Is Revisional Gastric Bypass as Effective as Primary Gastric Bypass for Weight Loss and Improvement of Comorbidities?

Revisional gastric bypass (R-RYGB) surgery is utilized for the management of inadequate weight loss or weight regain observed after some cases of bariatric surgeries. Data on the mid-term effectiveness of primary gastric bypass (P-RYGB) compared with R-RYGB (e.g., post sleeve gastrectomy/gastric banding) are controversial. Retrospective chart review of all patients who received P-RYGB and R-RYGB (January 2011–June 2015) at our center. One hundred twenty patients who underwent P-RYGB and 34 R-RYGB who completed 18 months follow-up were included. We compared the effectiveness of P-RYGB with R-RYGB by assessing four anthropometric, two glycemic, and four lipid parameters, as well as the control of type 2 diabetes (T2DM), hypertension, dyslipidemia (remission, improvement, persistence, relapse, de novo), mortality and complications rates. A comparison of the effectiveness of P-RYGB with R-RYGB at 18 months revealed no significant differences in patients’ age, gender, and preoperative BMI between groups. However, patients who received P-RYGB had lower mean weight (P = 0.001) and BMI (P < 0.001), reflected by a higher mean delta BMI (P = 0.02), total weight loss percentage (TWL%) (P < 0.0001) and excess weight loss percentage (EWL%) (P < 0.0001). No differences in glycemic parameters, lipid profiles, control of T2DM, hypertension, and dyslipidemia were observed. No death is reported and complication rates were comparable. Although R-RYGB effectively addressed inadequate weight loss, weight regain, and recurrence of comorbidities after restrictive bariatric surgery, R-RYGB resulted in inferior weight loss compared with P-RYGB. Neither procedure differed in their clinical control of T2DM, hypertension, and dyslipidemia. Both procedures exhibited comparable complication rates.


Introduction
Restrictive bariatric procedures, such as laparoscopic adjustable gastric band (LAGB) and laparoscopic sleeve gastrectomy (LSG), are technically simple, have a low surgical risk [1,2] and are effective in achieving weight loss and managing obesity-related comorbidities, e.g., hypertension (HTN), type 2 diabetes (T2DM) and hyperlipidemia [3,4]. Despite these benefits, surgical revision is performed in 20-60% of LABG patients due to insufficient weight loss or surgical complications and in 5.7% of LSG patients due to weight regain and a relapse of comorbidities [5,6].
Revision of LAGB and LSG to Roux-en-Y gastric bypass (RYGB) is a reasonable approach when conservative treatments fail [7,8]. The restrictive and potentially malabsorptive mechanisms of RYGB, as well as its physiological and metabolic effects resulting from the changes in gastrointestinal hormones collectively contribute to excellent weight loss and metabolic disease resolution [9]. RYGB is also less complicated than biliopancreatic diversion and duodenal switching, carries less risk of nutritional deficiencies, and is more effective than re-sleeving procedures [10,11]. Nevertheless, the results of studies that compared the effectiveness of primary gastric bypass (P-RYGB) with revisional gastric bypass (R-RYGB) remain inconclusive.
Most studies that assessed the effectiveness of P-RYGB and R-RYGB only examined the weight loss and surgical complications [12][13][14]. Moreover, these studies reported inconsistent findings, suggesting comparable weight loss and complications for patients who underwent P-RYGB and R-RYGB [12][13][14] or inferior weight loss and higher complications in patients who received R-RYGB [15,16]. Furthermore, few studies have assessed the evolution of comorbidities after P-RYGB and R-RYGB [8,17,18], despite the possibility of relapse of comorbidities after RYGB (e.g., T2DM, HTN, and dyslipidemia). Likewise, studies comparing P-RYGB to R-RYGB only reported the remission of the comorbidities, with no data provided on the improvements or relapses [19][20][21]. Additionally, some studies did not provide explicit definitions of a remission, improvement, or relapse [16]. Given this range of inconsistencies and deficiencies, an understanding of the effectiveness of R-RYGB will assist physicians and patients in making informed decisions as whether to proceed with R-RYGB after a failed restrictive surgery, particularly when R-RYGB is performed to treat the relapse or persistence of comorbidities rather than the correction of surgical complications [20].
Therefore, the current study assessed the effectiveness of P-RYGB and R-RYGB at 18 months by evaluating weight loss and the evolution (status and control) of three comorbidities (HTN, T2DM, and dyslipidemia) using 4 anthropometric parameters, 2 glycemic parameters, and 4 lipid parameters. The four specific objectives were to assess changes in: & Anthropometric parameters: weight, body mass index (BMI), delta BMI, excess weight loss percentage (EWL %), and total weight loss percentage (TWL %); & Glycemic parameters: glycosylated hemoglobin A1c (HbA1c) and fasting blood glucose (FBG) levels; & Lipid parameters: total cholesterol (TC), low-density lipoprotein (LDL), high-density lipoprotein (HDL), and triglyceride (TG) levels; and, & The status and control of comorbidities: remission, improvement, persistence, or relapse.
The study also assessed and compared the early and late mortality and complication rates of patients who underwent P-RYGB or R-RYGB.

