Obesity Surgery

, Volume 23, Issue 11, pp 1899–1914

Laparoscopic Roux-en-Y Gastric Bypass or Laparoscopic Sleeve Gastrectomy as Revisional Procedure after Adjustable Gastric Band—a Systematic Review

Authors

    • Department of Surgery, Obesity Center AmsterdamSint Lucas Andreas Hospital
  • Caroline J. Verveld
    • Department of Surgery, Obesity Center AmsterdamSint Lucas Andreas Hospital
  • Bart A. van Wagensveld
    • Department of Surgery, Obesity Center AmsterdamSint Lucas Andreas Hospital
  • Sjoerd M. Lagarde
    • Department of Surgery, Obesity Center AmsterdamSint Lucas Andreas Hospital
Review Article

DOI: 10.1007/s11695-013-1058-0

Cite this article as:
Coblijn, U.K., Verveld, C.J., van Wagensveld, B.A. et al. OBES SURG (2013) 23: 1899. doi:10.1007/s11695-013-1058-0

Abstract

The adjustable gastric band (L)AGB gained popularity as a weight loss procedure. However, long-term results are disappointing; many patients need revision to laparoscopic Roux-en-Y gastric bypass (LRYGB) or sleeve gastrectomy (LSG). The purpose of this study was to assess morbidity, mortality, and results of these two revisional procedures. Fifteen LRYGB studies with a total of 588 patients and eight LSG studies with 286 patients were included. The reason for revision was insufficient weight loss or weight regain in 62.2 and 63.9 % in LRYGB and LSG patients. Short-term complications occurred in 8.5 and 15.7 % and long-term complications in 8.9 and 2.5 %. Reoperation was performed in 6.5 and 3.5 %. Revision to LRYGB or LSG after (L)AGB is feasible and relatively safe. Complication rate is higher than in primary procedures.

Keywords

Bariatric surgeryConversionInsufficient weight loss laparoscopic adjustable gastric bandLaparoscopic Roux-en-Y gastric bypassLaparoscopic sleeve gastrectomyLRYGBLSGMalabsorptionRestrictiveRevisionWeight regain

Introduction

Obesity is a rising concern all over the world. In the USA alone, the prevalence is around 30 % in the adult population [1]. The World Health Organization (WHO) predicts that worldwide, in 2025, there will be 300 million obese people [2]. Obesity is associated with a wide array of comorbidities such as the development of metabolic syndrome, obstructive sleep apnea, cardiovascular disease, and early osteoarthritis [3].

In contrast to bariatric surgery, nonsurgical treatment is ineffective for sustainable weight loss and reduction of associated comorbidities [4]. Bariatric surgery is based on reducing gastric volume (restrictive surgery) or reducing the absorption capacity of the intestines (malabsorptive surgery) or a combination of both. Over the last decades, laparoscopic adjustable gastric band (L)AGB gained popularity because of its relatively low complexity and adjustability in combination with low perioperative morbidity (1–5 %) and mortality rate (0–0.05 %). (L)AGB has good results in the first postoperative period [57]. Furthermore, the procedure is considered reversible; after removing the band, the stomach regains its normal anatomy [8]. Worldwide, the (L)AGB represents approximately 42 % of all bariatric procedures and is the most common bariatric procedure performed in many countries [4, 5, 911].

Despite good results in the first postoperative period, the procedure has several limitations. Band-related complications such as esophageal dilatation, food intolerance, gastric necrosis, band slippage, band dilation, and pouch dilation are reported in 15–58 % of all patients [8, 1216]. Furthermore, a growing number have inadequate weight loss or weight regain after successful initial weight loss. Inadequate weight loss and weight regain is reason for a reoperation in 27–100 % of the patients [17, 18] [19].

There are different surgical options to treat late complications or inadequate weight loss. Removal of the band, without further intervention, is associated with high weight regain [20]. Although (L)AGB repositioning or replacement is technically possible, studies show mixed results (concerning feasibility and long-term weight loss especially when compared to revision into Roux-en-Y gastric bypass (RYGB) or sleeve gastrectomy (LSG) [11, 2124].

Although band placement does not create a permanent anatomic alteration, it does not leave the stomach region undamaged. Erosion, scar tissue, pouch dilation, and adhesions make the area more complex and vulnerable during further interventions; this makes revisional surgery technically demanding [17, 19, 20]. Laparoscopic Roux-en-Y gastric bypass (LRYGB) and LSG are more and more frequently performed as a rescue operation after failed (L)AGB. However, it is unclear which operation should be performed and when. Safety, effectiveness, and timing of revision are subjects of debate. In recent years, many groups have published their (small) series. The purpose of this systematic review was to assess the morbidity, mortality, and long-term results of revisional surgery to LRYGB or gastric sleeve after gastric banding.

Materials and Methods

Literature Search

The Cochrane Database of systematic reviews, the Cochrane central register of controlled trials, and MEDLINE databases were searched by using the keywords (conversion OR revision OR revisional OR revisionary) AND (gastric band OR gastric banding OR bariatric surgery OR gastric bypass OR sleeve) in order to identify studies published up to February 2012. Free-text words instead of MeSH terms were used to avoid missing recent articles that had not been given a MeSH label yet. Three investigators (CJV, UKC, and SML) independently performed the literature search. Electronic links to related articles and references of selected articles were hand searched as well. References were snowballed. A hand search of relevant journals and conference proceedings was not performed. The search was not restricted to any language; however, this systematic review only took studies published in English into account.

