The Evolution

Morbid obesity and type-2 diabetes are now recognized as one of the main challenges of the twenty-first century. Although their origin maybe multifactorial, genetic components are probably the main defects, and surgery is now considered an accepted treatment leading to meaningful, life-lasting, favorable impact in this population.

Duodenal switch is the version 2.0 from the intervention proposed by Dr. Nicola Scopinaro, the biliopancreatic diversion with distal gastrectomy BPD performed in Genoa in 1979 [1]. Bile and pancreatic secretions are diverted and shunted away from the stomach, with a duodenal switch, rediscovered to decrease biliary reflux by Tom DeMeester et al. in 1987, an amelioration from the older Mann–Williamson procedure [2]. The gastrectomy technique was modified in 1988–1990, to perform a longitudinal gastrectomy (instead of a distal gastrectomy), in which the greater curvature was removed, leaving the lesser curve and vagal innervation, antrum and pylorus, and at a fair distance from the gastroesophageal junction. On the intestinal side, the common channel length was increased from 50 to 100 cm, [3, 4] therefore leaving an alimentary limb of 150 cm and leaving a length of biliopancreatic limb, anywhere from 100 to 350 cm. Dr. Douglass Hess from Ohio has used a percentage from the total intestinal length, an impractical way to collect data in the long term, basically making the common channel at 10% of the total small bowel, and alimentary limb with common channel at 40%. The first laparoscopic duodenal switch (DS) , by Dr. Michel Gagner, was performed on July 2, 1999, at the Mount Sinai Hospital in New York City and published in 2000 [5]. These efforts had been supported prior to that from an animal study, using the porcine model at the research institute of Mount Sinai School of Medicine, with the help of Dr. Gregg Jossart and Dr. John DeCsepel who were clinical and research fellows at the time [6]. We have now reached a worldwide experience of 20 years for the laparoscopic DS.

What was described laparoscopically, included first a sleeve gastrectomy which provided minimal to moderate restriction, maintaining gastric emptying but most importantly decreased acid production to reduce the yield of marginal ulcers frequently seen in classic BPD operations: They had a gastroileostomy with a large gastric pouch of at least 200 ml, which in BPD had no gastric retention leading to severe dumping syndromes [7, 8]. However, gastric emptying and intestinal transit time after laparoscopic sleeve gastrectomy have been shown to be accelerated, by 20–50%. Secondly, the 150 cm alimentary limb and 100 cm common channel total an absorption length of 250 cm and decreases overall caloric absorption. Thirdly, the 100 cm common channel, where food mixes with biliopancreatic juices, results in diminished protein and fat absorption, which I like to call hypoabsorption, since the absorption mechanisms are intact, rather than awkwardly stating “malabsorption.” Laparoscopic DS has also evolved recently, into a version 3.0, with a simpler procedure called SADI, for single anastomosis duodeno-ileostomy , where a loop of ileum at 250–300 cm is anastomosed at the pylorus, but with out a Roux connection, avoiding an ileoileostomy [9]. This will not be discussed in this chapter, as it is an entirely new set of preliminary data involved.

The Technical Elements

The surgeon may stand on the right side or between the patient’s legs and the assistant to the left side, except for the intestinal part of the procedure where both are on the patient’s left. The first 12 mm trocar is then placed in the umbilicus, for a 30° laparoscope. Two 12 mm ports are placed at the same level in the left and right upper quadrants. Additional 5 mm ports are placed, in the epigastria for a liver retractor, one in the left subcostal area for the assistant and one in the left lower quadrant for the intestinal part of the procedure. The short gastric vessels and branches from the gastro-epiploic arcade are divided off from the greater curvature, using ultrasonic shears, from about 6 cm proximal to the pylorus up to the angle of His.

The assistant forceps are placed on the lateral antrum from the left subcostal port, applying a left and upward traction, giving an excellent duodenal exposure. The pylorus is identified, and the peritoneum at the inferior and superior edge of the duodenum is opened, the gastroduodenal artery serving as a lateral limit of the transection and dissection.

