Surgical Endoscopy

, Volume 22, Issue 7, pp 1690–1696

Roux-en-Y gastric bypass procedure performed with the da Vinci robot system: is it worth it?

Authors

    • Department of Abdominal SurgeryUniversity Hospital Antwerpen
  • L. Balliu
    • Department of Abdominal SurgeryUniversity Hospital Antwerpen
  • M. Ruppert
    • Department of Abdominal SurgeryUniversity Hospital Antwerpen
  • B. Gypen
    • Department of Abdominal SurgeryUniversity Hospital Antwerpen
  • T. Van Tu
    • Department of Abdominal SurgeryUniversity Hospital Antwerpen
  • W. Vaneerdeweg
    • Department of Abdominal SurgeryUniversity Hospital Antwerpen
Article

DOI: 10.1007/s00464-007-9698-6

Cite this article as:
Hubens, G., Balliu, L., Ruppert, M. et al. Surg Endosc (2008) 22: 1690. doi:10.1007/s00464-007-9698-6

Abstract

Background

The Roux-en-Y gastric bypass procedure (RYGBP) is in many countries the gold standard for obtaining long-lasting weight reduction and improvement of obesity-related comorbidities. However, performing this operation by standard laparoscopic techniques requires important surgical skills because of the anastomoses involved. The da Vinci surgical robot system with its enhanced degrees of freedom in motion and three-dimensional vision is designed to overcome the difficulties encountered in traditional laparoscopic surgery with suturing and delicate tissue handling.

Methods

For this study, 45 patients (9 men) with a mean body mass index (BMI) of 44.2 (range, 35.1–55.4) underwent RYGBP with the aid of the da Vinci robot system. They were compared with 45 consecutive patients with a mean BMI of 43.9 (range, 35.1–56.2) who underwent a laparoscopic RYGBP by the same surgeon during the same period.

Results

Overall, the total operating time was shorter for the laparoscopic cases (127 vs 212 min; p < 0.05). However, the last 10 robotic cases were performed in the same time span as the laparoscopic cases (136 vs 127 min). The total robotic setup time remained constant at about 30 min. There were no differences in postoperative complications between the two groups in terms of anastomotic leakage or stenosis. In the robotic group, more conversions to open surgery were noted. Early in the study, four patients (9%) had to undergo conversion to standard laparoscopic techniques due to inadequate setup of the robotic arms. Five patients (11%), however, had to undergo conversion to open surgery because of intestinal laceration during manipulation of the intestines with the robotic instruments. The costs were higher for robotic surgery than for standard laparoscopic RYGBP, mainly because of the extra equipment used, such as ultrasonic devices.

Conclusion

The RYGBP procedure can be performed safely with the da Vinci robot after a learning curve of about 35 cases. At this writing, however, it is not clear whether the da Vinci system offers a real advantage over standard laparoscopic techniques.

Keywords

BariatricHumanObesityRoboticTechnical

Morbid obesity is one of the major health problems in the Western world. At this writing, bariatric surgery offers the only possible treatment for long-lasting, sustained weight loss with improvement in obesity-related comorbidities. The number of these procedures worldwide is rising exponentially. In 2003, an estimated 100,000 bariatric surgical procedures were performed in the United States, a 40% increase compared with 2002 [1].

Recent years, in Belgium, as in other Western European countries, have seen a shift from purely restrictive procedures to the Roux-en-Y gastric bypass (RYGBP) as the gold standard for the surgical treatment of morbid obesity. This operation has been shown to produce a long-term excessive weight loss (EWL) of approximately 70% to 80%, with a remarkable reduction in obesity-related comorbidities [2]. Large series demonstrate an early postoperative complication rate of 3% to 11% [37], whereas long-term complication rates vary from 2% to 27% [8].

Despite the good results obtained by the experts in centers of excellence, many surgeons still consider the RYGBP to be one of the most difficult interventions to perform in laparoscopically. A learning curve of at least 75 to 100 interventions generally is acknowledged [4, 5].

Recently, the da Vinci robot system (Intuitive Surgical, Inc., Sunnyvale, CA, USA) has been introduced in the surgical world, and its feasibility for bariatric surgical procedures also has been documented [913]. With its three-dimensional vision, articulated instruments, and enhanced degrees of freedom in the motions of these instruments, this device offers theoretical advantages for performing difficult laparoscopic tasks such as the creation of a meticulous anastomosis.

