Magnetic vs stapled technique in jejuno-ileal bipartition: one-year outcomes Authors: Martin Fried, D. Michalsky, J. N. Buchwald, D. Charpentier, J. Schneider, Michel Gagner Published: 2026 ABSTRACT BackgroundMinimally invasive side-to-side magnetic jejuno-ileal (JI) bipartition (MagJI) is a novel treatment for patients with mild obesity and/or type 2 diabetes (T2D) that obviates complications from enterotomies and retained foreign materials and may provide efficacy equivalent to the standard stapled JI procedure.MethodsThe comparative, non-randomized, non-inferiority study evaluated the efficacy of magnetic vs stapled JI in adults with body mass index (BMI, kg/m2) ≥ 30.0– ≤ 39.9 and T2D matched by age and sex. After introduction of two linear magnets by swallowing and/or endoscopy beyond the ligament of Treitz, magnets were guided laparoscopically to target ileal and jejunal locations and apposed across bowel walls, initiating compression anastomosis. Endpoints: weight, BMI, HbA1C, and fasting plasma glucose (FPG) reduction.ResultsBetween March1, 2023, and June30, 2024, patients were enrolled and matched for comparison. No trocar site infection or severe adverse event related to the device or procedure in either group. At 1 year, regression-adjusted outcomes for an average baseline patient were as follows: weight 91.3 kg (stapled) vs 93.3 kg (magnetic); BMI 31.0 vs 31.3; HbA1C 56.2 vs 41.5 mmol/mol; FPG 8.6 vs 5.4 mmol/L. Non-inferiority was demonstrated for BMI (p = 0.017), HbA1C (p < 0.001), and FPG (p < 0.001), but not weight (p = 0.091).ConclusionsEarly quantitative evidence demonstrated that magnetic JI achieved metabolic outcomes (BMI, HbA1C, and FPG) non-inferior to those of stapled JI at 1 year. Minimally invasive MagJI is a safe, effective, anatomy sparing technique that may offer a patient-accessible metabolic/bariatric surgical option to treat mild obesity and T2D.Clinicaltrials.gov Identifier: NCT#06232746. Decades of addressing the global burden of obesity and type 2 diabetes (T2D) have resulted in robust evidence demonstrating that metabolic/bariatric surgery (MBS) achieves the most effective long-term reduction of body mass index (BMI) and T2D [1–4]. Yet, MBS is significantly underutilized in the worldwide treatment algorithm; < 1.0% of patients who may benefit from MBS receive it, regardless of continental variations in healthcare systems and reimbursement models [5, 6]. Multiple factors influence the deficient utilization of MBS. These include low physician referral rates and patients’ fear of undergoing the surgery due to the misperception that MBS is very invasive and carries a high risk of early and late complications without the option of reversibility [7–10]. While MBS, like all therapies for obesity and T2D, requires continued follow-up and dietary observation, it is predominantly—and increasingly—minimally invasive surgery (MIS) with a low perioperative mortality rate comparable to laparoscopic cholecystectomy (approximately 0.1%) [11]. In most cases, MBS is revisable if needed. Intensive technical innovation in MBS continues to lower the already modest risk profile and contribute to broader adoption of this most effective therapy. Recently, MIS single-anastomosis and magnetic procedures inaugurated a new era of less complex MBS operations. For example, use of gradual, incisionless, magnetic compression anastomosis (MCA) as an alternative to intraoperative cutting, stapling, and suturing to form an anastomosis has shown reduced invasiveness and increased safety in indirect comparisons [12]. In addition, early clinical outcomes of magnetic gastrointestinal (GI) procedures, such as side-to-side magnetic duodeno-ileostomy (MagDI), provide evidence of efficacy equivalent to that of MBS procedures that involve stapled/sutured anastomoses [13–16]. In 2003, Woods et al. published a preclinical study of stapled jejuno-ileostomy (JI) in a rodent model [17]. JI