Study Design and Ethics
The Medical Research Center at HMC approved this retrospective study (IRB #16181/16) that was conducted at the Bariatric and Metabolic Surgery Center, Hamad Medical Corporation (HMC) in Doha, Qatar (January 2011-June 2015).

Definitions of the Status and Control of Comorbidities
This study used the standardized American Society for Metabolic & Bariatric Surgery (ASMBS) definitions of the evolution of obesity-related comorbidities after bariatric surgery [21] (Table 1).

Participants, Procedures, and Data Collection
Eligible patients were individuals aged 18-60 who had undergone P-RYGB or R-RYGB during the study period. The inclusion criteria were patients with a BMI ≥ 35 and comorbidities (e.g., HTN, T2DM, and dyslipidemia). The exclusion criteria were patients receiving steroid therapy who had undergone > 1 revisional surgery or had received R-RYGB for the correction of surgical complications (e.g., stricture or band slippage). As shown in Fig. 1, 246 patients who underwent RYGB were identified from the database, but 30 were excluded because they met the exclusion criteria. Thus, patients who received RYGB were eligible, of which another 62 patients were excluded due to incomplete follow-up data. The remaining 154 patients who received RYGB had complete follow-up data at 18 months and were subsequently included in the analysis (120 patients who received P-RYGB and 34 patients who received R-RYGB).
Data were retrieved from medical charts and electronic records of patients who received P-RYGB or R-RYGB. The information included anthropometric (weight, BMI, EWL, EWL%, TWL%, and delta BMI), glycemic (FBG and HbA1c levels), and lipid (TC, LDL, HDL, and TG levels) parameters at baseline and 18 months. Data also included changes in the status and control of the three comorbidities (T2DM, HTN, and dyslipidemia) at 18 months. TWL%, EWL%, and delta BMI were calculated using the methods described in a previous study [23]. In addition, for patients who underwent R-RYGB, we retrieved information on the indications for R-RYGB, the initial procedure/s performed before the R-RYGB surgery, the percentage of initial procedure/s performed at HMC, and prior anthropometric parameters (e.g., weight prior to the initial procedure that was performed before the revisional surgery). Early and late mortality and complication rates of patients who underwent P-RYGB and R-RYGB were also retrieved, classified according to the Clavien-Dindo classifications, and compared [24].

Operative Techniques
All surgeries were performed laparoscopically. RYGB was performed in a uniform manner by creating a 3-4 cm gastric pouch and using a linear stapled antecolic retrogastric anastomosis with a 100 cm alimentary limb and 50 cm bilio-  Fig. 1 Flow diagram of recruitment of participants. * E.g., erosion, stricture, band slippage, and two prior primary procedures pancreatic limb, which was the standard technique performed at our center during the study period. In cases of conversion of sleeve gastrectomy to RYGB, pouch trimming was performed in selected patients when significant intraoperative pouch dilatation was observed. All patients in whom the gastric band was converted to RYGB at our institution received surgery that was performed in one stage.