A few centers in the United States have reported their experience with robotics and the RYGBP [1113]. In most of these centers, the gastrojejunostomy is performed with the aid of the robot, whereas the remaining procedure is performed by traditional laparoscopic techniques. To date, only one study on the feasibility of a complete robotic RYGBP has been published [12].

The current study is the first European series using the same technique for this advanced laparoscopic procedure. We compared our first 45 robotic RYGBPs with 45 laparoscopic RYGBPs performed by a senior registrar and supervised by the same surgeon (G.H.) during the same period.

Material and methods

Between October 2004 and April 2006, 152 patients underwent surgery for morbid obesity in our department. All the patients were part of a multidisciplinary preoperative approach and fulfilled the 1991 National Institutes of Health (NIH) criteria for bariatric surgery. A robotic RYGBP was performed for 45 patients. These interventions all were performed with the same surgeon (G.H.) at the console, but with varying surgical trainees at the operating table. All the trainees were able to perform a standard laparoscopic RYGBP.

For comparison, the last 45 consecutive patients treated in a standard laparoscopic manner supervised by the same surgeon (G.H.) in the same period were used. The laparoscopic technique described by Olbers et al. [14] is used in our department. Both the gastrojejunal anastomosis and the enteroenterostomy are performed using a linear stapler with closure of the resulting defects via a running Vicryl 3–0 suture. The characteristics of both patient groups are found in Table 1.
Table 1

Patient demographics

 

Robotic

Laparoscopic

N

45

45

Mean age: n (range)

42 (21–62)

39 (23–61)

M/F

9/36

8/37

Mean BMI: n (range)

44.2 (35.1–55.4)

43.9 (35.1–56.2)

ASA

    1–2

43

44

    3–4

2

1

Comorbidities

    IDDM

8

5

    NIDDM

11

12

    AHT

19

21

    OSAS

17

12

    Metabolic syndrome

35

31

    Joint problems

34

38

Previous abdominal surgery

4

2

Appendectomy

3

1

Lapasoscopic banding

1

1

BMI, body mass index; ASA, American Society of Anesthesiology; IDDM, insulin-dependent diabetes mellitus; NIDDM, non–insulin-dependent diabetes mellitus; AHT, arterial hypertension; OSAS, obstructive sleep apnea syndrome

We compared the operating times between the two operative techniques as well as the perioperative and short-term postoperative complications, then tried to estimate the costs of the robotic procedure. Only the costs of the equipment for the procedure were taken into account. No estimation of costs for eventual extra staff and hospital stay was made.

Statistical evaluation was performed with a one-way analysis of variance (least significant difference) for comparison of the operating times, and Fisher’s exact test was used to compare complications. The statistical program SPSS 13.0.1 was used, and a p value less than 0.05 was considered significant.

Da Vinci technique

Because it is difficult to replace the robotic system during the operation, optimal installation is crucial to a smooth procedure. This includes installation of the patient, the robotic chart, the trocars, and the robotic arms. Patients are installed in a 20° to 25° reversed Trendelenburg position with their right arm extended and their left arm to the side. A scrub nurse drapes both the patient and the robotic arms.

The anesthesiologists are placed at the patient’s right shoulder. The assistant surgeon stands between the patient’s legs, and the scrub nurse is situated on the patient’s right side.

After creating a 15-mmHg pneumoperitoneum, the surgical assistant introduces the necessary ports. Optimal port placement is essential to avoid collisions between the robotic arms and camera arm and to obtain sufficient access both high up in the abdomen and below the mesocolon at the angle of Treitz.

