Title: Initial results of reoperative bariatric surgery: a retrospective feasibility and safety study at a low-volume unit

Background This study assessed the feasibility and results of revisional bariatric surgery at a low-volume unit. Methods This retrospective study was conducted from January 2017 to August 2020; the revision group comprised patients treated for weight regain (WR, n = 6), insucient weight loss (IWL, n =3), and various complications (n=6). Clinical characteristics and 30-day outcomes were assessed and compared with those of primary bariatric procedures (control, n=173). Results A total of 8.0% (15/188) of patients underwent revisional procedures and tended to be elderly (40.1 vs. 38.2 years), be female (73.3% vs. 54.9%), and have a signicantly lower body mass index (33.1 vs. 39.9 kg/m 2 ) and fewer comorbidities than controls. Gastric bypass was the most prevalent revisional procedure. A signicantly longer operative duration (155 and 96 min; p < 0.001), longer length of stay (3.7 and 2.4 days) and higher 30-day complication rate (20.0% vs. 4.6%) were found in the revision group, including one case of leakage and another of jejunojejunostomy obstruction. There were no cases of conversion to open surgery or mortality. A total of 5/6 WR patients achieved excessive weight loss >50%, versus only one IWL patient who reached this goal. Complications, including marginal ulcer, stula and post-sleeve gastrectomy stenosis, were alleviated after revision. Conclusion surgery effective practice with WR complications after however, the benets should be weighed against the risks. More robust evidence is required to


Background
Parallel to the global trend of the exponentially increasing prevalence of obesity, bariatric surgeries are being increasingly utilized worldwide [1]. Bariatric surgery is well recognized as an effective solution for sustained weight loss and comorbidity resolution [2]. However, as obesity is a chronic relapsing condition, not all patients achieve long-term satisfactory results. In the United States, revision/conversional procedures account for 15.4% of all procedures, and this proportion continues to increase [3]. Because of the lack of established guidelines or de nite criteria, the American Society for Metabolic and Bariatric Surgery (ASMBS) Revision Task Force has highlighted the importance of careful patient selection and the necessity of surgeon expertise [4]. As literature has continued to emerge, these practices are largely based upon a multidisciplinary approach and are conducted after the exclusion of alterable causes, such as maladaptive lifestyles or eating disorders [5]. However, revisional procedures are generally considered more complex and technically challenging than primary procedures [6]. Although an acceptable safety pro le can be achieved for revisional procedures at accredited centers, the overall rate of complications is at least twice as high [7], and the rate of major complications can be 10 times as high [8]. Consequently, it is recommended that these procedures be carried out at highvolume centers [9]. On the other hand, low-volume centers deal with a nonnegligible proportion of all bariatric procedures [10] and are not immune to these challenges. To date, the literature remains insu cient to support such a plan in lowvolume practices; hence, a retrospective study was conducted at our unit enrolling all patients who underwent revisional surgery on a background of 50 procedures per year. The primary endpoint was the e cacy in terms of additional weight loss and resolution of various complications. The secondary endpoints were regarding perioperative safety, including the operative duration, length of hospital stay, and 30-day morbidity and mortality rates.

Methods
After approval by the Institutional Review Board, we retrospectively reviewed the electronic charts of all consecutive patients who underwent bariatric surgery performed by a single surgeon at a private university-a liated hospital. The requirement for informed consent was waived because no data regarding the cases were disclosed.
From January 2017 to August 2020, the control group included all patients who met the regional criteria and were eligible for a primary bariatric procedure [11]. Patients who underwent revisional surgery were identi ed according to the de nition [4] and strati ed into 3 subgroups: those who suffered from insu cient weight loss (IWL), weight regain (WR) and chronic complications. A body mass index (BMI) of 23.5 kg/m 2 was considered to represent the ideal body weight.
IWL was de ned as < 50% excess weight loss (EWL) at least 24 months postoperatively [12]. WR was de ned as the regain of >50% of lost weight. Patients with a BMI that still ful lled regional criteria at revision were also enrolled [11]. The exclusion criteria were emergency procedures, immediate postoperative complications and poor compliance or nonstabilized psychiatric disorders. The preoperative workup included routine panendoscopy and upper gastrointestinal series with re ux esophagitis according to the Los Angeles classi cation [13]. Patients were approved for surgery by a multidisciplinary assessment after a nutritional survey and exclusion of maladaptive eating problems.