Study Selection and Data Extraction

Of the potentially eligible publications, studies were included if they reported (based on a clear definition) on revision from (L)AGB into LRYGB or LSG. Studies were excluded if they reported on less than 10 patients or presented results of revisions to other bariatric procedures. Studies were also excluded if they presented results from revisions of primary RYGB, performed rebanding, if revisional procedure was (primarily) open, or if primary procedure was different than (L)AGB. The same investigators independently searched the list of abstracts according to the search results and selected articles for closer reading. Subsequently, two investigators (CVJ and UKC) extracted the following data, if reported, from the original articles using a preformatted sheet. Variables were operation technique, conversion and reoperation rate, mortality, number and type of early and late morbidity, indication for revision and follow-up in terms of weight loss. Data were retrieved from the articles only. No attempt was made to collect missing data by contacting the authors.

Each of the selected studies was critically appraised by two investigators (CJV and UKC) using a modified form as proposed by the Dutch Cochrane Collaboration. Duration of the data collection; study design (randomisation, prospective, or retrospective consecutive data collection); comparability of study groups; adequate follow-up; commercial interest and description of statistical analysis; and different technique (learning curve) and different kind of treatment were assessed. In case of retrospective analysis of data from a prospective consecutive collected database, the study was qualified as being prospective. Final inclusion was done after consensus was reached. Discrepancies in judgment, if any, were resolved by discussion between the investigators (CJV, UKC, and SML) in a consensus meeting.

Primary outcome measurement is safety of revision from (L)AGB into RYGB or LSG. Items scored included: one- or two-step revisions, number of reoperations and complications (anastomotic leakage, staple line leakage, wound infection, bleeding, perforation, and intra-abdominal abscess), conversions and duration of hospital stay Secondary outcome measurements were reason for revisional surgery in the first place and weight loss (in percentage excessive weight loss (EWL)).

Results

Included Studies

Nine hundred fifteen publications were identified. Seven hundred seventy-eight contained the search terms in a different context and were therefore deemed irrelevant. In total, 137 abstracts were selected for closer reading. One hundred eleven articles were excluded based on the abstract. Of the remaining articles, two were not written in English and were therefore discarded. Twenty-four articles remained. With snowballing, one extra article was identified.

These 25 articles were scrutinized and mined for data. Three articles had different primary procedures (e.g., vertical banded gastroplasty and sleeve gastrectomy), but more than 10 revisions of (L)AGB into RYGB were reported and for that reason the articles were included. In two articles, all data from different primary procedure were combined and for that reason, the articles were excluded (Fig. 1, flowchart). Finally, a total of 15 articles about revision into LRYGB and 8 articles about revision into LSG were used in this review. Table 1 (supplemental) contains the included studies and rates their level of quality.
https://static-content.springer.com/image/art%3A10.1007%2Fs11695-013-1058-0/MediaObjects/11695_2013_1058_Fig1_HTML.gif
Fig. 1

Flowchart of a systematic review about the complications and safety of conversion from laparoscopic gastric band to RYGB or sleeve gastrectomy

Table 1

Quality rate of included articles

Publication

Description of study

Follow-up median ± SD (range)

Quality points

Total

1

2

3

4

5

6

7

8

9

10

 

Abu-Gazala and Keidara [32]

Prospective consecutive series

14.6 ± 9.7

1

0

1

0

0

1

1

1

1

1

7

Ardestani et al.a [11]

Prospective consecutive series

44.4 ± 9.4

1

1

1

0

0

1

0

1

1

1

7

Hamza et al.b [26]

Retrospective consecutive studies

12.9 ± 7.9

1

0

1

0

1

1

1

1

0

1

7

Hii et al. [8]

Prospective consecutive series

36

1

1

1

0

0

0

1

1

1

1

7

Khoursheed et al.b [28]

Retrospective consecutive series

15.83(1–48)

1

0

1

0

0

1

1

1

0

1

6

Langer et al. [19]

Prospective consecutibe series

12

1

1

1

1

0

1

1

1

1

1

9

Mognol et al [20]

Prospective consecutive series

7.3 (3–18)

1

0

1

1

1

1

1

1

1

1

9

Moore et al. [17]

Prospective consecutive series

18

1

1

1

0

1

1

1

1

0

1

8

Muller et al.c [21]

Prospective consecutive series

36 (24–60)

1

1

1

0

1

0

1

0

1

1

7

van Nieuwenhove [18], 2 step

Retrospective consecutive series

9 (3–21)

1

1

0

1

1

0

1

1

0

1

7

van Nieuwenhove et al. [18], 1 step

Retrospective consecutive series

10 (4–23)

1

1

0

1

1

0

1

1

0

1

7

Robert et al. [27]

Prospective consecutive series

22 (3–72)

1

1

1

0

1

1

1

1

1

1

9

Spivak et al. [35]

Prospective consecutive series

15.7 (12–26)

1

1

1

0

0

1

1

1

1

1

8

Topart et al.d [33]

Retrospective consecutive series

12

1

1

1

0

1

1

1

1

0

1

8

van Wageningene [34]

Prospective consecutive series.

12 ± 12

1

1

1

0

1

0

1

1

1

1

8

Weber et al.c [22]

Prospective consecutive series

12

1

1

1

0

0

0

0

1

1

1

6

Acholonu et al. [36]

Prospective consecutive series

6 (2–24)

1

0

1

0

0

0

1

1

1

1

6

Berende et al. [29]

Prospective consecutive series

13.8 (2–46)

1

1

1

0

0

0

1

0

1

1

6

Dapri et al. [40]

Prospective consecutive series

18.6 ± 14.8 ( 1–59)

1

0

1

0

0

0

1

1

1

1

6

Foletto et al. [30]

Prospective consecutive series

20 (3–36)

1

1

1

0

0

0

1

0

1

1

6

Goitein et al. [37]

Prospective

17 (1–39)

1

0

1

0

0

1

1

1

1

1

7

Himpens et al. [31]

Retrospective

32.6 ± 14.8

1

0

0

0

0

0

1

1

0

1

4

Iannelli et al. [38]

Prospective

13.4 (1–36)