We start by placing five ports (Fig. 1), all 10/12-mm Ethicon Endopath trocars (Ethicon Endosurgery, Cincinnati, OH, USA) with a long shaft. Only when we cannot reach the angle of Hiss with port 4 (Fig. 1) do we add a sixth port at the Mio Clavicular Line (MCL), subcostally left. This port placement and the port types are the same as in the series by Mohr et al. [12]. We also use the double-cannulation technique, with the da Vinci ports placed inside conventional laparoscopic Ethicon ports because this permits the introduction of a stapler after retraction of the robotic arm with the cannula still attached.
https://static-content.springer.com/image/art%3A10.1007%2Fs00464-007-9698-6/MediaObjects/464_2007_9698_Fig1_HTML.jpg
Fig. 1

Trocar placement. (1) da Vinci robotic arm port. (2) da Vinci optic port. (3) Assistant port. (4) da Vinci robotic arm port. (5) Assistant port (liver retractor). (6) Optional da Vinci port

After port placement, the robotic chart is brought in over patient’s left shoulder, making a 20° angle with the patient’s midline. The arms then are attached to the da Vinci cannulas and introduced into the ports. The position of the arms also is very important for maximum range of motions without collisions. Again, we are happy with the arm positions described in the U.S. study by Mohr et al. [12]. Basically, the right robotic arm is brought in low, and the proximal part of the left arm is placed vertically with the U-shaped metal frame facing the camera arm (Fig. 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs00464-007-9698-6/MediaObjects/464_2007_9698_Fig2_HTML.jpg
Fig. 2

Robotic arm placement. Note the vertical alignment of the left arm with the metal frame facing the camera

Once the setup is complete, the procedure starts, with the console surgeon looking for the angle of Treitz and grasping the jejunum with two bowel clamps (Intuitive Surgical, Inc.) while the assistant holds the transverse colon and mesocolon upward. The jejunum is presented by the console surgeon to the assistant, who divides it with a 45-mm Ethicon Endostapler (white cartridge) approximately 40 to 50 cm from the angle of Treitz. The mesentery is the further divided by the assistant using ultrasonic scissors. We mark the biliopancreatic site with a small stitch to avoid creating a Roux-en-O, which is easily possible due to the lack of overview the surgeon has working close to the intestine. A Roux limb of 80 to 120 cm is measured by the console surgeon using the Intuitive bowel clamps. At this site, an enteroenterostomy with the biliopancreatic loop is prepared.

The two bowel clamps are replaced by two needle drivers, and the console surgeon first places two stay stitches to align the bowel segments. These stitches are left long so the assistant can retract the bowel segments toward the patient’s right upper side by means of a clamp introduced in port 5. By means of an electrocautery hook, the console surgeon makes two enterotomies, after which a complete hand-sewn enteroenterostomy is made using a running Vicryl 3–0 suture. Once this is completed, the two needle drivers are changed again for two bowel clamps, and the edge of the omentum is presented to the assistant. The omentum is split using ultrasonic scissors or the electrocautery hook.

We move next to the upper abdomen. If we cannot reach the angle of Hiss with the left robotic arm, an additional da Vinci port is placed subcostally. The liver is retracted by a liver retractor in port 5. The angle of Hiss is dissected with an electrocautery hook by the console surgeon, who then starts dissecting on the lesser curvature between the second and third veins.

While the assistant surgeon holds the lesser curvature elevated, the retrogastric space is entered so there is room to place the stapling device. The assistant surgeon fires one stapler (blue) cartridge horizontally from the lesser curvature and three to four stapler cartridges upward toward the angle of Hiss to create the pouch. Then the console surgeon brings up the alimentary loop in an antecolic and antegastric manner and fixes it to the gastric pouch with two stay stitches at the corners. A running one-layer anastomosis is created with Vicryl 3–0 sutures. Finally, the anastomosis is tested with a methylene blue test to check for leaks and patency. The robot then is withdrawn, and the skin incisions are closed.

Results

The times needed to perform the procedures are shown in Table 2. For the robotic procedures, we measured both the setup time, including draping of the robot, installation of the chart, connecting of the robotic arms, and the actual operating time. Overall, the laparoscopic RYGBP was performed faster than the procedures performed with the aid of the da Vinci robot system (mean, 127 vs 212 min; p < 0.05).
Table 2

Mean operating times in minutes

 

Robot setup n (range)

Robotic n (range)

Laparoscopic n (range)

Overall

30.2 (17–42)

212 (110–325)

127 (90–160)a

Cases 1–15

32.4 (22–42)

231 (190–280)

 

Cases 16–25

29.6 (17–39)

215 (180–247)

 

Cases 26–35

28.8 (22–41)

257 (210–325)

 