Statistical Analysis
Statistical analyses were performed using the statistical packages SPSS 22.0 (SPSS Inc. Chicago, IL) and Epi-info (Centers for Disease Control and Prevention, Atlanta, GA).
All P values presented here are two-tailed, and P values < 0.05 were considered statistically significant. Descriptive statistics summarized the demographic, anthropometric, clinical, biochemical, and other related characteristics of the participants. Continuous data were reported as the means and standard deviations (SD); the remaining results were reported as frequencies and percentages. The primary outcome was to compare changes in anthropometric parameters, glycemic parameters, lipid parameters, and the status and control of comorbidities (remission, improvement, persistence, or relapse) at 18 months, as well as mortality and complication rates between patients who received P-RYGB and R-RYGB. Associations between two or more qualitative variables were assessed using the Chi square (χ 2 ) test or Fisher's exact test, as appropriate. An unpaired t test or Mann-Whitney U test (depending on the normality of the data distribution) was used to compare quantitative data and outcome measures (age, anthropometric, glycemic and lipid parameters, etc.) at baseline and 18 months. In addition, we compared the same outcomes at baseline and 18 months for each individual RYGB type. Table 2 compares the preoperative characteristics of patients who were excluded from the study due to incomplete data (n = 62) with patients with complete data who were included in the study (n = 154). The comparisons were performed for the whole sample and for the P-RYGB and R-RYGB groups individually. The 62 excluded patients were not different from the 154 included patients in terms of all baseline characteristics, with two exceptions: patients included in the study were significantly older than the excluded patients (P < 0.0001), and the P-RYGB group generally comprised more females (P = 0.029). Table 3 shows the preoperative characteristics of patients stratified by gastric bypass type. No significant differences in the preoperative characteristics of patients who received P-RYGB and R-RYGB were observed for most of the parameters examined. However, the mean HDL level was significantly lower in patients who underwent P-RYGB (P = 0.037). In addition, more patients were diagnosed with diabetes in the P-RYGB group compared with the R-RYGB group (n = 78 vs. 15 patients, P = 0.028). A significantly higher percentage of patients were diagnosed with hypothyroidism in the R-RYGB group. Table 4 depicts the baseline characteristics of patients who underwent R-RYGB prior to the initial (LSG or LAGB) procedure. Patients' mean baseline weight was 128.07 ± 46.75 kg, EWL% was 55.66 ± 32.14%, and minimal weight achieved after the initial procedure was 63.33 ± 46.76 kg. After the initial procedure, the mean weight gain was 34.39 ± 29.27 kg. Table 4 also shows the type of initial procedure performed prior to R-RYGB (22 patients received LSG and 12 patients received LAGB). Forty-four percent of these initial procedures were performed at our institution (data not presented). Table 5 illustrates the indications for R-RYGB, where 35.2% of patients underwent R-RYGB because of weight regain, 32.3% for inadequate weight loss, and 32.3% to treat GERD. Approximately, 73% had > 1 indication for R-RYGB, e.g., the persistence or relapse of T2DM or HTN, in addition to the indications listed above. Table 6 compares the outcomes of P-RYGB with R-RYGB at 18 months. In terms of anthropometric parameters, patients who underwent P-RYGB had a significantly lower mean weight, BMI, and minimal weight compared with patients who underwent R-RYGB (P = 0.001, P = 0.0001, and P = 0.004, respectively). In addition, patients who received P-RYGB had a greater TWL%, EWL%, and delta BMI reduction than patients who received R-RYGB (P < 0.0001, P < 0.0001, and P = 0.002, respectively). No differences is observed in glycemic or lipid profile between P-RYGB and R-RYGB patients at 18 months. Table 6 further compares the outcomes between the baseline and follow-up at 18 months for each individual RYGB type. Regarding the anthropometric parameters, patients who underwent P-RYGB exhibited a significantly lower weight and BMI at 18 months compared with baseline (P < 0.0001 for both), where delta BMI was − 14.29 ± 5.77 kg/m 2 , EWL% was 69.99 ± 23.61%, and minimum weight was 81.87 ± 16.84 kg. Patients who underwent R-RYGB similarly achieved a significantly lower weight and BMI at 18 months compared with baseline values (P < 0.0001), where the delta BMI was − 11.36 ± 7.41 kg/m 2 , EWL% was 50.62 ± 24.47%, and minimum weight was 94.32 ± 26.59 kg. Regarding the glycemic and lipid parameters, an analysis of the glycemic parameters of patients who underwent P-RYGB showed significantly lower FBG and HbA1c levels (P = 0.0001 for both), and the lipid profile indicated lower TC, LDL, and TG levels (P < 0.0001 for each) but significantly higher HDL levels (P = 0.006) at 18 months compared with the baseline. In contrast, patients who underwent R-RYGB displayed a significant reduction in HbA1c levels (P = 0.007), but no significant improvements in the FBG level or lipid profile at 18 months compared with the baseline. Table 7 depicts the changes in the medication use and the disease status of three comorbidities at 18 months. Regarding medication use, the P-RYGB and R-RYGB groups did not significantly differ in the percentages of patients who stopped, reduced, maintained the same dose, or resumed their medications. Regarding the disease status, again, the P-RYGB and R-RYGB groups did not differ significantly in the percentages of patients who achieved complete remission, improvement, persistence, or relapse of their T2DM, HTN, or dyslipidemia. Two patients who underwent P-RYGB and 1 patient who underwent R-RYGB developed de novo dyslipidemia. Table 8 shows the early and late complications of patients stratified by RYGB type. Early and late mortality were not observed both in the P-RYGB and in the R-RYGB groups. Furthermore, the P-RYGB and R-RYGB groups did not display significant differences in the overall, early or late complication rates.