1

0

1

0

0

1

1

1

1

1

7

Jacobs et al. [39]

Prospective

26 (5–40)

1

0

0

0

0

1

1

1

1

1

6

Quality rate

Points: 1

Points: 0

           

Definition of study objectives

Clear

Unclear/no

           

Statistical method described

Yes

No

           

Possible bias in inclusion/exclusion

Not present

Present/unclear

           

Different types of treatment besides the evaluated one

Not present

Present/unclear

           

Different technique in patients from same series (learning curve)

No

Yes/not defined

           

Differences in population of compared groups

No

Yes/not defined

           

Measures of outcome

Defined

Had to be calculated

           

Eventual commercial interest related to techniques and/or certain devices

Devices not cited

Devices cited

           

Prospective data collection

Yes

No

           

More than 10 patients

Yes

No

           

Maximum points 10

             

aRevision to multiple procedures (e.g., gastric sleeve, RYGB, biliopancreatic diversion)

bMultiple primary procedures

cComparing LRYGB and rebanding

dComparing primary and secondary RYGB

eComparing open and laparoscopic revisions

Data synthesis

No randomized clinical trials were found. For this reason, our data collection merely exists on observational studies. Performing a meta-analysis or a pooled analysis on this type of studies is not reliable.

Patients

All patients included in this review met the International Federation for the Surgery of Obesity criteria for morbid obesity before the first bariatric procedure. Most of them still did at the time of revision [25].

Five hundred eighty-eight patients underwent revision of (L)AGB into LRYGB. The number of patients in each study varied between 11 and 85 (Table 2) [26, 27]. Only two (0.003 %) patients underwent revision as emergency operation [17, 28]. All other operations were in an elective setting. In two studies (66 patients), male/female ratio was not mentioned. Of the remaining 522 patients, 83 % was female (Table 2). Age and weight are normally distributed in most studies. The mean age was reported in 14 studies. It varied between 37 and 49 years (range, 18–68; Table 2) [8, 20, 28].
Table 2

Patient demographics and revision procedure

Author

N

Age

Sex (M/F)

BMI before first band

BMI at revision

Months since first operation

Two-step revision

One-step revision

BMI at conversion

Abu-Gazala and Keidar [32]

18

43.7 ± 13.2

10/8

41.6 ± 5.3

 

Ardestani et al. [11]

19

42.3 ± 12.3

2/17

44.1 ± 5.7

27.6 (11–48)

1

18

 

Hamza et al. [26]

11

46.5 ± 8.6

2/9

42.8 ± 6.7

29.7 ± 14.3

2

9

 

Hii et al. [8]

82

49 (18–66)

12/70

47(32–73)

43 (31–70)

46 (9–168)

18

64

 

Khoursheed et al. [28]

36

36.95 ± 8.97 (21–53)

45.15 ± 7.95

2

34

 

Langer et al. [19]

25

43 ± 11

1/24

51.0 ± 8.1

47.6 ± 7.7 (34–70)

53 (17–118)

0

25

 

Mognol et al [20]

70

41 ± 10.14 (21–68)

22/58

44.9 ± 10.8 (26.9–81)

42 ± 18 (7–74)

23

47

 

Moore et al. [17]

26

46

3/23

45

40

29

2

24

 

Muller et al. [21]

30

47.1

41.9

 

van Nieuwenhove et al. [18], 2 steps

14

40 ± 11

4/10

45.0 ± 6.7

43.8 ± 5.8

80 (28–120)

14

0

 

van Nieuwenhove et al. [18], 1 step

23

44 ± 12

5/18

41.9 ± 6.1

41.4 ± 6.7

90 (40–144)

0

23

 

Robert et al. [27]

85

39.3 (20–56)

13/72

47.2 (33–67)

42.9 (27–72)

17

68

 

Spivak et al. [35]

33

43.8 (31–62)

3/30

45.8 (39.9–53; SD, 3.4)

42.8 (33.1–50; SD, 4.4)

28.2 (11–46; SD, 11. 3)

0

33

 

Topart et al. [33]a

58

42.5 ± 9.9

5/53

46.3 ± 7.2

43.2 ± 7.0

46.1 ± 17.4

8

50

 

van Wageningen [34]

26

42.7 ± 8.8

11/15

49.7 ± 9.3

43.8 ± 9.5

50.4 ± 24

1

25

 

Weber et al. [22]

32

46 ± 9.5 (28–60)

9/23

47.8 ± 7.6

42.0 ± 6.7 (30.7–59.3)

42 (17–73)

0

32

 

Total

588

     

88

452

 

Acholonu et al. [36]

15

46.6

3/12

 

34.7 (16–60)

2

13

38.66

Berende et al. [29]

28

39.2 (19–65)

45.3 (35–77)

 

13

15

30.6 (23–61)

Dapri et al. [40]

27

43.6 ± 11.4 (25–66)

10/17

45 ± 8.1 (35–64)

 

51.2 ± 30.1 (22–132)

0

27

39 ± 9.6 (24–61)

Foletto et al. [30]

57

49.9 ± 11.9

20/37

51.2 ± 11.1 (39–85)

 

90.48 ± 57.6

16

41

45.7 ± 10.8 (36–77)

Goitein et al. [37]

46

40 (20–60)

12/34

 

20

26

43.1 (33–57)

Himpens et al. [31]

40

47.2 ± 12.4 (25–66)

17/23

46 ± 8.0 (35–64)

 

60.2 ± 30.1 (22–132)

2

38

43.5 ± 8.1 (25.3–61)

Iannelli et al. [38]

41

42 (19–63)

7/34

53.1 (35.9–63)

 

49.9 (35.9–63)

Jacobs et al. [39]

32

45.5

7/25

45.2 (36.2–59.1)

 

67

0

32

42.69

Total

286

44.5

76/182

   

53

192

 

aEight two-step revisions, data could not be seperately extracted

A total of 8 studies, including 286 patients, reported revision into LSG. All operations were elective. In one study (12 patients), the male/female ratio was not mentioned. Of the remaining 258 patients, 70.5 % was female. Age and weight are normally distributed in most research groups. The mean age was reported in all studies and varied between the 39 and 50 (range, 19–66; Table 2) [2931].