Indications, rationale and surgical details
For WR/IWL patients, the procedures were as follows: (i) After sleeve gastrectomy (SG): As reached by consensus, Rouxen Y gastric bypass (RYGB) was our preferred choice [14]. The technique comprises the construction of a 30-mL gastric pouch, linearly stapled antecolic gastrojejunostomy (GJ), and jejunojejunostomy (JJ) with both alimentary (A)/biliopancreatic (BP) limbs measuring 100 cm. The alternative was one-anastomosis gastric bypass (OAGB) in the case of a higher weight loss demand without pouch dilatation or concomitant re ux [15]. The technique involves transverse crow's foot stapling with the BP limb set at 200 cm. (ii) After OAGB: As a commonly utilized revisional procedure after OAGB [9], modi ed RYGB involves BP limb sectioning proximal to the GJ, gastric pouch resection/trimming and GJ and JJ recreation. Only in this case did the BP limb vary between 180 cm and 200 cm due to the primary procedure. For patients with an MU after OAGB, the technical details involved dividing the GJ from the gastric side, trimming the pouch, and recreating the GJ and JJ with a 100 cm BP/A limb [17]. (vi) Stenosis/ stula after SG: RYGB served as a salvage procedure after failed nonoperative treatment. (vii) Malnutrition after OAGB: RYGB was chosen for this indication with a tailored short limb [15].
Patients were followed for at least 6 months postrevisional procedures. Additional weight loss, complication resolution and clinical characteristics, such as the type of primary/revisional procedure, time lapse, body weight (BW)/BMI at each time point together with the operative duration, length of stay (LOS) and 30-day rate of complications graded according to the Clavien-Dindo (C-D) classi cation [18], were collected and compared with those of primary procedures. Statistical analysis was performed using free R software (R Project for Statistical Computing).

Results
From January 2017 to August 2020, a total of one hundred eighty-eight patients underwent bariatric surgery at our hospital. Among them, 15 patients (8.0%) underwent revisional procedures, all of which were the rst revision. As shown in Table 1 110.3±18.6 kg; p<0.001) than those in the control group. Diabetes mellitus (0 vs. 31.2%; p=0.01) and dyslipidemia (26.7% vs. 61.8%; p=0.008) were signi cantly less prevalent in the revision group.
Another patient who visited the ER 23 days postoperatively was diagnosed with JJ obstruction by a phytobezoar after conversion from OAGB to RYGB. Both patients recovered uneventfully after the rescue operation. A third patient who underwent OAGB after open VBG experienced postoperative bleeding, which was resolved with proper medication. There were no cases of mortality. At the one-year follow-up, the EWL (50.1±24.4% vs. 75.4±19.2%; p=0.006) and percentage of total weight loss (17.6±12.7% vs. 29.6±7.5%; p<0.001) were signi cantly lower in the revision group (n=6) than in the control group (n=129, retention rate 87%).
The details for the revision group are summarized in Table 3. Eight patients (8/15, 53.3%) had initially undergone surgery at another hospital. While SG (6/15, 40%) was the most common initial surgery, the most common revisional surgery was RYGB (13/15, 86.7%). Eight patients presented with concomitant gastroesophageal re ux disease (GERD). In the WR group, 5/6 patients regained > 50% of the lost weight (range: 52.9% to >100%) at the mean follow-up of 62 months; patient 5 had a BMI of 37.8 kg/m 2 at revision and met the criteria. RYGB was performed for all but one patient who was converted to OAGB after SG. The initial mean BMI was 42.2 kg/m 2 and 36.4 kg/m 2 at the time of revision. The mean BMI decreased to 28.6 kg/m 2 after revision, with most (5/6) patients achieving EWL ≥ 50% at the mean follow-up of 14 months. For three patients with IWL, the EWL ranged from 16.1% to 44.1% after the initial operation, and two of them regained >100% of the lost weight after a mean interval of 9 years. The mean BMI was 39.3 kg/m 2 initially and 38.2 kg/m 2 at revision. The mean BMI decreased to 32.8 kg/m 2 at an average of 11 months after revision. Only one patient who underwent conversion from Band+GCP to RYGB achieved acceptable EWL (68.9%). A 54-year-old woman who was converted from open VBG to OAGB achieved merely 39.7% EWL at 9 months postoperatively. Another 36-year-old man who was converted to RYGB after SG failed to achieve further meaningful weight loss at 6 months postoperatively.
Among those with chronic complications, three patients underwent revision for an intractable MU, one after RYGB and the other two after OAGB. No recurrence of ulcer-related symptoms was reported among all of the patients at the mean follow-up of 7 months. Two other patients were converted to RYGB for stenosis after SG; one experienced repeat balloon dilatation failure, and the other had a concomitant stula that failed to resolve after repeat stenting. This particular patient then underwent 4 endoscopic internal drainage procedures after the revisional procedure and achieved complete healing of the stula. Finally, a 64-year-old woman was found to have malnutrition after OAGB (albumin: 2.6 mg/dl) and recovered uneventfully after conversion to RYGB with 75-cm A/50-cm BP limbs.
Out of the 15 patients, four had hypertension. Two (50%) had complete remission, and the other two (50%) had partial remission. Three out of four patients with hyperlipidemia achieved remission after revision, and the fourth patient showed signi cant improvement.