bipartition creates a side-to-side anastomosis between a proximal jejunal loop and a distal ileal loop, establishing two alimentary paths: one, the native intestinal route, and one through a bypass to the ileum. Partially digested nutrients arriving rapidly in the distal ileum stimulate L-cell release of glucagon-like peptide-1 (GLP-1) and peptide YY (PYY), enhancing glucose-dependent insulin secretion, suppressing glucagon, and activating the ileal brake—a neuroendocrine satiety signal that reduces caloric intake [18, 19]. Glucose-dependent insulinotropic polypeptide (GIP) release further stimulates insulin secretion and influences lipid metabolism. As food continues to flow through the native intestinal route, total intestinal continuity is preserved, avoiding the malnutrition and bacterial overgrowth associated with total small bowel exclusion procedures. These mechanistic principles were validated in the clinical studies of stapled side-to-side JI by Melissas et al. [20, 21] and Fried et al. [22] from 2012 to 2017. The safety profile and statistically significant T2D improvement of JI resulted in its approval for use in standard MBS practice since 2017. In 2024, our team initiated and, in 2025, reported [23] the first-in-human prospective open-label study to investigate the feasibility, safety, and preliminary efficacy of the biofragmentable linear magnetic anastomosis system (BMAS) in creating side-to-side magnetic JI (MagJI) in patients with class I-II obesity (BMI 30.0–39.9 kg/m2) and T2D (glycated hemoglobin [HbA1C] ≥ 6.5%). The MagJI study provided the basis for the current investigation, the first comparative study of magnetic vs stapled JI efficacy with 1-year follow-up. The primary aim of this investigation was to compare magnetic versus stapled JI anastomotic technique; stapled JI served as an established institutional comparator for the novel magnetic approach, rather than as a ‘gold standard’ MBS procedure. We hypothesized that magnetic compression anastomosis, by creating a physiologically equivalent side-to-side JI anastomosis without enterotomy or retained foreign material, would achieve weight and metabolic outcomes non-inferior to those of the established stapled technique at 1 year. Methods Study design and endpoints This single-center clinical investigation was a non-randomized, comparative, non-inferiority study of the efficacy of magnetic vs stapled side-to-side JI technique in patients with mild obesity and T2D. Prospective interim outcomes for the MagJI cohort (a continuous series, n = 10) previously reported at 3 and 6 months [23] were scheduled to be followed through 1 year and compared with 10 stapled JI patients (historical control group) randomly selected from our institutional JI database (n = 158). Treatment evaluation measurements were recorded at multiple time points: days 30, 60, 90, 180, and 360 (days 14 and 270, MagJI only). The primary analysis assessed whether magnetic JI technique was non-inferior to stapled JI technique with respect to baseline-adjusted 1-year outcomes for weight, BMI, HbA1C, and fasting plasma glucose (FPG). The study was approved by the medical center’s Ethics Committee and the State Authority for Drug and Device Approval. Patient inclusion Eligible candidates were introduced to the objectives, technical principles, and risks and benefits of the investigational operative technique. All interested patients engaged in discussion with the surgical team in order to make an informed decision regarding consent to study participation. Patients underwent standard preoperative multidisciplinary evaluation prior to enrollment including nutritional consultation with a registered dietitian consistent with MBS practice guidelines. No differential lifestyle intervention was applied to either group; both received identical standard postoperative institutional counseling regarding nutrition, physical activity, and medication compliance. Included