Indication for Revisional Surgery

Indications for revisional surgery are displayed in Table 3 and supplemental Table 4. Some patients had more than one reason for revisional surgery. Five of the 15 studies did not mention the reason for revision [21, 22, 28, 3234]. Main reason for revisional surgery in the LRYGB as well as the LSG groups was insufficient weight loss or weight regain in 62.6 and 63.9 % [8, 11, 1720, 26, 27, 30, 3539].
Table 3

Reason for revision into RYGB or gastric sleeve

Author

N

Insufficient weight loss/weight regain

Food intolerance/vomiting

Pouch dilation

Band erosion

Necrosis of the stomach

Reflux oesophagitis/oesophageale dysmotility

Intra-abdominal abscess

Duodenal fistula

Band slippage

Gastric bypass

388

243

23

65

14

2

39

3

1

24

Total

388/588

243/388 = 62.6 %

23/388 = 5.9 %

65/388 = 16.8 %

14/424 = 3.3 %

2/424 = 0.5 %

39/388 = 10.1 %

3/388 = 0.8 %

1/388 = 0.8 %

24/388 = 6.2 %

Gastric sleeve

191

122

18

2

3

0

4

1

2

28

Total

191/286

122/191 = 63.9 %

18/191 = 9.4 %

2/191 = 1.0 %

3/191 = 1.6 %

0

4/191 = 2.1 %

1/191 = 0.5 %

2/191 = 1.0 %

28/191 = 14.7 %

Table 4

Reason for revisional surgery

Author

N

Insufficient weight loss/weight regain

Food intolerance/vomiting

Pouch dilation

Band erosion

Necrosis of the stomach

Reflux esophagitis/esophageal dysmotility

Intra-abdominal abscess

(Total) obstruction

Band slippage

Duodenal fistula

Abu-Gazala and Keidar [32]

18

 

Ardestani et al. [11]

19

16

0

0

1

0

0

0

1

2

 

Hamza et al. [26]

11

7

0

1

1

0

1

1

0

0

 

Hii et al. [8]

82

42

0

11

7

0

10

1

0

11

 

Khoursheed et al. [28]

36

1

1

 

Langer et al. [19]

25

10

0

15

0

0

0

0

0

0

 

Mognol et al [20]

70

17 + 34

1

15

3

0

0

0

0

0

 

Moore et al. [17]

26

7

9

0

0

1

4

1

0

5

 

Muller et al. [21]

30

 

van Nieuwenhove [18]

37

37

0

0

0

0

0

0

0

0

 

Robert et al. [27]

85

48

3

2

1

0

24

0

0

2

 

Spivak et al. [35]

33

25

10

21

0

0

0

0

0

4

 

Topart et al. [33]

58

 

van Wageningen [34]

26

 

Weber et al. [22]

32

 

Total

388/588

243/388 = 62.6 %

23/388 = 5.9 %

65/388 = 16.8 %

14/424 = 3.3 %

2/424 = 0.5 %

39/388 = 10.1 %

3/388 = 0.8 %

1/388 = 0.3 %

24/388 = 6.2 %

 

Acholonu et al. [36]

15

10

0

0

0

0

4

0

 

6

1

Berende et al. [29]

28

Dapri et al. [40]

27

 

Foletto et al. [30]

57

27

4

1

2

0

0

0

 

17

0

Goitein et al. [37]

46

23

13

0

0

0

0

0

 

0

0

Himpens et al. [31]

40

 

Iannelli et al. [38]

41

41

0

0

0

0

0

0

 

0

0

Jacobs et al. [39]

32

21

1

1

1

0

0

1

 

5

1

Total

191/286

122/191 = 63.9 %

18/191 = 9.4 %

2/191 = 1.0 %

3/191 = 1.6 %

0

4/191 = 2.1 %

1/191 = 0.5 %

 

28/191 = 14.7 %

2/191 = 1.0 %

Sometimes, there are more than one reason for revision

One- or Two-Step Revision

In the gastric bypass group, 88 patients (16.3 %) were operated upon in two steps, compared to 53 patients (21.6 %) who underwent sleeve gastrectomy. The reason for two-step revision was referral from another institution after band removal, pouch dilation, or poor tissue quality at band removal as judged by the individual surgeon [20, 26, 32]. Only three articles mentioned the interval between band removal and second stage LRYGB. This interval ranged from 2 to 102 months [18, 37]. Table 5 is added as supplement for details about operation technique.
Table 5

Operative techniques

Article

Capsule resection

Type band

Pouch size (ml)

Stapler

GJ technique

GJ technique

Pouch creation above or below scar tissue

Route of the Roux limb

Size alimentary limb (cm)

Abu-Gazala and Keidara [32]

25

Linear

Double layered

Stapled and hand-sewn

Above or below

Antecolic

150

Ardestani et al. [11]

-

Inamed

-

Linear

Stapled

Antecolic

100

Hamza et al. [26]

30

No

Single

Hand-sewn

Antecolic

75–150

Hii et al. [8]

Yes

 

20–30

Sometimes

Stapled and hand-sewn

Below

100–150

Langer et al. [19]

Circular and linear

Stapled

Khoursheed et al. [28]

No

25–30

Linear

Single

Stapled and hand-sewn

Below

Antecolic

100–150

Mognol et al [20]

Yes

Circular

Single

Stapled

Above or below

Antecolic

100

Moore et al. [17]

Muller et al.c [21]