This is called the inferior approach, which consists in a complete mobilization of the inferior and posterior attachments of the duodenum. The inferior antrum, pylorus, and first part of the duodenum are isolated using ultrasonic sears and clips. A window is then created on the upper aspect of the duodenum, to accommodate the anvil of a 60 mm linear stapler with a blue cartridge, and most often I use staple line reinforcement using an absorbable membrane, decreasing bleeding and avoiding the need for the duodenal stump oversewing. It also helps in straightening the duodenal wall for suturing during the anastomosis and provides a full opening preventing a stenosis.

A 60-French bougie is introduced, and gastric transection is started 6 cm from the pylorus using black and or green cartridges for the first two to three firings. As the transection progresses toward the fundus, the height of the staples is decreased from green to gold cartridges. The gastrectomy specimen is then extracted trough the umbilical trocar.

The patient is then placed in a Trendelenburg position with the left side down, and measurements are initiated retrograde at the ileocecal junction; a common channel of 100 cm is measured, the future site of the ileoileostomy, using clips on the mesentery. I prefer an umbilical tape of 50 cm for consistent measurement, with flat atraumatic bowel fenestrated forceps. The ileum is then run another 150 cm proximally, and an end-to-side duodeno-ileostomy handsewn is performed, making sure that no twists have occurred (250 cm from the ileocecal valve): It habitually needs a second verification from the ileocecal valve.

The first running posterior layer is made, joining the anti-mesenteric side of the small bowel to the proximal duodenum with 3-0 monofilament absorbable sutures , going from cephalad to caudad. A 2 cm enterotomy, as wide as possible, is made on each intestinal side, and another running suture is used to close the anterior part of the anastomosis, using a similar suture material. The ileum left to the anastomosis is divided with a laparoscopic stapler, and then an ileoileostomy created, side to side joining the end of the biliopancreatic limb with the common cannel at 100 cm from the ileocecal valve. The stump of the biliary limb should be on the left, and an anastomosis is created using a white cartridge of a 60 mm linear stapler, and the common enterotomy closed a single layer of 3-0 absorbable sutures.

The two mesenteric defects created are closed from the left using a 2-0 silk suture, starting at the ileoileostomy transversely. The patient is then placed head up with the left side up, and the transverse colon is lifted up to expose the Petersen’s defect, which is then closed using also 2-0 silk suture on the left side, uniting the transverse mesocolon to the mesentery of the ileum going antecolic. Some surgeons will do a retrocolic passage, and in this case the meso-window, a third mesenteric defect, has to be judiciously closed circumferentially .

Outcomes

The DS has proven to be very effective in achieving and maintaining meaningful weight loss in the morbid and super-obese population (BMI >50 kg/m²). In Dr. Henry Buchwald’s systematic review, comparing weight loss surgical procedures, he suggested that BPD and duodenal switch are the most effective operations giving excess body weight loss (EBWL) of 73% at 2 years, which is maintained for 15–20 years; the Roux-en-Y gastric bypass stays behind in the 50–60% long-term, vertical banded gastroplasty at around similar rates 56% and gastric banding giving the poorest results below 50% [10]. BPD has been abandoned in favor of the DS, due to an unacceptably high rate of revisions for malnutrition, from a too undersized common channel of 50 cm, and to a higher marginal ulceration rate, from a too sizeable gastric pouch. VBGs have also fell tremendously, barely done these days, and replaced by the sleeve gastrectomy, avoiding the foreign body problem (mesh made of polypropylene or PTFE) that frequently eroded into the gastric wall. Sleeve gastrectomy was not part of this Buchwald review, as it started in 2000, but 10-year results show similar findings as the vertical banded gastroplasty, that is, around 50% EWL. This has been confirmed by systematic meta-analysis from a group in Australia [11].

In terms of level I evidence, Sovik et al. performed a randomized study of 60 super-obese patients (BMI 50–60 kg/m²) to undergo either RYGB or DS and found EBWL to be 22.3, 44.0, and 54.4% following RYGB at 6 weeks, 6 months, and 1 year, respectively, compared to 28.1, 59.9, and 74.8% following DS [12]. This was republished with 5-year results confirming the superiority of DS and its great effect on type-2 diabetes [13].

