Efficacy and Safety of One Anastomosis Gastric Bypass Versus Roux-en-Y Gastric Bypass for Obesity: a Meta-analysis and Systematic Review

The inclusion criteria are as follows: (1) randomized controlled trials comparing OAGB and RYGB; (2) include previous English-language articles; (3) the follow-up time must be ≥ 1 month; (4) follow-up content must include one of the following: BMI, percent of excess weight loss (%EWL), or percent of excess BMI loss (%EBMIL).

The exclusion criteria are as follows: (1) articles were withdrawn; (2) articles on animal experiments; (3) full text not.

We searched the PubMed, Embase, and Cochrane Library databases, under the title and abstract contexts using the keywords “one anastomosis gastric bypass,” “mini gastric bypass,” “single anastomosis gastric bypass,” or “omega loop gastric bypass,” and “Roux-en-Y gastric bypass” in pairwise combinations. The search period spanned from the start of the study until July 4, 2022. Two members independently conducted the search process and included articles. A third reviewer was consulted if there was a disagreement concerning the inclusion of an article.

Data was extracted independently by two members and included the following:

If verification is required by a third reviewer, we may choose to contact the author or editor for further information.

Two authors independently performed the quality assessment. The quality of randomized controlled trials was evaluated with the Cochrane collaboration’s tool to assess the risk of selection including random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other biases. The risk of bias was assessed for these seven items. The decision of whether it was “low risk of bias,” “high risk of bias,” or “unclear” was made according to the risk of bias assessment criteria [19].

Endpoint definitions in the literature included BMI [weight/height²] and %EWL = [(initial weight) (postoperative follow-up weight)]/[preoperative overweight weight] × 100%. The target weight was set to be the weight corresponding to a BMI equal to 25 kg/m². %EBMIL = [ΔBMI/(preoperative initial BMI 25)] × 100%.

Statistical analysis was performed using RevMan 5.4.1. The risk ratio (RR) was used as an effect analysis statistic for categorical variables, and the mean difference (MD) was used as an effect analysis statistic for numerical variables. All statistics were calculated using a 95% confidence interval (CI). When P > 0.05 and I² < 50%, the fixed effects model was used, whereas when P ≤ 0.05 or I² ≥ 50%, the random effects model was used.

A total of 1217 articles were initially retrieved through PubMed, Cochrane, and Embase databases. We excluded 743 articles comparing non-OAGB and RYGB procedures, leaving 474. After screening the inclusion and exclusion criteria, 38 articles were retained. Three of them were the same articles withdrawn from the three databases, one article was the same article that did not include follow-up data in the three databases, and there were 24 duplicate articles. Finally, eight randomized controlled trials were included in this study (Fig. 1).

A total of eight prospective randomized controlled trials were included in our meta-analysis [20–27]. The papers were published between 2005 and 2022, and a total of 931 patients participated. The basic characteristics of the included studies are summarized in Table 1, with mean ages ranging from 30.7 to 45 years and mean preoperative BMI ranging from 42.6 to 53.5 kg/m². In the study providing information on metabolic syndrome, there were 288 patients with type 2 diabetes, 321 patients with hypertension, 188 patients with hyperlipidemia, 101 patients with gastroesophageal reflux syndrome, and 133 patients with apnea syndrome. The surgical technique characteristics of the included studies are summarized in Table 2. Figure 2 shows the risk of bias in the included studies. Almost all the articles had risk of bias due to blinding, which was not described in the text.

Since the outcome measures described in each article were inconsistent, we decided to perform a meta-analysis of postoperative weight loss outcomes using BMI, %EWL, or EBMIL. First, the preoperative BMI was analyzed. In the eight included articles, there was low heterogeneity in preoperative BMI between OAGB and RYGB, and the difference was not statistically significant (MD = − 0.08, I² = 0%, 95% CI = − 0.64 to 0.49, P = 0.79) (Fig. 3). Regarding the BMI at 6 months after surgery, four articles were included, and the difference was not statistically significant (MD = − 0.74, I² = 69%, 95% CI = − 2.98 to 1.51, P = 0.52) (Fig. 4). Regarding the %EWL at 6 months after surgery, four articles were included, and the difference was not statistically significant (MD = 2.19, I² = 77%, 95% CI = − 4.14 to 8.51, P = 0.50) (Fig. 5). A total of four articles were included for %EWL at 12 months after surgery, and OAGB was considered superior to RYGB (MD = 3.55, I² = 0%, 95% CI = 0.90 to 6.19, P = 0.009) (Fig. 6). A total of five articles were included for BMI at 12 months after surgery, and it was considered that OAGB had better weight loss effect than RYGB (MD = − 2.03, I² = 87%, 95% CI = − 3.63 to − 0.44, P = 0.01) (Fig. 7). Two articles were included for BMI at 2 years after surgery, and the weight loss effect of OAGB was better than that of RYGB (MD = − 2.97, I² = 21%, 95% CI = − 3.33 to − 2.60, P < 0.00001) (Fig. 8). Two articles were included for the %EBMIL comparison between OAGB and RYGB at 2 years after surgery and there was no statistical significance (MD = 9.89, I² = 96%, 95% CI = − 4.79 to 24.58, P = 0.19) (Fig. 9). Two articles were included for BMI at 5 years after surgery and there was no statistical significance (MD = − 3.02, I² = 96%, 95% CI = − 6.64 to 0.61, P = 0.10) (Fig. 10).