Discussion
The rapidly increasing number of bariatric procedures performed is likely to be accompanied by increasing numbers of revisional procedures when other treatment modalities fail. At our bariatric center, RYGB accounted for 9.7% (246/2529) of all bariatric surgeries performed during the study period. In the current study, weight regain was the most frequent indication for R-RYGB, consistent with other authors who reported that 46.15% of their R-RYGB surgeries were due to weight regain [25]. Weight regain is attributed to anatomical, physiological, psychosocial, and behavioral (nonadherence to healthy lifestyle) factors [22]. Likewise, 41.02% of our patients underwent R-RYGB due to inadequate weight loss, consistent with another study in which 32.3% of revisions were performed due to inadequate weight loss [26]. A total of 32.3% of patients in the current study underwent R-RYGB as a treatment for GERD, supporting the findings of other studies in which 11 of 81 patients received R-RYGB for GERD [17]. GERD is a common long-term sequela of LAGB and LSG [26,27], and RYGB is an effective procedure when other modalities fail [27]. Notably, 26.4% of our patients underwent R-RYGB for one indication and 73.5% for > 1 indication. Documenting the number of indications for  revisional bariatric surgery is important, as indications for revisional surgery usually do not exist in isolation and are interrelated, e.g., weight regain after primary surgery that eventually leads to the recurrence of HTN and T2DM. In terms of anthropometric parameters, some studies have not observed differences in EWL% between patients who received P-RYGB and R-RYGB [12,13]. Conversely, patients who underwent P-RYGB have been reported to achieve significantly better EWL% [16,18], consistent with the findings of the current study. Likewise, the patients who received P-RYGB in our study had a significantly higher TWL% than patients who received R-RYGB, similar to the findings from other studies [20,28]. In the current study, patients who underwent P-RYGB achieved a significantly lower mean BMI and higher delta BMI, consistent with other studies [17,18]. The inferior weight loss we observed in the R-RYGB group is likely attributed to complex factors but we could speculate that the characteristics of the patients and/or surgical technique are likely the main factors. Regarding patient characteristics, patients who undergo R-RYGB are likely to have advanced/ refractory obesity due to genetic causes [8], and lifestyle/psychological factors after repeated surgery also potentially reduce the efficacy of R-RYGB [29,30]. In addition, the reduced efficacy of R-RYGB might be attributed to changes in bile flow and the modulation of enteric hormone levels. This is based on evidence suggesting that patients who achieved a sustained weight loss after RYGB had higher leptin, glucose-dependent insulinotropic polypeptide, and glucagon-like peptide 1 levels than patients who experienced weight regain [31,32]. Regarding the characteristics of the surgical technique, the inferior weight loss observed in the R-RYGB group may be due to the lack of precise measurements of the stomach pouch size during revisional surgery, resulting in a larger stomach pouch [16]. However, recent study suggests that pouch size does not play a critical role in weight regain unless the pouch is very large [33]. There is also a growing evidence suggesting that the differences in the total alimentary limb length (TALL) and the length of the biliopancreatic (BPL) could play a role [34,35]. At the time of the current study, the standard procedure at our institution was RYGB performed without consideration of total bowel length and TALL (a point seen today as a shortcoming). Future research would benefit from assessing these factors and their effects.
We examined each type of RYGB individually by comparing the parameters at baseline and 18 months. At 18 months, patients who underwent P-RYGB in the present study achieved a BMI and EWL% that were consistent with another study [8]; similarly, patients who underwent R-RYGB achieved a BMI and EWL% comparable to other studies [8,36]. In terms of comorbidities, the T2DM remission rate in the current study was not significantly different between the P-RYGB and R-RYGB groups (53.8% vs. 62.5%), similar to the data reported in a previous study (23.1% vs. 50.1%) [37]. Although other studies have compared P-RYGB with R-RYGB exclusively by measuring T2DM remission [8,20], we also compared T2DM evolution (improvement, persistence, or relapse) between patients who received the two procedures and no significant differences were observed. Regarding HTN, the P-RYGB and R-RYGB groups displayed similar HTN remission rates, consistent with studies employing shorter (1.5 years) and longer (3-5 years) followup durations [8,19,37]. However, another 3-year follow-up study found that patients who underwent P-RYGB experienced a significantly better HTN remission rate than patients who underwent R-RYGB [16]. In terms of the remission of dyslipidemia, remission rates in the P-RYGB group were not different from the R-RYGB group, although our rates were much lower than the values reported in another study [18], which was probably attributed to their longer follow-up duration (3 years) that may