Discussion
IWL, WR and long-term complications remain signi cant challenges despite the validated e cacy of bariatric surgery [5]. With the complex nature of obesity, there are likely many interrelated factors, and not all complications can be attributed to the surgery [19]. Nevertheless, current evidence supports the application of additional procedures if patients present with undesirable results [4]. Our case distribution is consistent with that of the aforementioned systematic review [5]. Indications usually do not exist in isolation, and over half of our patients presented with concomitant re ux esophagitis. Therefore, RYGB was our preferred choice [8,14]. The current study reveals that most patients with WR can achieve acceptable weight loss after revision, but only 1/3 of those with IWL can reach this goal. For various chronic complications, the corresponding treatment effectively solves most problems, including symptoms of GERD.
While revisional surgeries are on the rise and have become the third most common type of bariatric surgery [6], the incidence of revisional bariatric surgery varies and largely depends on the original procedure [20]. Herein, we found a rising but relatively low revision rate of 8.0% compared to 15.4% in the U.S. [3]. This difference could be due to the shortterm follow-up period and the fact that our primary operations consisted mainly of RYGB and OAGB [21,22]. Moreover, as illustrated in a previous report [8], most of our patients underwent their original procedure at another institution. Recently, a meta-analysis revealed an increasing trend of revision being required after SG [23], and these patients accounted for a large part of our revision group. In addition, ve of 120 consecutive patients (4.2%) required revision after OAGB: one patient for malnutrition, two for an intractable MU, and another two for IWL. Though within the scope of earlier large series [22], our revision rate still merits long-term monitoring.
In line with ASMBS data, patients in the revision group tended to be older, have a lower BMI and have a lower incidence of comorbidities than those in the control group [6]. Furthermore, in accordance with the Metabolic and Bariatric Surgery Quality Improvement Program Data Registry, the revision surgeries showed a longer operative duration, longer LOS and higher complication rate [7]. However, perioperative results often cannot be compared with those of preceding studies due to the presence of heterogeneity [24]. The overall complication rate has been reported to be up to 39.3% for revisional surgery and as high as 47.6% for conversion surgery, with a nonnegligible mortality rate of 2% [5]. Our result in terms of the operative duration (155 min vs. 106.7 min) and LOS (3.7 days vs. 1.8 days) is far inferior to that of reports from certi ed bariatric centers [6]. However, as our main revisional procedure was RYGB, we attained no inferior results compared to those of a former case-matched analysis with respect to the LOS (3.7 days vs. 3.8 days), conversion rate (0 vs. 10.8%) and overall complication rate (20% vs. 27%) [25]. In contrast, in a smaller study of 24 cases, Amiki et al reported a serious complication rate of 16.7% and a reoperation rate of 8.3% [26]. Although the result of the present study are still far from ideal, with a leakage rate of 6.7% and no cases of conversion or mortality, the current study supports the feasibility of conducting such a plan in a low-volume setting.
Speci cally, while revision after purely restrictive procedures that involve an additional malabsorption component generally leads to optimal weight loss [27], revisional RYGB yielded acceptable results in the current study. In case 6, who were converted from OAGB to RYGB failed to achieve desirable results. In contrast to a prior method that utilized 50-cm A/150-cm BP limbs without pouch trimming that resulted in a mean EWL of 14% and BMI loss of 3.2 kg/m 2 [28], a more complex approach including pouch trimming and lengthening of both the A (100 cm) and BP (180-200 cm) limbs was adopted. A noticeable reemergence of maladaptive eating behavior was deemed the main culprit for both the signi cant WR after the initial OAGB and the poor response after the revisional procedure.
Remarkably, none of the three patients with IWL reached ≥ 50% EWL after the initial restrictive procedure. As previously reported, revisional surgery for nonresponders is suggested to be less than ideal [29], and only one patient who was converted from Band+GCP to RYGB achieved 68.9% EWL. In case 8, the patient was converted to OAGB after VBG with the consideration that a 200-cm BP limb and a trimmed gastric sleeve would be su cient, and the result was clearly inferior to that of a previous report [16]. Regarding the total bowel length, which was measured to be 8 m in this patient, more evidence is required to guide the application of a tailored limb approach. In case 9, conversion to RYGB after SG was arguably more effective for GERD than for further weight loss [30]. Moreover, applying a GJ diameter of 2.5 cm instead of the standard 2 cm to relieve symptoms of re ux and dysphagia may have also contributed to this unfavorable outcome [31].
In summary, the current study showed a mean EWL of 50.1% among 6 out of 9 patients who underwent revision surgery for WR and IWL and at least 12 months of follow-up, while 4/6 achieved > 50% EWL. Although this result is inferior to that in the control group, since a systematic review of conversional RYGB reported a mean EWL of 46-70% [5], we consider our initial results to be appropriate. Various negative factors, such as primary nonresponsiveness, an unde ned ideal bowel length and the re-emergence of maladaptive eating behavior, need to be carefully addressed via a more comprehensive approach in the future.
While MU was a common late complication after RYGB [32] and OAGB [33], two out of three patients in this study were still recognized as active smokers. Similarly, previous studies in the literature have also noted this problem [34]. It is worth mentioning that there are two major complications in this study, one case of GJ leakage and another of JJ obstruction, compared to no such complications among 136 consecutive primary RYGBs. Since optimal techniques and strategies to minimize complications remain inconclusive [35], we have adopted a more cautious perioperative approach and more vigorous strategies to ensure successful smoking cessation. We also had one patient who reported malnutrition after OAGB. Considering that the index patient was initially treated with a 180-cm BP limb and a 300-cm common limb, more evidence is required to determine the ideal limb length [36].
Last, in accordance with a systematic review, remission of comorbidities, namely, hypertension and hyperlipidemia occurred in 50% of our patients; therefore, revisional surgery should be considered in patients with persistent metabolic disease [37].