A total of two articles were included for intraoperative complications, and there was no statistical significance between OAGB and RYGB (RR = 2.31, I² = 0%, 95% CI = 0.78 to 6.79, P = 0.13) (Fig. 11). Regarding intraoperative complications, the OAGB group included four cases of bleeding, three cases of intestinal injury, and two cases of iatrogenic suture of nasogastric tube. The RYGB group included three cases of bleeding and one case of intestinal injury. Three articles were included in the early postoperative complications and there was statistical significance. It was observed that OAGB had fewer complications (RR = 0.45, I² = 0%, 95% CI = 0.21 to 0.97, P = 0.04) (Fig. 12). Among the early postoperative complications, the OAGB group included three serious complications (one of which required surgery) and six cases of general complications. The RYGB group included eight serious complications (two of which required surgery) and twelve cases of general complications. Late complications included all complications that were not staged in the article. Regarding late complications from the two surgical methods, a total of six articles were included and there was no difference (RR = 0.91, I² = 0%, 95% CI = 0.61 to 1.35, P = 0.65) (Fig. 13). Among those with late postoperative complications, the OAGB group included three serious complications (two of which required surgery), 17 general complications, and 19 unspecified stage complications. The RYGB group included 13 serious complications (nine of which required surgery), 15 general complications, and 15 unspecified stage complications. In addition, according to Robert et al., there were 24 cases of RYGB and 42 cases of OAGB surgery-related adverse events up to 2 years of follow-up.

Diabetes, hypertension, hyperlipidemia, and gastroesophageal reflux disease (GERD) were all obesity-related complications in our meta-analysis, and postoperative remission of all complications was based on the study’s last follow-up time. Postoperative diabetes remission included patients with partial and complete remission, which included five articles, and there was no difference between the two surgical methods (RR = 1.11, I² = 0%, 95% CI = 0.99 to 1.26, P = 0.08) (Fig. 14). Four articles were included for postoperative hypertension remission, and there was no difference between the two surgical methods (RR = 1.10, I² = 0%, 95% CI = 0.94 to 1.29, P = 0.21) (Fig. 15). Two articles were included for postoperative hyperlipidemia remission with statistical significance, and OAGB had a higher remission rate (RR = 1.39, I² = 0%, 95% CI = 1.20 to 1.61, P < 0.0001) (Fig. 16). Two articles were included for postoperative GERD remission, and no difference between the two surgical methods was observed (RR = 0.35, I² = 0%, 95% CI = 0.11 to 1.18, P = 0.09) (Fig. 17).

A total of seven articles were included for operation time, and there was statistical significance. The operation time of OAGB was significantly shorter than that of RYGB (MD = − 36.95, I² = 76%, 95% CI = − 44.56 to − 29.33, P < 0.00001) (Fig. 18).

Regarding postoperative revision surgery, Robert et al. excluded four patients from the population who changed from OAGB to RYGB [20]. In the study of Ruiz-Tovar et al., two patients changed to RYGB due to bile acid reflux after OAGB, and three patients underwent revision surgery because of weight gain after RYGB [23]. Two patients died in the study by Robert et al.; hence, they were excluded from the study [20]. There were no deaths in the remaining studies.