patients were aged 18–65 years, had a BMI between 30.0 and 39.9 kg/m2, and had not undergone previous MBS. Patients had T2D (HbA1C ≥ 6.5% [≥ 48 mmol/mol]). Patients were requested to forego additional MBS or reconstructive surgery for 1 year; females were asked to use contraception and refrain from becoming pregnant through the 1-year study follow-up period. Key exclusion criteria were type 1 diabetes, latent autoimmune diabetes of adulthood, Barrett’s disease, hiatal hernia ≥ 5 cm, and any condition that might prevent cooperation with all follow-up assessments. Jejuno-ileostomy (JI) procedure In both the stapled and magnetic JI technique (using magnets that may be swallowed), the procedure comprises the side-to-side isoperistaltic anastomosis of two loops of small bowel at a point 150 cm proximal to the ileocecal valve and 100 cm distal from the ligament of Treitz (Fig. 1). At the end of each procedure, the mesenteric defect is closed with a running 2–0 non-absorbable braided polyester suture coated with PTFE (Tervalon®, Vitrex Medical A/S, Herlev, Denmark).Fig. 1Laparoscopic side-to-side jejuno-ileostomy (JI) bipartition using a 60-mm stapler has been performed successfully since 2017. Since 2023, a novel magnetic compression anastomosis technique using two 50-mm linear neodymium magnets that can be swallowed has also been used to perform JI (the MagJI procedure). If the patient can comfortably swallow two magnets, two endoscopic procedures are eliminated; if one magnet is swallowed, one endoscopy is averted. If neither magnet is swallowed, both are placed beyond the ligament of Treitz by endoscopy Laparoscopic side-to-side jejuno-ileostomy (JI) bipartition using a 60-mm stapler has been performed successfully since 2017. Since 2023, a novel magnetic compression anastomosis technique using two 50-mm linear neodymium magnets that can be swallowed has also been used to perform JI (the MagJI procedure). If the patient can comfortably swallow two magnets, two endoscopic procedures are eliminated; if one magnet is swallowed, one endoscopy is averted. If neither magnet is swallowed, both are placed beyond the ligament of Treitz by endoscopy Magnetic JI technique The magnetic JI technique described in detail by Gagner et al. in 2025 [23] begins with the patient swallowing the first (ileal) magnet. After approximately 2–3 h, depending on an individual patient’s gastric and intestinal motility, when the magnet is detected on X-ray beyond the ligament of Treitz, the second (jejunal) magnet is swallowed; when both magnets have arrived in the left mesogastrium, the patients is taken to the operating room. Alternatively, if only the first magnet is swallowed, upon its arrival in the mesogastrium, the patient is taken to the operating room for placement of the second magnet by endoscopy under general anesthesia. Patients who elect endoscopic placement of both magnets are taken directly to the operating room. Laparoscopic assistance is employed to refine magnet positions (ileal, 150 cm proximal to the ileocecal valve; jejunal, 100 cm distal from the ligament of Treitz) using the magnetic positioning tool. The tool’s base is placed on the serosal surface of the bowel above the intraluminal magnet to connect to it by attraction; the tool’s handle is used to gently ‘surf’ the magnet through the bowel to its target position (Fig. 2). Finally, the intraluminal ileal magnet is lifted into apposition with the intraluminal jejunal magnet, connecting with it through the bowel walls and initiating gradual compression anastomosis (Fig. 3). There is no enterotomy, bowel transection, or permanent fixation of the anastomosis with staples, suture, or other foreign material.Fig. 2When both magnets have arrived at the left mesogastrium via swallowing or endoscopic placement, the laparoscopic magnetic positioning device is brought into extraluminal connection with the first magnet and, by traction, ‘surfs’ it through the lumen to the target ileal position, followed by surfing the second magnet to the jejunal positionFig. 