No

25

Circular

Single

Stapled

Above or below

Antecolic

150

van Nieuwenhove [18]

Yes

No

Single

Hand sewn

Antecolic

100–150

Robert et al. [27]

Yes

30

Linear

Single

Stapled

Antecolic

150

Spivak et al. [35]

No

Inamed

Circular

Single

Stapled

Below

Antecolic

Topart et al. [33]

No

Linear

Single

Stapled

Below

Antecolic

150

van Wageningen [34]

No

Inamed

50

Below

Antecolic

75–25

Weber et al. [22]

25

Circular

Single

Stapled

Above

Antecolic

150

Primary Outcomes

Mortality and Morbidity

Thirteen studies on LRYGB and eight on LSG reported perioperative and long-term mortality rate. The perioperative mortality rate was zero in the LRYGB studies. Hii et al. describes one patient who died 60 days postoperatively (unrelated to surgery; Table 6) [8]. The LSG group tells about three deaths. Foletto et al. is the only study which describes mortality [30]. Three patients died during the perioperative period. One patient died of multi-organ failure because of septic shock. The second patient, suffering from chronic obstructive pulmonary disease, died after 6 months of sepsis after an uneventful discharge at home. The third patient died of pulmonary embolism 24 months after revisional LSG.
Table 6

Operative characteristics and short-term complications

Author

N

Conversions

Operating time, min.

Reoperations

Wound infection

Intra-abdominal abscess

Staple line leak

Bleeding

Acute gastric outlet obstruction (ileus)

Subphrenich hematoma

Fever

Death

Hospital stay, days

Anastomotic leakage

Bleeding/spleen laceration

Herniation port site

Pneumonia

Perforation of stomach

Acholonu et al. [36]

15

0

120

1

0

0

1

0

1

0

0

0

5.5

     

Berende et al. [29]

28

0

99 (54–221)

2

0

0

5

4

0

0

0

0

3 (2–38)

     

Dapri et al. [40]

27

0

120.6 ± 32.4 (65–195)

1

0

0

0

0

0

1

0

0

3.2 ± 1.4 (2–8)

     

Foletto et al. [30]

57

3

120 (90–180)

1

0

0

3

3

1

0

0

3 (>30 dagen)

     

Goitein et al. [37]

46

2

118 (70–250)

2

0

0

2

1

0

0

0

0

3 (1–100)

     

Himpens et al. [31]

40

0

95.6 ± 30.4 (35–195)

2

0

0

3

0

1

1

2

0

4.2 ± 1.2 (2–77)

     

Iannelli et al. [38]

41

0

117 (90–165)

1

0

3

1

0

1

0

0

0

7.9 (5–71)

     

Jacobs et al. [39]

32

0

0

0

0

1

0

0

0

0

0

1.5 (1–3)

     

Total

286

5

 

10

0

3

16

8

4

2

2

3

4.1

     

Total percentage

100 %

1.7 %

 

3.5 %

0 %

1.0 %

5.6 %

2.8 %

1.4 %

0.7 %

0.7 %

1.0 %

4.1

     

Abu-Gazala and Keidar [32]

18

0

195 ± 59

0

0

0

    

0

0

3.9 ± 1.5

0

1

0

0

0

Ardestani et al. [11]

19

0

    

Hamza et al. [26]

11

0

165 ± 75.8

1

0

0

    

0

0

2.0 ± 1.3

0

1

0

0

0

Hii et al. [8]

82

3

132 (70–236)

10

11

0

    

0

0

4.5 (3–39)

1

1

0

0

0

Khoursheed et al. [28]

36

1

145 ± 35 (120–240)

4

0

0

    

0

0

3.36 ±1.2 (3–10)

0

1

1

0

0

Langer et al. [19]

25

0

219 ± 52 (135–375)

0

0

0

    

0

0

5 (4–20)

0

0

1

0

0

Mognol et al [20]

70

3

240 ± 40 (210–280)

4

3

0

    

3

0

7.2

0

3

0

1

0

Moore et al. [17]

26

0

160 (140–195)

1

0

0

    

0

0

3

1

0

0

0

0

Muller et al. [21]

30

0

2

    

van Nieuwenhove et al. [18], 2 steps

14

0

116 ± 31

0

0

0

    

0

0

3 (2–7)

0

0

0

0

0

van Nieuwenhove et al. [18], 1 step

23

0

150 ± 39

1

0

0

    

0

0

3 (3–8)

0

0

0

0

0

Robert et al. [27]

85

2

166 (110–360)

3

0

2

    

1

0

5.2 (3–35)

0

0

0

0

0

Spivak et al. [35]

33

0

105 (85–175)

2

0

0

    

0

0

2.8 (1–10)

0

1

0

0

0

Topart et al. [33]

58

1 (2 open)

128 ± 25.9

3

3

0

    

0

0

7.7 ± 2.8

0

0

0

0

0

van Wageningen [34]

26

4

194 ± 45

2

0

1

    

0

0

7.9 ± 5.6

2

2

0

0

1

Weber et al. [22]

32

0

215 ± 62.7 (135–380)

4

2

1

    

0

0

8.9 ± 4.9(4–24)

1

0

0

0

0

Total

588

14/588 = 2.4 %

167

37

19

4

    

4

0

5.3

5

10

2

1

1

Total percentage

569 = 100 %

 

167

6.5 %

3.3 %

0.7 %

    

0.7 %

0

5.3

0.9 %

1.8 %

0.4 %

0.2 %

0.2 %

The amount of complications varied according to the study cited. Considering LRYGB, the short- and long-term complication rate varied between 3.0 and 29.3 % [8, 35]. Altogether, a total of 45 perioperative complications occurred (8.5 %) and ranged from minor complications as wound infection to major complications as bleeding and perforation (Table 6). The most common short-term complication was wound infection (3.5 %). Anastomotic leakage and bleeding (bleedings combined with splenic injuries) appeared in 0.9 and 1.8 % of the operations, respectively. Three patients needed reoperation for anastomotic leakage, accountable for 7.3 % of reoperations. Bleeding was responsible for 17.1 % of reoperations [8, 1720, 22, 2628, 3235].