We analyzed preoperative BMI, and found no significant difference among the included studies, implying that comparing postoperative BMI is feasible for evaluating the effect of weight loss. We selected multiple indicators for comparison at each time point since the indicators reflecting the effect of weight loss were different in each study. In the meta-analysis of %EWL at 6 months, BMI at 6 months and 12 months, %EBMIL at 2 years, and BMI at 5 years, there was no significant difference in weight loss between OAGB and RYGB. The results of the %EWL and 2-year BMI analysis at 12 months after surgery showed that the weight loss effect of OAGB was more satisfactory than that of RYGB. In the included study, Ruiz-Tovar et al. believed that OAGB had a better weight loss effect since the biliopancreatic limbs were longer than required during digestive tract reconstruction [23]. According to the 2018 World Bariatric and Metabolic Surgery consensus, obese patients undergoing OAGB should have a biliopancreatic limb length of 150 cm when their BMI is < 50 and a biliopancreatic limb length of 200 cm when their BMI is > 50. The biliopancreatic limb length of obese patients undergoing RYGB should be between 50 and 150 cm, while the biliopancreatic limb length of Ruiz-Tovar et al.’s study participants for OAGB was 250–350 cm [9]. The duodenum and jejunum are important organs for food digestion and absorption; however, the biliopancreatic limb is a bypass segment of the small intestine, and ingested food does not pass through this channel, limiting nutrient absorption. In the long biliopancreatic limb, the absorption of more nutrients is restricted [32].

Some previous systematic reviews and meta-analyses have also shown that the weight loss effect of OAGB is not inferior to that of RYGB, and it is believed that this may be due to the longer indwelling biliopancreatic limbs during reconstruction. However, the included articles are few and bias is not well controlled; hence, conclusions are questionable [33, 34]. Similarly, we should realize that the length of the biliopancreatic limbs using OAGB is not as long as possible. While considering the effect of weight loss, the length of the common channel should be ensured to reduce the incidence of postoperative malnutrition [35, 36]. RYGB surgery has shorter biliopancreatic limbs than OAGB but a smaller stomach volume and more obvious restriction on food intake. Therefore, the long-term weight loss effects of the two surgery methods may be similar. This is also consistent with our study, which observed no significant differences between the two surgical methods in terms of %EBMIL at 2 years and BMI at 5 years. Therefore, the standard of biliopancreatic limb length and the influence of gastric pouch size on patients can be further explored and standardized in large-scale clinical studies to achieve the effect of weight loss without causing malnutrition.

Studies on early postoperative complications suggest that OAGB has fewer complications than RYGB, with lower a heterogeneity, larger sample size, and higher reliability. The study found that RYGB had more major complications than OAGB, among which anastomotic leakage was the most common. OAGB had no anastomotic leakage, while RYGB had a total of five. There are four possible reasons for the high incidence of anastomotic leakage in RYGB. First, OAGB has only one gastro-jejunal anastomosis, while RYGB has gastro-jejunal anastomosis and jejunal-jejunal anastomosis, increasing the incidence of anastomotic leakage in RYGB [9]. Second, whereas both RYGB and OAGB are involved in removing the right gastroepiploic and left gastroepiploic arteries, only one or two branches of the left gastroepiploic artery remained when the gastric pouch was formed after RYGB [37, 38]. This leads to differences in the blood supply of the anastomotic stoma between the two surgical methods, and since gastrojejunostomy leakage is more common in bariatric surgery, anastomotic leakage caused by blood supply cannot be excluded [39]. Third, RYGB is more difficult to operate than OAGB. Moreover, the learning curve is longer and negatively correlated with the occurrence of complications such as intraoperative bleeding, postoperative leakage, anastomotic stenosis, and marginal ulcers [30]. Fourth, the large gastric capacity and wide anastomotic stoma of OAGB make it less likely to cause increased intragastric pressure [9].

Regarding late postoperative complications, our study showed no significant differences between the two procedures. In addition, when we added 24 cases of RYGB and 42 cases of OAGB that had surgical complications 2 years after surgery, which were separately mentioned in the study of Robert et al., the results of the meta-analysis still showed no significant differences between the two groups. Common complications of OAGB include nutritional deficiencies, gallstones, and diarrhea; common complications of RYGB include anastomotic ulcers, gallstones, and abdominal pain. In particular, Rober et al. reported nine malnutrition complications in OAGB. This is consistent with the conclusion that OAGB has more nutritional complications as mentioned in some studies, and the most common cause of nutritional deficiency in OAGB may be related to our presumed longer biliopancreatic limbs, while RYGB longer biliopancreatic limbs inhibited more nutrient absorption than OAGB [14, 40, 41]. The anastomotic ulcer of RYGB may also be related to the increase in the number of anastomotic stoma, and it is related to factors such as less vascular supply, difficult operation, and increased intragastric pressure due to small gastric volume.

The remission of diabetes after the two surgical methods was P = 0.08, and there was no significant difference between the two; however, the remission of diabetes was still considered a high remission rate. In 2015, the International Diabetes Organization confirmed the good blood sugar control after metabolic surgery, and suggested that for diabetic patients with a BMI range of 30.0‒34.9 kg/m², obesity surgery should be considered when blood sugar control is poor [42].