have facilitated the resolution of dyslipidemia [18]. Although other studies have assessed the effectiveness of P-RYGB and R-RYGB by evaluating dyslipidemia remission alone [17,20], the current study also compared the effectiveness of the two procedures in improving dyslipidemia, and again no significant differences were observed between groups. Likewise, no previous studies have reported the emergence of de novo dyslipidemia, but the present study observed two de novo cases in the P-RYGB group and one case in the R-RYGB group at 18 months. Collectively, the findings of improvements in comorbidities support the hypotheses that although R-RYGB is associated with inferior weight loss, it successfully manages obesity-   Chi square test used for comparisons associated comorbidities, e.g., T2DM and HTN, and displays the same potential as P-RYGB. We also examined each type of RYGB individually by comparing the resolution of comorbidities (baseline vs. 18 months). Although P-RYGB produced a significant improvement in HbA1c levels, consistent with a previous study [38], it did not improve the lipid profile, in contrast to another study [39]. Patients in the R-RYGB group achieved a significant improvement in HbA1c levels, consistent with a previous study showing that R-RYGB significantly improved HbA1c levels at 1 year after LAGB, but not after LSG [20]. Although R-RYGB improved the FBG levels at 18 months, this difference was not statistically significant, similar to other studies [20]. The probable explanation is our small sample size.
In terms of complications of P-RYGB and R-RYGB, R-RYGB is reported to be a safe procedure, with similar early and late complication rates to P-RYGB, supporting the findings of the current study [37]. No patients died in the present study, whereas other studies reported a low mortality rate in patients who underwent P-RYGB and R-RYGB (0.2% and 1.3%, respectively) [40].
The study has limitations. We included fewer patients in the R-RYGB group than in the P-RYGB group (34 vs. 120), which reflects the observation that revisional surgeries are generally less common than primary surgeries; the inclusion of equal numbers of patients might have provided more precise estimates. Due to the smaller number of patients in the R-RYGB group, we combined patients who underwent the two initial procedures (LSG and LAGB) in the R-RYGB group, and although both LSG and LAGB are restrictive surgeries, LSG has some metabolic component that differentially affects the resolution of comorbidities than the purely restrictive LAGB [41]. A longer follow-up period would have been beneficial in terms of assessments of weight regain because the nadir weight loss is usually achieved after approximately 2-3 years, after which weight regain takes place. A longer followup period would have also allowed an extended assessment of the resolution, improvement, relapse or persistence of comorbidities. We did not compare the effects of the two procedures on the patients' nutritional status and quality of life; this information would have provided a more comprehensive assessment of the effects of each procedure.
The study also has strengths. Studies comparing the effects of P-RYGB and R-RYGB on comorbidities (T2DM, HTN, and dyslipidemia) have only focused on remission [19,20], falling short of assessing the effects of the procedures on the improvement, persistence, and relapse of these comorbidities. The study was also novel in assessing the de novo rates of these comorbidities in patients who underwent each of the two procedures, a point that is usually omitted [18], despite the importance of continued follow-up of de novo cases. Previous studies did not use explicit definitions for remission, improvement, and relapse [16]; we employed international ASMBS definitions for the disease status in the comparison of P-RYGB with R-RYGB. Furthermore, the current study is the first to compare the anthropometric, glycemic, and lipid parameters for each individual type of RYGB at baseline and 18 months and explore the effectiveness of each type of RYGB per se, where no other study performed this type of analysis [8,17]. We also assessed the complications of both procedures for an impartial comparison of P-RYGB with R-RYGB. Finally, due to incomplete follow-up data, 62 patients were excluded from the analysis. Although these excluded patients were significantly younger than the included group, their exclusion might not have affected our findings, as reports advocate that improvements in obesity-associated comorbidities after RYGB are similar across age groups [42].

Conclusions
Based on the findings of the current study, R-RYGB results in an inferior outcome in terms of weight loss. However, R-RYGB is similar to P-RYGB in the clinical effects on the control of T2DM, hypertension, and dyslipidemia (remission, improvement, persistence, relapse, and de novo rates). Therefore, R-RYGB potentially represents an effective approach to address the relapse of comorbidities following restrictive bariatric surgeries (e.g., LAGB & LSG). An understanding of the outcomes of R-RYGB after restrictive bariatric surgeries in terms of its potentially lower weight loss but comparable resolution of comorbidities will assist physicians and patients in making appropriate decisions.