Limitations
It is beyond the scope and not our intention to provide a full discussion of indications. Similar to many reports [4,24], the present study had a limited sample size, had a short follow-up period and was retrospective in nature, and these disadvantages hinder the interpretation of our results. However, concerning the lack of reports regarding the e cacy and perioperative safety of revisional bariatric surgery from a low-volume perspective, we believe that the current study can still provide some insight.

Conclusion
In conclusion, our results suggest that revisional procedures can feasibly be conducted in a low-volume setting with proper patient and procedure selection. Although these procedures are less e cient and riskier than primary procedures, most patients with WR can achieve satisfactory weight loss, and most complications can be resolved. Given the small study population and short-term follow-up, robust data are required to support a more standardized and safer practice. Abbreviations WR, weight regain; IWL, insu cient weight loss; ASMBS, American Society for Metabolic and Bariatric Surgery; BMI, body mass index; EWL, excess weight loss; SG, sleeve gastrectomy; RYGB, Roux-en Y gastric bypass; GJ, gastrojejunostomy; JJ, jejunojejunostomy; A, alimentary; BP, biliopancreatic; OAGB, one-anastomosis gastric bypass; Band+GCP, adjustable gastric banding + greater curvature plication; VBG, vertical banded gastroplasty; MU, marginal ulcer; BW, body weight; LOS, length of stay; C-D, Clavien-Dindo; GERD, gastroesophageal re ux disease; Declarations Authors' contributions HC designed the study, performed the surgical procedures, followed the patients, and participated in the data analysis and writing of the manuscript. SM contributed to the data analysis. All authors read and approved the nal version to be published.

Funding
The study was not sponsored and funded by any funding.

Availability of data and materials
The datasets generated and/or analyzed during the current study are not publicly available due to restrictions by the local Institutional Review Board but are available from the corresponding author upon reasonable request and with permission from the local Institutional Review Board.
Ethical approval and consent to participate All procedures performed in studies involving human participants were in accordance with the ethical standards of institutional and/or national research committees and the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards. The research project was approved by the Taipei Medical University-Joint Institutional Review Board. Informed consent was waived by the Taipei Medical University-Joint Institutional Review Board (No.: N202103140) because no data of the cases were disclosed.

Consent for publication
Written informed consent for publication was waived by the Taipei Medical University-Joint Institutional Review Board because no clinical details and/or clinical images of the cases were disclosed.