For large cohorts, especially followed assiduously in countries with national healthcare system, the province of Quebec in Canada has been ideal, where all patients are easily tracked as they rarely move to other provinces or countries. Biertho et al., from Laval University in Quebec City, showed in a study of 810 morbidly obese patients with mean initial BMI of 44.2 + 3.6 kg/m² that EBWL plateaued at 76%, with a mean follow-up of 8.6 years [14]. Most US and European surgeons would think that DS is appropriate for super-obese patients, but in Quebec they are also done for the same indication as RYGB. It was concluded that DS was appropriate for morbidly obese patients as well (BMI >40 kgm²) [14]. Concordantly, Anthone et al. in a review of 701 DS patients with preoperative BMIs ranging from 34–95 kg/m² found an EBWL of 69% after 1 year, 73% after 3 years, and 66% after 5 or more years follow-up [15]. Overall, studies examining the outcome of DS suggest analogous results with EBWL ranging from 61% to 85% with moderate term follow-up [16–22].

The DS has also a striking effect on obesity-related comorbidities. Mingrone and Rubino had randomized 60 obese patients with T2DM to receive medical therapy (lifestyle adjustments and hypoglycemic agents) or surgical intervention (RYGB or BPD ). They reported no remission of T2DM in the medical therapy group, compared to 75% in the RYGB group and 95% in the BPD group after 2 years of follow-up [23]. That study was followed and reported in Lancet with 5-year results, showing a tremendously maintained effect of the BPD , almost by twofold, when compared to RYGB [24]. Overall, 19 (50%) of the 38 surgical patients (37% of the gastric bypass group and 63% of the biliopancreatic diversion group) maintained diabetes remission at 5 years, compared with none of the 15 medically treated patients (p = 0·0007). Eight (42%) patients who underwent gastric bypass and 13 (68%) patients who underwent biliopancreatic diversion had an HbA1c concentration of 6.5% or less, compared with 4 (27%) medically treated patients (p = 0·04) [24].

Iaconelli et al. [25] and Tsoli et al. [26] showed resolution of T2DM in all BPD patients 12 months after surgery. From the Cornell study, no patients had T2DM after 9 years [27]. A systematic review and meta-analysis confirmed that diabetes resolution was greatest for patients undergoing DS, followed by RYGB, and slightest for banding procedures [28]. Bariatric surgery has now been recommended for management of T2DM for selected obese patients (BMI >35 kgm²) by the International Diabetes Federation; however, no specific surgical procedure was recommended [29, 30]. This panel was overrepresented by surgeons who did only RYGB, and very few DS surgeons were present at the conference, hence biases are such that a DS option was not well defended and represented. This may change in the future with the SADI operation, a version 3.0 of DS [30]. Astiarraga et al. assessed the effect of DS on T2DM in nonobese patients demonstrating amelioration of metabolic control and remission in one-third of patients, suggesting a weight-independent effect of the intervention [31].

Other cardiometabolic parameters, such as hypertension, nonalcoholic fatty liver disease, hypercholesterolemia, and hypertriglyceridemia, have also shown discernable correction ensuing DS [16–22]. Due to the profound lipid hypoabsorption, remission rates for dyslipidemia have been shown to be between 95 and 100% [16, 17, 32]. Furthermore, obstructive sleep apnea and hypopneas resolved in a majority of patients [32].

