictive and a malabsorption procedure. However, neither of these procedures are performed to an extreme degree. The restriction is related only to reducing the size of the stomach. There is no constricting band or narrowed stoma. We use a vertical gastrectomy which preserves the pylorus, a portion of the antrum, some of the mid and upper stomach, and removes most of the acid producing fundus. If in the future any revision needs to be performed on these patients it would be unusual to have to re-operate on the stomach. Surgery in this area becomes difficult due to adhesions between the stomach, liver, and the upper abdominal area on the second surgeries.

The malabsorption portion of this operation consists of an alimentary canal of 250 to 350 cm, with a common channel portion measuring 50 to 100 cm. of the distal ileum, which practically always gives adequate absorption and nutrition. If there is some difficulty with malabsorption, the length of the alimentary canal and common channel can be extended without much difficulty and without disturbing the stomach or the duodenal anastomosis. Liver failure, renal failure, severe electrolyte imbalances etc. do not seem to be a problem with this operation, if the patients have adequate follow-up and proper supplementation.

Since the pylorus is still intact a functional reversal of this operation can be performed quite satisfactorily. The volume of the stomach, 100 to 175 cc, will enlarge with time, and is always adequate in size. Shortening of the roux-en-y or anatomical reversal would work without the formation of an ulcer or the need of a vagotomy.

It is known that the gastric bypass with both a short or long limb roux-en-y may be an ulcergenic operation. By the addition of the duodenal switch procedure the possibility of a marginal ulcer is remote9. We have never had a marginal ulcer since using the duodenal switch procedure in all of our cases, which including our redo surgeries, number more than 600 procedures. Since we do not remove the pylorus and do not have marginal ulcers there is little need for a vagotomy, and in turn, no dumping syndrome, We have never had a dumping syndrome in any of our cases.

In our 20 years of experience, the biliopancreatic bypass with a duodenal switch has shown to be the most effective weight loss procedure, for both the morbidly obese and the super morbidly obese patient. For the super morbidly obese patient, restrictive procedures alone will probably not be successful. The biliopancreatic bypass with a duodenal switch, however, is a procedure that has shown to be a successful method of treatment for the super obese patient.

back to top

ACKNOWLEDGEMENTS

The authors would like to acknowledge the help of Julia A. Smart with the preparation of the manuscript.

back to top

REFERENCES

1. Scopinaro N, Gianetta E, Civalleri D, et al. Biliopancreatic Bypass for Obesity: initial experience in man. Br J Surg 1979; 66: 618-20.

2. DeMeester TR, Fuchs KH, Ball CS, et al. Experimental and Clinical Results with Proximal End-to-End Duodenjejunostomy for Pathologic Duodenogastric Reflux. Ann Surg 1987; 206: 414-24.

3. Scopinaro N, 14th annual meeting for the American Society for Bariatric Surgery. Chicago, Illinois, June 4-7, 1997.

4. Mason EE, Amaral J, Cowen GSM, et al. Standards for Reporting Results. Obes Surg 1994; 4: 56-65.

5. Baltasar A, delRio J, Escriv C, et al. Preliminary Results of the Duodenal Switch. Obes Surg 1998; 7: 500-04.

6. Pories WJ, Swanson MS, MacDonald KG, et al. Who Would Have Thought It?: ii. an operation proves to be the most effective therapy for adult-onset diabetes mellitus. Ann Surg 1995; 222: 339-52.

7. Robert E. Brolin, from the proceedings of the 14th annual meeting of the American Society for Bariatric Surgery. Chicago, Illinois, June 4-7, 1997.

8. Mark T Jaroch. Micronutrient Status After Distal Roux Y Gastric Bypass. Poster presentation from the 83rd annual Clinical Congress of the American College of Surgeons, Chicago, Illinois, October 12-17.

9. Mason EE. Ulcerogenesis in Surgery for Obesity. NBSR Newsletter 1995, 10 (4): 1-3.

The entire contents of this page are used with permission, and are copyright ©1998 by Douglas S. Hess, MD, FACS, et al, all rights reserved. Text, graphics, and HTML code are protected by US and International Copyright Laws, and may not be copied, reprinted, published, translated, hosted, or otherwise distributed by any means without explicit permission. Correspondence should be directed to Douglas S. Hess, M.D., 640 South Wintergarden Road, Bowling Green, Ohio 43402, USA. Telephone: 419-352-1452; Fax: 419-352-1244