A total of 35 (12.2 %) perioperative complications occurred in the patients who underwent LSG. Most were minor but staple line leakage had an incidence of 5.6 %, thereby being the most frequent complication. Almost all complications could be handled conservatively [30, 31, 3640]. Three patients needed reoperation because of staple line leakage, accounting for 30 % of reoperations [29, 31, 3638].

Long-term complications are those who occur later than 30 days postoperative. Details are displayed in supplemental Table 7. Ten of the 15 articles that comprised RYGB revision scored long-term complications. In a total of 478 patients, 42 (8.9 %) complications were seen; some patients developed more than one. Stenosis at the gastrojejunostomy was by far the most common complication at 6.5 %. Marginal ulceration followed at 1.0 % [8, 1822, 27, 3335]. Only three studies about LSG mentioned long-term complications. Just one occurred; this was an internal herniation which required reoperation [31].
Table 7

Long-term complications

Author

Total of patients

Gastro-gastric fistulae

Stenosis at the gastro-jejunostomy/small bowel obstruction

Marginal ulcer

Port-site hernia

Internal hernia

Chronic anemia

Abu-Gazala and Keidar [32]

 

 

Ardestani et al. [11]

 

 

Hamza et al. [26]

 

 

Hii et al. [8]

82

0

22

2

0

0

 

Khoursheed et al. [28]

 

 

Langer et al. [19]

25

1

1

0

1

0

 

Mognol et al [20]

70

0

3

3

0

0

 

Moore et al. [17]

 

 

Muller et al. [21]

30

0

1

0

0

0

 

van Nieuwenhove [18]

37

0

3

0

0

0

 

Robert et al. [27]

85

0

1

0

2

0

 

Spivak et al. [35]

55

0

0

0

0

1

 

Topart et al. [33]

58

0

0

0

1

0

 

van Wageningen [34]

26

0

1

0

0

0

 

Weber et al. [22]

32

0

2

0

0

0

 

Total

478

1

31

5

4

1

 

Total percentage

100 %

0.2 %

6.4 %

1.0 %

0.8 %

0.2 %

 

Acholonu et al. [36]

 

Berende et al. [29]

 

Dapri et al. [40]

 

Foletto et al. [30]

 

Goitein et al. [37]

46

1 (2× PC)

Himpens et al. [31]

40

0

0

0

0

1

0

Iannelli et al. [38]

 

Jacobs et al. [39]

 

Total

40/86

0

0

0

0

1

1

Total percentage

100 %

0 %

0 %

0 %

0 %

2.5 %/1.2 %

2.5 %/1.2 %

PC packed cells

Median length of stay was reported in 13 studies and varied from 1 to 39 days with a mean of 5.3. This included patients with and without complications. Readmission was not mentioned in all studies [27, 35].

Conversion Rate

All articles mentioned the rate of conversion to open surgery. Conversion to open surgery was performed in 14 (2.4 %) LRYGB patients; and in 5 (1.7 %) LSG patients, most common reason was (expected) adhesions and large incisional herniations [30, 37].

Reoperation

The mean incidence of reoperation was respectively 6.5 and 3.5 % [29, 31, 38]. Besides already mentioned bleeding and staple line leakage, intestinal obstruction, stenosis, and internal herniations were reasons also for reoperation (Table 8 and 9) [8, 17, 18, 20, 22, 2628, 3335]. In only two studies, none of the patients who underwent LRYGB required reoperation [19, 32].
Table 8

Reasons for reoperation

Author

N

Number of reoperations

Perforation

Anastomotic leakage

Bleeding

Wound infection

Small bowel obstruction/stenosis/hiatial hernia

Intra-abdominal abscess

Atelectasis

Fistulae

Fever

Staple line leak

Subphrenic hematoma

Abu-Gazala and Keidar [32]

18

0

0

0

0

0

0

0

0

0

0

  

Ardestani et al. [11]

19

 

Hamza et al. [26]

11

1

0

0

1

0

0

0

0

0

0

  

Hii et al. [8]

82

10

0

0

1

3

6

0

0

0

0

  

Khoursheed et al. [28]

36

4

1

0

2

0

1

0

0

0

0

  

Langer et al. [19]

25

0

0

0

0

0

0

0

0

0

0

  

Mognol et al [20]

70

4

0

0

1

0

0

0

0

0

3

  

Moore et al. [17]

26

1

0

0

0

0

1

0

0

0

0

  

Muller et al. [21]

30

2

0

0

0

0

2

0

0

0

0

  

van Nieuwenhove [18]

14

0

0

0

0

0

0

0

0

0

0

  

van Nieuwenhove [18]

23

1

0

0

1

0

0

0

0

0

0

  

Robert et al. [27]

85

3

1

0

0

0

1

0

0

1

0

  

Spivak et al. [35]

33

2

0

0

1

0

1

0

0

0

0

  

Topart et al. [33]

58

3

0

0

0

1

1

0

1

0

0

  

van Wageningen [34]

26

2

0

2

0

0

0

0

0

0

0

  

Weber et al. [22]

32

4

0

1

0

0

2

1

0

0

0

  

Total

569

37

2

3

7

4

15

1

1

1

3

  

Total percentage

100 %

6.5 %

0.4 %

0.5 %

1.2 %

0.7 %

2.6 %

0.2 %

0.2 %

0.2 %

0.5 %

  

Acholonu et al. [36]

15

1

0

 

0

0

0

0

0

0

 

1

0

Berende et al. [29]

28

2

0

 

2

0

0

0

0

0

  