Cases 36–45

28.0 (22–34)

136 (110–150)b

 

Nonsignificant difference (NS)

ap < 0.05 vs robot

bp < 0.05 vs cases 26–35 vs cases 16–25 vs cases 1–15; NS vs laparoscopic

To determine whether there was a learning curve, we divided our cases into four groups. The first 15 cases required a mean of 231 min (range, 190–280 min), whereas the mean time for the last 10 cases was 136 min (range, 110–150 min). This was significantly less time (p < 0.05) than required for the first 15 cases, but not significantly different from the mean operating time of 127 min (range, 90–160 min) for the laparoscopically RYGBP performed by the same surgeon. This implies a learning curve of about 35 patients before a robotic RYGBP can be performed in the same time span as an average laparoscopic RYGBP. The mean setup time of the robotic system remained fairly constant at about 30 min throughout all the procedures.

Interestingly, however, we saw a rise in the operating time for the third group of robotic patients (cases 26–35). This was due to the occurrence of some perioperative complications, which prolonged the intervention and eventually caused conversion to open procedures. Tables 3 and 4 give an overview of our complications. There were no mortalities in either group. Although conversion is not a complication as such, the reason for conversion often was a complication.
Table 3

Perioperative complications

 

Robotic n (%)

Laparoscopic n (%)

Conversion

9 (20)

0

To laparoscopy

4 (9)a

To open surgery

5 (11)

Leakage on M-B test

0

2 (4)

M-B, methylene blue

ap = 0.056 vs laparoscopic

Table 4

Postoperative complications

 

Robotic n (%)

Laparoscopic n (%)

Leakage G-E anastomosis

0

1 (2.2)

Leakage E-E anastomosis

0

1 (2.2)

Stenosis G-E anastomosis

2 (4.4)

2 (4.4)

Bleeding gastric remnant

0

1 (2.2)

Perforation gastric remnant

0

1 (2.2)

Torsion alimentary limb

1 (2.2)

0

Total

3 (6.6)

6 (13.3)

Revisional surgery needed

2 (4.4)

2 (4.4)

In the laparoscopic group, there were no conversions to open surgery. In the robotic group, four patients (9%) underwent conversion to laparoscopic surgery. All but one conversion occurred in the first 25 procedures and were due to improper placement of the da Vinci ports, causing abundant collisions of the arms with the optics.

More seriously, however, five patients had to undergo conversion to open surgery (11%; p = 0.056, compared with laparoscopic group). Three of these conversions happened rather late in our experience (cases 27, 29, and 30). Conversion to open surgery was necessary because of substantial jejunal tears that occurred during manipulation of the alimentary loop. Two of the converted patients had a previous open appendectomy with some adhesions. In one patient (case 34), the biliopancreatic limb was anastomosed to itself, creating an O loop. A tear occurred during manipulation while the clinicians tried to figure out the anatomy. Therefore, the robotic procedure was completed for 36 patients (80%).

Two patients (4.4%) experienced tears on the gastric pouch due to manipulation of the pouch with the da Vinci instruments. These tears were easily sutured with a running suture and caused no postoperative problems.

We considered major postoperative complications those necessitating revisional surgery or causing a prolonged hospital stay including anastomotic leaks, stenosis, or intestinal obstruction.

In the robotic group, we saw no leaks at the gastrojejunostomy site either immediately when a methylene blue test of the gastrojejunostomy was performed, or at a standard gastrografin swallow 3 days postoperatively, or clinically during the first 2 postoperative weeks. In the laparoscopic group, extravasation of the methylene blue dye was noted in two patients, necessitating an additional stitch. In one of these patients, a small postoperative clinical leak was noted on the postoperative gastrografin swallow, which was treated conservatively.

Stenosis of the gastroenterostomy due to an anastomotic ulcer was noted in two patients (4.4%) each in the robotic and laparoscopic groups during a follow-up period of 3 to 20 months. Both patients in the laparoscopic group could be treated with endoscopic dilations, whereas one patient in the robotic group needed revisional surgery 3 weeks after the primary intervention. One patient in the robotic group was readmitted because of heavy vomiting 5 weeks after the primary intervention. Reexploration showed torsion of the alimentary limb, which was adjusted.