3The magnet in the ileal loop is lifted into apposition with the magnet in the jejunal loop. The magnets ‘click’ into alignment, initiating gradual compression and necrosis of tissue to form an anastomosis. In 7–21 days, the fused BMAS magnets drop into the intestine and are expelled by peristalsis When both magnets have arrived at the left mesogastrium via swallowing or endoscopic placement, the laparoscopic magnetic positioning device is brought into extraluminal connection with the first magnet and, by traction, ‘surfs’ it through the lumen to the target ileal position, followed by surfing the second magnet to the jejunal position The magnet in the ileal loop is lifted into apposition with the magnet in the jejunal loop. The magnets ‘click’ into alignment, initiating gradual compression and necrosis of tissue to form an anastomosis. In 7–21 days, the fused BMAS magnets drop into the intestine and are expelled by peristalsis The linear biofragmentable magnet anastomosis system (BMAS [GT Metabolic Solutions, Inc., San Jose, CA, USA]) magnets are 50-mm long × 10.2-mm wide × 4.4-mm thick (GT Metabolic Solutions). The BMAS magnet package incorporates a flexible orogastric delivery catheter used to advance magnets intraluminally when one (or both) magnets are not swallowed. Laparoscopic positioning tools of various magnetic strengths are also included. Stapled JI technique The laparoscopic stapled technique described by Melissas et al. and Fried et al. in 2017 [21, 22] is conducted under general anesthesia using a 60-mm linear stapler to create the JI anastomosis (ileal, 150 cm proximal to the ileocecal valve; jejunal, 100 cm distal from the ligament of Treitz). The linear stapler is inserted into the bowel lumen through two enterotomies created by an advanced bipolar energy device on the antimesenteric side of the bowel wall. After the stapler is fired, the anastomosis is inspected for bleeding and the JI enterotomy site is closed by hand in two layers of running absorbable suture. Statistical analysis Assignment of controls Ten MagJI patients were enrolled. Each was matched by sex and age (± 3 years) with 10 control patients selected from our institutional database of 158 cases who had undergone stapled JI between March 1, 2023, and June 30, 2024. When multiple potential matches were identified for a given MagJI patient, one control was randomly selected using a random number generator. Analysis All analyses were performed using IBM SPSS Statistics version 31.0 (IBM Corp., Armonk, NY) or MATLAB R2023a (The MathWorks, Inc., 2023). Quantitative variables were expressed as mean ± standard deviation (SD), ranges, and/or 95% confidence intervals (CI). Between-group differences were evaluated using independent-samples t-tests for quantitative variables and chi-square or Fisher’s exact tests for categorical variables. Within-group changes in quantitative variables were analyzed using paired-samples t-tests; statistical significance was set at p < 0.05 (two-tailed). Weight loss was calculated as total weight loss percentage (%TWL = [(Initial Weight − Follow-up Weight)/Initial Weight] × 100) and excess weight loss percentage (%EWL = [(Initial BMI − Follow-up BMI)/(Initial BMI − 25.0)] × 100), where 25.0 kg/m2 represents the upper limit of normal BMI. The proportion of patients achieving > 5.0% TWL was reported. Metabolic outcomes included FPG (mg/dL and mmol/L) and HbA1C (% and mmol/mol). Changes in T2D were evaluated in relation to American Diabetes Association diagnostic values: HbA1C ≥ 6.5% (≥ 48.0 mmol/mol); FPG ≥ 126.0 mg/dL (≥ 7.0 mmol/L) [24]. The primary analysis evaluated whether 1-year outcomes following MagJI were non-inferior to those following stapled JI with respect to weight, BMI, HbA1C, and FPG. In the absence of established minimal clinically important