0

Dapri et al. [40]

27

1

0

 

0

0

0

0

0

0

  

1

Foletto et al. [30]

57

1

1

 

0

0

0

0

0

0

  

0

Goitein et al. [37]

46

2

0

 

0

0

0

0

0

0

  

0

Himpens et al. [31]

40

2

0

 

0

0

1

0

0

0

  

1

Iannelli et al. [38]

41

1

0

 

0

0

1

0

0

0

  

0

Jacobs et al. [39]

32

0

0

 

0

0

0

0

0

0

  

0

Total

286

10

1

 

2

0

2

0

0

0

  

2

Total percentage

100 %

3.5 %

0.3 %

 

0.7 %

0 %

0.7 %

0 %

0 %

0 %

  

0.7 %

Table 9

Reasons for reoperations difference Sleeve and LRYGB

Author

N

Number of reoperations

Perforation

Staple line leak/anastomotic leakage

Bleeding

Wound infection

Small bowel obstruction/stenosis/hiatial hernia

Intra-abdominal abscess

Atelectasis

Fistulae

Subphrenich hematoma

Fever

Sleeve

286

10

1

3

2

0

2

0

0

0

2

0

Total percentage

100 %

3.5 %

0.3 %

1.0 %

0.7 %

0 %

0.7 %

0 %

0 %

0 %

0.7 %

0 %

LRYGB

569/588

37

2

3

7

4

15

1

1

1

0

3

Total percentage

100 %

6.5 %

0.4 %

0.5 %

1.2 %

0.7 %

2.6 %

0.2 %

0.2 %

0.2 %

0 %

0.5 %

Secondary Outcomes

Follow-Up and Weight Loss

The BMI at revision varied from 40 to 47.6 in the LRYGB group and from 30.6 to 49.9 in the LSG group [17, 19, 29, 38]. The amount of months between the first operation and revision varied between 28 and 90 months (Table 2 and 10) [11, 18, 30].
Table 10

BMI and excessive weight loss

Author

Number of patients

BMI at revision

BMI at follow-up

Excessive weight loss at follow-up (EWL %)

Time of follow-up at weight measurement (months)

BMI at follow-up/lost weight at follow-up

Abu-Gazala and Keidar [32]

18

41.6 ± 5.3

31.8 ± 5.1

52 ± 44.3

14.6 ± 9.7

 

Ardestani et al. [11]

19

53.7 ± 21.9

24

 

Hamza et al. [26]

11

43.9 ± 7.4

34.3 ± 8.1

62.1 ± 24.0

12.9 ± 7.9

 

Hii et al. [8]

82

43 (31–70)

33 (27–54)

74 (16–85)

36

 

Khoursheed et al. [28]

36

45.15 ±7.95

35.23 ± 6.70 (22.94–51.85)

41.19 ± 20.22 (0.00–89.66)

15.83 ± 13.43 (1–48)

 

Langer et al. [19]

25

47.6 ± 7.7 (37–70)

56.9 ± 15.0

12

 

Mognol et al [20]

70

44.9 ± 10.8

32.2 ± 6.3

70.2 ± 21

18

 

Moore et al. [17]

26

40

23.0

18

 

Muller et al. [21]

30

41.9

35.2

36

 

van Nieuwenhove [18]

14

43.8 ± 5.8

35.3 ± 7.55

35.4 ± 13.2

9 (3–31)

 

van Nieuwenhove [18]

23

41.4 ± 6.7

28.7 ± 10.8

26.5 ± 5.5

10 (4–23)

 

Robert et al. [27]

85

42.9

34.8 (22–50)

22 (3–72)

 

Spivak et al. [35]

33

42.8 (33.1–50.0, SD 4.4)

30.7 (22–39 SD 5.3)

15.7 (12–26)

 

Topart et al. [33]

58

43.2 ± 7.0

66.1 ± 26.8

12

 

van Wageningen [34]

26

43.8 ± 9.5

37.4 ± 8.6

51.9 ± 23.9

12

 

Weber et al. [22]

32

42.0 ± 6.7 (30.7–59.3)

31.8

12

 

Total

588

     

Acholonu et al. [36]

15

 

64.2 (46.5–80.1)

6

Berende et al. [29]

28

39.8

 

32.5

Dapri et al. [40]

27

-

 

16.7

-

34.6 ± 8.7 (21–50.4)

Foletto et al. [30]

57

45.7 ± 10.8 (36–77)

 

-

20 (3–36)

39.0 ± 8.5

Goitein et al. [37]

46

43.1 (33–57)

 

48

36

Himpens et al. [31]

40

43.5 ± 8.1 (25.3–61)

 

31

32.6 ± 14.8 (3–83)

36.2 ± 8.7 (25–54)

Iannelli et al. [38]

41

53.1 (35.9–63)

 

42.7 (4–76)

13.4 (1–36)

42.7

Jacobs et al. [39]

32

42.69

 

60 (13.5–120)

26

33.3 (23–50)

Total

286

     

The mean follow-up after revision to LRYGB was reported in all studies and the mean varied between 7.3 and 44.4 months [11, 20]. The results were reported in BMI at follow-up by four studies [21, 22, 27, 35], in EWL at follow-up by four studies [11, 17, 19, 33], and by seven studies in both [8, 18, 20, 26, 28, 32, 34]. The BMI at follow-up varied between 30.7 and 37.4 [34, 35]. Comparing the first BMI with the BMI at follow-up, the difference was 12.7 points in two studies [18, 20]. The time difference of follow-up was 8 months; 10 months follow-up by van Nieuwenhove and 18 months by Mognol. The smallest descent in BMI appeared in Muller et al., a decrease of 6 points, with the longest follow-up of 36 months. Of the studies where EWL were reported, Moore had the least with only 23 % in 18 months of follow-up. Hii reported EWL of 74 % in 36 months (Table 10).