Problems with the gastric remnant were encountered in two patients of the laparoscopic group. The one patient had bleeding from the stapler line needing revisional surgery, whereas the other patient had a perforation of a duodenal ulcer 4 months after the laparoscopic RYGBP.

Altogether, revisional surgery was necessary for two patients in each group. There was no statistically significant difference in the complication rates between the robotic and laparoscopic groups.

The postoperative hospital stay did not differ between the two groups. The patients stayed a mean of 4.7 days after surgery.

A attempt was made to estimate the costs of the robotic RYGBP procedures and to compare them with the costs of traditional laparoscopic RYGBP (Table 5). The costs for a robotic procedure in our institution varies from 1437€ to 4244€ (mean, 2761€). This includes the costs for the robotic instruments, stapling cartridges, and additional laparoscopic instruments. The variation is mainly because of the decision to use or not to use ultracision instruments. For a traditional laparoscopic procedure, the costs are fixed because we make use of “tailor-made” RYGBP kits negotiated with the manufacturer and reusable trocars.
Table 5

Costs of laparoscopic and robotic procedures in € (only for material; no estimation of costs for operating time, staff, etc.)

 

Robotic

Laparoscopic

 

Minimal

Mean

Maximal

 

Total costs

1,437

2,761

4,244

1,766

Robotic costs

834

1,929

2,405

Extra laparoscopic material

69

832

2,162

Discussion

The introduction of laparoscopic techniques is a factor that has contributed to the growing numbers of bariatric procedures performed worldwide. In many countries, the laparoscopic Roux-en-Y gastric bypass procedure has become the operation of choice for obtaining long-term weight reduction and significant improvement in obesity-related comorbidities. However, many surgeons still consider a laparoscopic RYGBP to be a difficult operation requiring substantial laparoscopic skills. With traditional laparoscopic RYGBP, the gastrojejunostomy can be performed by a linear stapling technique [15], a circular stapling technique [16], or a completely hand-sewn approach [8], whereas the enteroenterostomy usually is performed by a linear stapling technique or a completely hand-sewn approach. In large series, the postoperative complication rate is reported to reach 3% to 10.5% [8].

Depending on the technique, anastomotic leaks are noted in 0% to 3% [8] and gastrojejunal stenosis in 1.6% to 6.3% of cases [8]. However, these numbers reflect the experience in centers of excellence, and it is probable that the actual numbers are much higher.

As robotic techniques were introduced into the surgical armentarium, an attempt was made to combine the advantages of minimally invasive surgery (less trauma to the patient, quicker recovery, less postoperative morbidity) with the advantages of open surgery (three-dimensional vision, more precise tissue handling, more degrees of freedom in manipulating instruments, and better ergonomics for the surgeon). We [17] and others [18, 19] have shown that with the aid of the da Vinci robot system, various surgical tasks can be performed significantly faster and more precisely than with standard laparoscopic techniques.

Surgical robots have been used in bariatric surgery in both Europe [10] and the United States [9, 1113]. An Austrian study [10] used the da Vinci robot system for 10 patients undergoing a silicone adjustable gastric banding or an implantable gastric stimulator system. With these very simple procedures, they found no clear advantage with the surgical robot. Theoretically, the more difficult RYGBP might benefit much more from the robotic advantages. Some U.S. groups [1113] have published their results on the subject. In a survey conducted by Jacobsen et al. [9], 6 of 11 U.S. surgeons using the da Vinci system for bariatric surgery had experience in robot-assisted RYGBP. They all found the robotically assisted hand-sewn gastrojejunostomy superior and technically easier to perform than any standard laparoscopic technique. However, all these groups used the robot system only to create the anastomosis, performing the rest of the procedure in a traditional laparoscopic way.

To date, only one study [12] on a totally robotic Roux-en-Y gastric bypass (taking into account that a person familiar with the procedure and stapling techniques is still present at the operating table) has been published. In this article, Mohr et al. [12] described in detail the feasibility and technical aspects for their first 10 patients treated with a complete da Vinci robot RYGBP and compared them with the first 10 laparoscopic procedures performed by the same surgeon. They should be credited for their extensive research on optimal trocar and robotic arm placement, which is indeed of crucial importance for a successful procedure.