differences for JI procedures, a data-driven approach was adopted, setting a conservative threshold requiring MagJI to perform within the demonstrated range of the established stapled technique. Non-inferiority assessment was conducted using a three-step regression procedure. First, interaction terms (treatment × baseline) were tested to assess whether baseline–outcome relationships differed by treatment group. Second, the non-inferiority margin (U) for each outcome was defined as the upper 95% CI of the intercept from a linear regression model fitted within the stapled JI group using baseline values as the predictor. Third, the MagJI intercept was estimated using a parallel regression model in which the slope for baseline values was fixed to the common coefficient identified in step one. Baseline values were mean-centered at the stapled JI group average, allowing each intercept to represent the predicted 1-year outcome for a patient with average baseline characteristics. Non-inferiority was tested using one-sided t-tests comparing MagJI intercepts to their corresponding margins. The null hypothesis (H0: MagJI intercept > U) represented inferiority. Non-inferiority was concluded when the upper 95% CI bound for the MagJI prediction fell below U, equivalent to rejecting H0 (p < 0.05). In the absence of established minimal clinically important differences for JI procedures, a prespecified, empirically anchored approach was adopted, setting a conservative threshold requiring MagJI to perform within the observed range of the established stapled technique. This approach enables comparison conditional on baseline covariates and accounts for observed group imbalances under the specified model assumptions. Results Patient baseline characteristics Between March 1, 2023, and June 30, 2024, 10 patients were enrolled in a prospective study of magnetic JI, and for comparison, matched with 10 patient controls who underwent stapled JI. One intended MagJI patient withdrew before undergoing the procedure; thus, the study included nine MagJI patients (six male [66.7%], mean age 49.7 ± 7.9 years) and ten historical controls who underwent stapled JI (six male [60.0%], mean age 48.0 ± 8.8 years). Despite matching by sex and age, baseline anthropometric and metabolic parameters differed between groups. The MagJI group had a higher BMI (37.3 ± 3.3 kg/m2 vs 33.5 ± 4.3 kg/m2; p = 0.047) and a lower HbA1C (53.7 ± 6.2 mmol/mol [7.1%] vs 69.7 ± 10.4 mmol/mol [8.5%]; p < 0.001) compared with the stapled JI group. Baseline weight and FPG showed no significant difference (Table 1). Table 1Pre-, peri-, and postoperative stapled vs magnetic jejuno-ileostomy (JI) treatment group comparisonTreatmentStapled JIMagnetic JIp-value*(n = 10)(n = 9)PreoperativeMales660.999Age, yrs, mean ± SD (range)48.0 ± 8.849.7 ± 7.90.672(32.0–58.0)(32.0–59.0)Weight, kg, mean ± SD (range)99.1 ± 20.0115.6 ± 17.50.073(78.0–140.2)(92.0–137.2)Body mass index, kg/m2 (range)33.5 ± 4.337.3 ± 3.30.047(27.0–41.9)(31.0–41.1)HbA1C, mmol/mol (range)69.7 ± 10.453.7 ± 6.2 < 0.001(56.0–84.0)(48.0–65.0)Fasting plasma glucose, mmol/L (range)10.2 ± 2.28.0 ± 2.40.056(6.3–13.0)(5.5–13.4)PerioperativeOperative min, mean ± SD (range)102.5 ± 21.4104.4 ± 35.90.886(75.0–130.0)(70.0–190.0)Both magnets swallowed (n = 3)–86.7 ± 11.0–(78.0–99.0)1 magnet swallowed, 1 endoscopy (n = 6)–113.3 ± 41.7–(70.0–190.0)Hospital days, mean ± SD (range)3.8 ± 0.64.3 ± 0.70.101(3.0–5.0)(4.0–6.0)Magnet expulsion, days, mean ± SD (range)–21.0 ± 8.2–(10.0–31.0)Magnet expulsion, days, median–19.0–Postoperative1 yr mean ± SD (range)Mean change (95% CI)P-value†1 yr mean ± SD (range)Mean change (95% CI)P-value†P-value*Weight, kg91.3 ± 16.47.7 ± 8.30.017105.8 ± 16.99.8 ± 6.20.0010.077(66.9–122.0)(1.8, 13.7)(83.9–130.0)(5.1, 14.6)Body mass index, kg/m231.0 ± 4.12.5 ± 2.90.02334.1 ± 3.83.1 ± 1.8 < 0.0010.107(23.7–36.4)(0.4, 4.5)(28.0–40.4)(1.7, 4.5)HbA1C, mmol/mol56.2 ± 18.113.5 ± 20.90.07241.7 ± 6.612.0 ± 6.1 < 0.0010.036(28.0–40.4)(− 1.5, 28.5)(32.0–50.0)(7.3, 16.7)Fasting plasma glucose, mmol/L8.6 ± 3.91.6 ± 4.90.3496.5 ± 1.11.6 ± 2.20.0610.131(4.5–17.8)(− 2.0, 5.1)(4.5–8.1)(− 0.1, 3.2)*P-value for one-sided non-inferiority t-test Table 1Pre-, peri-, and postoperative stapled vs magnetic jejuno-ileostomy (JI) treatment group comparisonTreatmentStapled JIMagnetic JIp-value*(n = 10)(n = 9)PreoperativeMales660.999Age, yrs, mean ± SD (range)48.0 ± 8.849.7 ± 7.90.672(32.0–58.0)(32.0–59.0)Weight, kg, mean ± SD (range)99.1 ± 20.0115.6 ± 17.50.073(78.0–140.2)(92.0–137.2)Body mass index, kg/m2 (range)33.5 ± 4.337.3 ± 3.30.047(27.0–41.9)(31.0–41.1)HbA1C, mmol/mol (range)69.7 ± 10.453.7 ± 6.2 < 0.001(56.0–84.0)(48.0–65.0)Fasting plasma glucose, mmol/L (range)10.2 ± 2.28.0 ± 2.40.056(6.3–13.0)(5.5–13.4)PerioperativeOperative min, mean ± SD (range)102.5 ± 21.4104.4 ± 35.90.886(75.0–130.0)(70.0–190.0)Both magnets swallowed (n = 3)–86.7 ± 11.0–(78.0–99.0)1 magnet swallowed, 1 endoscopy (n = 6)–113.3 ± 41.7–(70.0–190.0)Hospital days, mean ± SD (range)3.8 ± 0.64.3 ± 0.70.101(3.0–5.0)(4.0–6.0)Magnet expulsion, days, mean ± SD (range)–21.0 ± 8.2–(10.0–31.0)Magnet expulsion, days, median–19.0–Postoperative1 yr mean ± SD (range)Mean change (95% CI)P-value†1 yr mean ± SD (range)Mean change (95% CI)P-value†P-value*Weight, kg91.3 ± 16.47.7 ± 8.30.017105.8 ± 16.99.8 ± 6.20.0010.077(66.9–122.0)(1.8, 13.7)(83.9–130.0)(5.1, 14.6)Body mass index, kg/m231.0 ± 4.12.5 ± 2.90.02334.1 ± 3.83.1 ± 1.8 < 0.0010.107(23.7–36.4)(0.4, 4.5)(28.0–40.4)(1.7, 4.5)HbA1C, mmol/mol56.2 ± 18.113.5 ± 20.90.07241.7 ± 6.612.0 ± 6.1 < 0.0010.036(28.0–40.4)(− 1.5, 28.5)(32.0–50.0)(7.3, 16.7)Fasting plasma glucose, mmol/L8.6 ± 3.91.6 ± 4.90.3496.5 ± 1.11.6 ± 2.20.0610.131(4.5–17.8)(− 2.0, 5.1)(4.5–8.1)(− 0.1, 3.2)*P-value for one-sided non-inferiority t-test Pre-, peri-, and postoperative stapled vs magnetic jejuno-ileostomy (JI) treatment group comparison *P-value for one-sided non-inferiority t-test Feasibility and safety All control group operations were completed laparoscopically. In the magnetic group, 100.0% swallowed the first magnet; three patients (33.3%) were also able to swallow the second magnet, and in six patients (66.7%), the second magnet was introduced by flexible endoscopy. Operative time and hospital stay were not significantly different. Return of bowel function, confirmed by auscultation of peristalsis, occurred on average 16 h post-procedure in the MagJI group and 20 h in the stapled JI group. In the control group, there were 0.0% trocar site infection and no intraoperative complications. All magnets were expelled from the body with no adverse events (Table 1). In the magnetic JI group, no severe adverse events related to the device or the procedure occurred intraoperatively or over the short term follow-up through 1 year. Within-group changes Both groups achieved significant reductions in absolute weight and BMI at 1-year follow-up. In the stapled JI group, weight decreased from 99.1 ± 20.0 kg to 91.3 ± 16.4 kg (p = 0.017), and BMI decreased from 33.5 ± 4.3 to 31.0 ± 4.1 kg/m2 (p = 0.023). HbA1C decreased from 69.7 ± 10.4 to 56.2 ± 18.1 mmol/mol (8.5 → 7.3%), approaching statistical significance (p = 0.072). FPG decreased from 10.2 ± 2.2 to 8.6 ± 3.9 mmol/L (184.0 → 155.0 mg/dL; p = 0.349) (Table 1). In the MagJI group, weight decreased from 115.6 ± 17.5 to 105.8 ± 16.9 kg (p = 0.001), BMI decreased from 37.3 ± 3.3 to 34.1 ± 3.8 kg/m2 (p < 0.001), HbA1C decreased from 53.7 ± 6.2 to 41.7 ± 6.6 mmol/mol (7.1 → 5.9%, p < 0.001), and FPG decreased from 8.0 ± 2.4 to 6.5 ± 1.1 mmol/L (144.0 → 117.0 mg/dL; p = 0.061) (Table 1). The regression-adjusted mean HbA1C of 41.5 mmol/mol (5.9%) in the MagJI group at 1 year fell below the ADA diagnostic threshold for T2D (48 mmol/mol [6.5%]), consistent with T2D remission by current criteria [24]. Longitudinal outcome trends for both groups are presented in Figs. 