There were differences in postoperative hyperlipidemia remission in the included article analysis, and the remission rate after OAGB was higher. However, only two articles were included, and one of them accounted for 96.2% of the weight; therefore, the evidence was weak, and no differences in the management of hyperlipidemia between the two surgical methods were observed. Moreover, there was no significant difference in the postoperative remission rate of patients with hypertension, and both surgical methods were considered effective in improving obesity and hypertension.

Bariatric surgery was previously considered a viable option for the treatment of morbidly obese patients with refractory GERD [43]. In our study, we found no significant difference in GERD remission; however, the sample size was small, included literature were few, evidence strength was weak, and conclusion questionable. RYGB improves some patients’ symptoms over time; however, the sample size was small and there are many subjective indicators; hence, the conclusion is not adequately reliable [21]. The study by Eskandaros et al. showed that the alkaline reflux resulting from RYGB was significantly lower than that from OAGB only at 6 and 12 months, while the GERD, percentage of esophageal acid exposure time, and number of acid reflux were significantly improved; however, there were no significant differences between them [25]. Analysis of the reasons for the improvement of alkaline reflux in the initial stage of RYGB may be due to the establishment of a bypass and the direct entry of biliopancreatic juice into the jejunal output loop, which is not directly connected with the gastric cavity. Moreover, RYGB can be used as a surgical option for obese patients with GERD symptoms [44]. However, in the long run, the treatment of GERD using OAGB is closely related to the long gastric pouch, large gastric volume, and wide anastomotic stoma, which can reduce the intragastric pressure or increase the gastroesophageal pressure gradient [30, 31, 45]. Because the gastric pouch is long, the alkaline body fluid in the intestine is not easily refluxed to the cardia and above. In addition, the gastric volume is large, and when there is an increased pressure in the stomach, there is a lesser chance of the food in the stomach flowing back into the esophagus. The width of the anastomotic stoma will not cause reflux of gastric contents into the esophagus because of the restricted discharge of gastric contents. Therefore, in our analysis, OAGB is no worse than RYGB for GERD remission.

Seven articles were included in a comparative analysis of the length of surgery and it was concluded that the operation time of OAGB was significantly shorter than that of RYGB. The heterogeneity of the analysis of the length of operation was relatively high. We performed the heterogeneity analysis by excluding articles apiece, and the study remained statistically significant after excluding the articles with high heterogeneity. The high heterogeneity screened out may be the reason why the mean operation time of this surgical method differs from the mean operation time of other articles.

Only one article that was included had postoperative revision, and the quality of the evidence was insufficient to effectively compare the two surgical modalities. In some obese patients with poor weight control, bile reflux, anastomotic ulcers, and other complications after other surgical methods, OAGB and RYGB can be used as revision methods [46]. Moreover, patients who need revision surgery after OAGB can continue with RYGB [40, 47]. However, patients who need a revision of RYGB surgery can only be changed to distal RYGB surgery or biliopancreatic diversion with duodenal switch (BPD-DS) [48].

In addition, Rober et al. used the Bariatric Analysis and Reporting Outcome System (BAROS) and the Impact of Weight on Quality of Life (IWQOL) to assess quality of life; Katayama et al. used the medical outcomes of a 36-item short-form health survey (SF-36) to assess quality of life; and Lee et al. used the Gastro-Intestinal Quality of Life Index (GIQLI) to assess the quality of life. All three articles indicated that the quality of life after the two surgeries was significantly improved compared to before the surgery; however, there was no significant difference between them. In addition, in the study of Ibrahim et al., GIQLI was used to assess the quality of life, and it was believed that the quality of life in the RYGB group was higher than that in the OAGB group at 6 and 12 months after surgery. However, since the methodology of quality of life evaluation in several articles is not uniform, and there is a large bias in reading one article alone, it cannot be determined which type of postoperative quality of life is better.

Our meta-analysis has some limitations. First, the number of articles evaluating the effect of weight loss after surgery is sufficient; however, the different expressed indicators lead to fewer articles that can be included in data analysis at various time points, which weakened the quality of the evidence. We hope that adding more articles in the future will improve the strength of the evidence, or suggest that researchers unify the indicators of postoperative weight loss. Second, some articles did not mention blinding, which may lead to an unreliable assessment of risk of bias. Third, the large difference in postoperative follow-up time resulted in fewer articles included in the analysis, which influenced the results of the analysis. Finally, the postoperative follow-up time was short; hence, no conclusion could be drawn on the treatment results of the sustained effect of weight loss and postoperative complications.

This article does not contain any studies with human participants or animals performed by any of the authors. Informed consent does not apply.

The authors declare no competing interests.

Data openly available in a public repository.