Because it is difficult and time consuming to change the robotic settings during the procedure, clinicians should look for optimal positioning that allows access to both the upper and lower left abdominal quadrants. We agree with Mohr et al. [12] that the special position of the left robotic arm with a vertical external yaw axis is absolutely necessary to obtain the widest range of motion possible and to have an easy access in both quadrants. Moving the right robotic arm from port position 4 to a higher subcostal position 6 (Fig. 1) really depends on the patient’s torso and should be done only after the trocar, pushed as deep as possible inside the patient, still does not reach the angle of Hiss with the da Vinci instruments.

Our technique differs from that by Mohr et al. [12] in that we also do a complete hand-sewn anastomosis for the enteroenterostomy instead of using the linear stapler method, because we think that the real benefit of the robot lies in facilitating the anastomosis. To do this in a comfortable way, it is necessary to place a stay stitch on both bowel ends, allowing for the assistant surgeon to pull the anastomotic site upward and to the right of the patient. This can easily be done by means of a clamp placed in port 5 (Fig. 1), which in the second stage of the procedure will be used as a port for the liver retractor.

The setup of the robotic system invariably took about 30 min and did not change significantly with the number of procedures. A possible explanation may be that the nursing team responsible for setting up the system varied considerably. With such variety, it is more difficult to obtain a large experience. In contrast, we noted a decrease in total operating time over the cases. After about 35 cases, the time needed to perform the procedure was comparable with the time needed by a senior registrar to do an average laparoscopic RYGBP.

In a series of robotically performed Nissen procedures and cholecystectomies, Giulianotti et al. [20] found that about 20 cases were necessary to obtain operating times comparable with those for traditional laparoscopic techniques. Of course, this reflects the more complex nature of the RYGBP. Interestingly, we saw that after a decline in operating times for the second group of robotically treated patients (cases 16–25) compared with those for the first group (although not statistically significant), the operating times for the third group were again longer. This could be explained by the occurrence of some important perioperative complications. Probably, at that moment in our learning curve, we were becoming too confident. Installation of the trocars and arms was fairly consistent, and we had the feeling that “we had a grip” on the robotic procedure. Nevertheless, in our view, the actual robotic instruments for grasping and handling delicate tissue still are less than optimal. This together with the loss of tactile sensation makes it very easy to create bowel or stomach lesions while manipulating tissue, as can be seen by our 11% conversion rate due to jejunal tears . Although this just did not reach statistical significance and we did not encounter this problem any more after case 34, it still remains a cause for strict attention.

Another problem with the da Vinci system is that especially in the first (intestinal) part of the procedure, the surgeon works very close to the bowel, sometimes finding it difficult to have a general overview. If no marked stitches are placed on bowel segments, it is very easy to create a Roux-en-O instead of the desired Roux-en-Y. The newest da Vinci robots, however, have a “fish eye option,” which might overcome this problem.

No statistical difference in postoperative complications rates or revisional surgery was noted between the patients who had surgery with the da Vinci system and those who underwent standard laparoscopic surgery. The anastomotic leakage and stenosis rates were comparable with those in the literature.

With regard to the costs for this type of surgery, we can say that in our institution, the robotic RYGBP is more expensive than the standard laparoscopic RYGBP. This increased cost is caused by the ultrasonic equipment and disposable trocars used in the robotic cases and not in the laparoscopic procedures. The costs of the da Vinci instruments are comparable with those for standard laparoscopic instruments.

In conclusion, the Roux-en-Y gastric bypass can be performed safely in a standardized manner using the da Vinci robot system, with a learning curve of about 35 cases. In our view, robotic techniques certainly facilitate the creation of a completely hand-sewn gastrojejunostomy and enteroenterostomy. Nevertheless, the technique has its own problems, mostly associated with the actual equipment, which still is more adaptable to abdominal surgery, and the loss of tactile sensation. Also this technique still requires a surgical registrar familiar with the procedure and stapling techniques scrubbed at the operating table. However, it is to be foreseen that newer robotic instruments overcoming these problems will be developed. At the moment, we believe it still is too early to say that there is a real advantage in using the robot to perform these types of difficult digestive procedures. Larger studies comparing this technique with the standard laparoscopic technique are necessary.

Copyright information

© Springer Science+Business Media, LLC 2007