Discussion

This systematic review shows that laparoscopic gastric bypass and laparoscopic gastric sleeve as revisional procedures after gastric banding are relatively safe with a small amount of complications and a very low mortality rate.

The (L)AGB used to be one of the most performed weight loss operations, but long-term results are disappointing. LRYGB and LSG have a better outcome in long-term weight loss and reducing comorbidities [4, 41, 42]. The main reason for revision is insufficient weight loss or weight regain after (L)AGB, but also a wide array of complaints may lead to revision. For this reason, more and more bands are reversed into LRYGB or LSG [8, 20, 33, 40]. In potential, the revisional operation is more difficult than primary LRYGB or LSG due to adhesions around the (L)AGB and often the area around the stomach has been damaged and scarred. In recent years, many study groups have published their results of conversion surgery; however, the numbers of performed procedures remain small.

The present review shows that laparoscopic revisions can be performed safely in around 98 %. The need for conversion was low. However, some operations were primary open because of expected adhesions in the first place. Open procedures were excluded from this review. This can lead to a potential bias about the difficulty of the revision procedure [33]. Nevertheless, results seem similar if compared with primary LRYGB (conversions rates, 0–23 %) [4, 43].

It should be realized that there is always a potential bias in reporting mortality and morbidity since studies with disappointing results might not be published (publication bias). However, mortality was not different between the procedures (one patient died). All studies have a relatively good outcome in morbidity and mortality [43].

In LSG, a staple line leakage percentage of 5.6 % was found, which is higher than reported in primary surgery (0.35–2.40 %) [30, 44]. Remarkably, LRYGB after gastric banding showed a leakage rate of 0.9 %. After primary LRYGB, this number is reported between 0.4 and 5.0 % [4548] and can thus been seen as equal. It is difficult to draw definite conclusions, although the higher leakage rate after LSG may indicate that perhaps the tissue of the stomach (e.g., scar tissue) after banding needs time to recover. Literature concerning one- or two-step revisions is scarce and conclusions cannot be drawn. Future studies should focus on doing a staged approach which may decrease the leak and overall complication rate.

Despite the fact that the rate of (the feared) anastomotic leakage is comparable in revisional and primary LRYGB, the percentage of reoperations after revisional surgery seems higher. The amount of reoperations after revisional surgery was 6.5 %. This is relatively high compared with reoperation rates reported in the literature after primary LRYGB (3.2 %) [43]. The reasons for reoperation were diverse (e.g., small bowel obstruction, stenosis, or hiatal hernia). This extreme variety in reported complications raises questions about the way the complications were collected and scored. However, a reoperation rate of almost 7 % is high and patients should be informed preoperatively accordingly. It is unclear if a staged procedure would decrease these numbers and future studies should focus on this. Furthermore, centralizing such surgery in large centers may further decrease reoperation rates.

Although in most studies the pouch was created below the scar tissue, in some studies it was created above the scar tissue according to the surgeon’s preference and local situation. No studies compared the position of the pouch. Based on the results of this review, the assumption can be made that perioperative decision making, based on the surgeon’s expertise is save. Numbers are too small to define a reliable conclusion about stapled or hand sewn anastomosis and its influence on occurrence of complications. From primary gastric bypass surgery, it is known that hand-sewn anastomosis with absorbable material may reduce the risk of late-onset complications such as the development of marginal ulceration [49, 50].

Mean hospital stay of revision gastric bypass was 5.3 days which is comparable with primary LRYGB between 2 and 7 days [4, 15, 5154]. The mean hospital stay of the revisional LSG is 1 day less (4.1 days) than that of revisional LRYGB.

Revisional surgery after restrictive procedures is effective. All articles that reported on RYGB mentioned weight loss (some in percentage of BMI others as EWL). Eleven reported weight loss after revision in EWL. The mean ranged from 23 to 74 % [8, 17] of the EWL (range, 0–89 %) [28]. To interpret the results of revisional surgery, the time of follow-up and the initial BMI is important for good interpretation. Standardized report systems should be made from variables as BMI and EWL at a standard follow-up time. Although LGS has no malabsorptive component, the effects are promising also compared to primary gastric band [5557]. In this review, the EWL and BMI after revisional LSG are almost comparable to those after revisional RYGB. A randomized controlled trial may answer the question about the different types of revisional surgery and its successes. Another important issue to pay attention to is that possibly there is a difference (and thus a difference in weight loss after revision) between those patients whose indication for revision was weight regain or inadequate weight loss compared with those whose indication for revision was due to a complication of the band-like erosion or band slippage. This would give a better idea of what the expected weight loss would be after a revision.

The current review has its weaknesses. All studies included are cohort studies, both retrospective and prospective consecutive series. This is the best evidence available at the moment. Additional work needs to be done to unravel the indications and options for revision. This review only comprises the revision from (L)AGB to LRYGB and LSG. More options of revisional surgery are possible such as minigastric bypass, biliopancreatic diversion, and duodenal switch. It remains unclear if there are differences between those procedures.

Revisions are becoming increasingly common and numbers will dramatically increase. There are no clear data to identify the risk of revision, what the best procedures will be, how patients should be screened, and how the follow-up should be performed. Randomized controlled trials would be favored to answer some of these questions in the near future, but since this can be difficult to realize, prospective cohort studies with a large number of patients could be valuable as well.

Conclusion

The present review shows that laparoscopic surgical revision of a gastric band into gastric bypass or gastric sleeve is relatively save, although reoperation rate is higher. Both have good outcomes. RYGB seems to have better long-term results regarding at weight loss. However, the articles published about gastric sleeve as a revision procedure lacks reliable information about the long-term complications and benefits. The decision to perform revision and type of procedure should be based on individual patient characteristics.

Copyright information

© Springer Science+Business Media New York 2013