Treatment of Related Complications

A recent paper by Buchwald and Oien revealed that the proportion of DS procedures in relation to all bariatric surgeries declined from 6.1% to 4.9% to 2.1% in 2003, 2008, and 2011, respectively, and in the USA, it is at about 1% or less [33, 34]. This raises questions on why the procedure with the highest weight loss, lasting effects, and greatest reversal of obesity related comorbidities is the least performed bariatric intervention. The answer is likely multifactorial and complex. Firstly, the technical complexity of this procedure is a consideration, with the procedure being time-consuming and requiring a skilled laparoscopic bariatric surgeon. The laparoscopic approach, introduced by me in 1999, sought the benefits of DS weight loss and reduced morbidity associated with laparoscopic surgery [5]. In many studies, this has proven to be true, with lower postoperative complication rates [35, 36]. Likely, learning curves and operative volumes may be important considerations, with a majority of DS being performed at focused bariatric centers [37, 38]. Many fellowships do not have DS in their armamentarium, and duodenal dissection and transection are seen as a risky procedure for bile duct injury, hepatic arterial injury, or associated with high-leak rate at this upper anastomosis. These fears have not been confirmed. Also, it is the sleeve gastrectomy effect, as more and more bariatric surgeons are doing this procedure, now 70% of all primary bariatric procedures in the USA; it has reduced the use of DS which is now mostly used as a second-stage approach. The last 2 years have seen a resurgence of DS by 60%, mostly due to weight regain after sleeve gastrectomy [39].

In a meta-analysis of 361 studies including 85,048 patients, overall mortality within 30 days of bariatric surgery was found to by 0.28%. DS had the highest early mortality with a rate of 0.29% to 1.23% for open and 0.0% to 2.7% for laparoscopic procedures [40]. But this has changed in the last 20 years, where one sees the mortality of second-stage DS being similar to RYGB, that is, in the range of 0.1–0.2%.

One-year complication rates have been reported in the Bariatric Outcomes Longitudinal Database (BOLD) ; they are 4.6, 10.8, 14.9, and 25.7%, respectively, after LAGB, LSG, RYGB, and DS [41]. This comprises of minor complications such as gastrointestinal side effects including flatulence, malodorous stools, and major complications like anastomotic leak, the most common serious early surgical complication, which should be less than 1%. One reason for a higher complication rate in DS is the fact that this procedure is often selected for super-obese patients, notoriously known for a higher complication rate. Avoidance of tension at the anastomosis is paramount, and the intestinal reconstructive part can be deferred as a second stage using good judgement. Hamoui et al. reviewed 701 DS cases performed over a 10-year period and reported that 5% of patients developed complications necessitating revisional surgery [42]. Protein malnutrition was the most common indication for reoperation. A postoperative complication rate of 15% was then seen in their revisional surgery group, with wound infections being the most common complication in this surgical team [42]. This often necessitates an elongation procedure in which the common channel is elongated by at least 100 cm at the expense of the biliopancreatic limb. This is not a complex laparoscopic procedure, as it involves transection of the alimentary limb at the connection and moving this proximally on the BP limb and making a side-to-side anastomosis. The end result is a common channel of 200 cm and a total distance between the duodenum and ileocecal valve of 350 cm.

Biertho et al. analyzed a series of 1000 DS patients, in which major complications occurred in 7% of patients, readmission was required in 12.7% of patients, and reoperations occurred in 6% of patients [36]. My own series from Cornell has shown no operative mortality, a 1.1% leak rate, 2.5% of surgical site infection, and 12.4% of reoperations over 9 years, including cholecystectomy, laparoscopy for bowel obstructions from adhesions or internal hernias, etc. [27] BMI was 30.1 kgm² at 1 year and 32.0 kgm² at 9 years, a pretty flat line. Body fat was reduced to 26% from >50% after 2 years. Complications requiring surgery were significant, and nutritional problems developed in 29.8% of patients over the course of observation. The baseline Beck Depression Index (BDI) was 13.9 and 7.2 in year 1, and from year 1 through 9, it remained unchanged [27]. There were significant positive changes in quality of life between baseline and year 1 for most domains. These positive changes were maintained for the follow-up cohorts. After surgery the resolution of comorbidities continued for the 9 years. Weight regain is still a possibility, and most would recommend a resleeve resection, not touching any limb lengths, unless a gross error has been done at initial surgery with too long limbs [43].

In a randomized trial, Sovik et al. compared mean operating time, median length of stay, and complication rates between RYGB and DS. On average RYGB required 91 min compared to 206 min for DS. Median length of stay was 2 days post-RYGB and 4 days post-DS. Perioperative complication rates were comparable between the two groups [12, 13]. But one has to mention that Sovik and their groups were not familiar with DS when they started this randomized trial, and equivalent experience might have been lacking. My own practice sees less than 24-hour stay these days, just like gastric bypass or sleeve gastrectomy, with operative time in the range of 2 hours.

DS is the bariatric procedure linked with some of the grandest perioperative malnutrition-/metabolic-related complications [44]. All patients must begin mineral, vitamin, and protein supplementation postoperatively and sometimes preoperatively; however, there should be a standardized approach to replacement to avoid deficiencies using dietary supplements. Following DS, patients may need 3000 Kcal per day and ingest 80 to 120 g of proteins, to overcome the hypoabsorption; otherwise they may become malnourished [45]. Iron-deficiency anemia, protein calorie malnutrition, hypocalcemia, and deficiency of fat-soluble vitamins ADEK, vitamin B1, vitamin B12, and folate are common [46]. DS has proven to be more hypoabsorptive compared to other bariatric surgeries; thus close follow-up is essential. Supplementation is of paramount importance; regrettably in this patient population compliance is lacking [47].

Aasheim et al. randomized 60 super-obese patients to receive either RYGB or DS compared vitamin D , vitamin A, and vitamin B1 levels postoperatively [48]. DS patients had lessened mean vitamin D and A serum concentrations, as well as an abrupt decline in vitamin thiamine compared to RYGB. But in this trial, insufficient supplements were given to DS patients owing to the early learning curve of this group. Decreased vitamin D and calcium levels with associated secondary hyperparathyroidism have been demonstrated [49–51]. Marceau’s group at Laval has confirmed with bone biopsies, with maintained serum PTH, that overall bone mineral density and fracture risks were unchanged 10 years after DS [52]. Clinically there have been case reports of DS-related vitamin A deficiency and associated night-blindness, post-DS peripheral neuropathies associated with B12 deficiencies, although rare, Wernicke’s encephalopathy as a result of B1 deficiencies can be seen, especially if there is a stenosis in the sleeve or a stricture at the duodeno-ileostomy anastomosis [53–57]. From my own series of DS patients operated at Cornell, between 1999 and 2010, 274 patients were operated, but 190 patients have been able to be followed through (70% women): age 42.7 years, BMI 53.0 kg/m²; year 1, 189 were available; year 3–193; year 5–132; year 7–98; year 9–68. BMI was 33.3 kgm² at year 1 and 31.5 kgm² at year 9. Baseline vitamin D was low and PTH high [58]. All patients took some supplements. Fat-soluble vitamins remained low, and protein deficiency appeared at year 3 and increased to 30% at year 9. At baseline serum zinc was normal, but at year 5, 45% were low. Hematocrit was low for 40% and hemoglobin for 46%. Iron deficiency continued through year 9, more marked in males. Calcium deficiency increased from year 3 and became steady. Half of patients had abnormal PTH at baseline and the percentage increased. Twenty percent had abnormal baseline magnesium values. This fluctuated during yearly observations [58]. I think most groups are not giving enough supplementations, and compliance is a huge problem, but when serum levels are low, it remains an opportunity to educate and correct.

Conclusions

Duodenal switch offers one of the best long-term controls of obesity-related diseases such as type-2 diabetes and is associated with one of the lowest risk of weight regain long-term. It should be part of surgeons’ armamentarium, particularly for the management of weight regain following sleeve gastrectomy. Like any advanced surgical procedure, there is a learning curve associated with laparoscopic DS, but standardization of the different surgical steps allows keeping complication rates low. The use of a hand-sewn anastomosis allows keeping the risk of anastomotic leak to the lowest. In case of intraoperative difficulties, the procedure can be aborted to a stand-alone sleeve gastrectomy, and a duodenal switch can be performed 18 months to several years after the initial surgery, without losing its efficacy.