4 and 5, which illustrate greater inter-individual variability in the stapled group, particularly in baseline weight. Mean EWL was 29.3 ± 41.9% (stapled) vs 27.8 ± 15.8% (MagJI) (p = 0.922). Mean total weight loss (TWL) was 7.2 ± 8.7% (stapled) vs 8.5 ± 4.5% (MagJI) (p = 0.683). The proportion of patients achieving TWL > 5.0% at 1 year was 70.0% (stapled) vs 78.0% (MagJI).Fig. 4Evolution of jejuno-ileostomy (JI) treatment evaluation parameters: individual trajectories and group averages. The left column displays data from patients treated with the stapled JI; the right column, the magnetic JI group. Thin colored lines represent individual patients; the solid black line indicates the group mean for each timepoint; dashed red lines depict th 95% CI of the mean. Body mass index (BMI) (Color figure online)Fig. 5Evolution of jejuno-ileostomy (JI) treatment evaluation parameters: individual trajectories and group averages. The left column displays data from patients treated with the stapled JI; the right column, the magnetic JI group. Thin, colored lines represent individual patients; the solid black line indicates the group mean for each timepoint; dashed red lines depict th 95% CI of the mean. Fasting plasma glucose (FPG), glycated hemoglobin (HbA1C) (Color figure online) Evolution of jejuno-ileostomy (JI) treatment evaluation parameters: individual trajectories and group averages. The left column displays data from patients treated with the stapled JI; the right column, the magnetic JI group. Thin colored lines represent individual patients; the solid black line indicates the group mean for each timepoint; dashed red lines depict th 95% CI of the mean. Body mass index (BMI) (Color figure online) Evolution of jejuno-ileostomy (JI) treatment evaluation parameters: individual trajectories and group averages. The left column displays data from patients treated with the stapled JI; the right column, the magnetic JI group. Thin, colored lines represent individual patients; the solid black line indicates the group mean for each timepoint; dashed red lines depict th 95% CI of the mean. Fasting plasma glucose (FPG), glycated hemoglobin (HbA1C) (Color figure online) Non-inferiority analysis The interaction terms between baseline values and treatment group were non-significant for all outcomes (weight p = 0.404; BMI p = 0.395; HbA1C p = 0.550; FPG p = 0.284), indicating consistent baseline effects across groups and supporting the validity of the adjusted between-group comparisons. Non-inferiority was demonstrated for BMI (p = 0.017), HbA1C (p < 0.001), and FPG (p < 0.001), but not for weight (p = 0.091). Predicted 1-year outcomes for an average pre-treatment patient are summarized in Table 2. Specifically, model-adjusted 1-year weight predictions were 91.3 kg (95% CI 86.2, 96.5) for stapled JI and 93.3 kg (88.2, 98.5) for MagJI. Weight did not meet the non-inferiority criterion because the upper limit of the MagJI 95% CI (98.5 kg) exceeded the stapled JI upper limit (96.5 kg). Table 2Non-inferiority analysis of jejuno-ileostomy (JI) efficacy outcomesOutcomeStapled JI predicted mean (95% CI)Magnetic JI predicted mean (95% CI)Treatment difference(Δ)One-sided P-value*Baseline effect (slope)Non-inferior?Weight91.3 kg (86.2, 96.5)93.3 kg (88.2, 98.5)2.0 kg0.0910.75 kgper 1 kgbaselineNoBody mass index31.0 kg/m2 (28.9, 33.1)31.3 kg/m2 (29.7, 32.9)0.3 kg/m20.0170.75 kg/m2per 1 kg/m2 baselineYesHbA1C56.2 mmol/mol (42.2, 70.2)/7.3% (6.0, 8.6)41.5 mmol/mol (36.2, 46.7)/5.9% (5.5, 6.4) − 14.7 mmol/mol/ − 3.5% < 0.001 − 0.01 mmol/mol per 1 mmol/mol baselineYesFasting plasma glucose8.6 mmol/L (5.7, 11.5)/155.0 mg/dL (103.0, 208.0)5.4 mmol/L (3.8, 6.9)/97.0 mg/dL (69.0, 125.0) − 3.2 mmol/L/ − 58.0 mg/dL < 0.001 − 0.51 mmol/L per 1 mmol/L baselineYesP-value for one-sided non-inferiority t-test Table 2Non-inferiority analysis of jejuno-ileostomy (JI) efficacy outcomesOutcomeStapled JI predicted mean (95% CI)Magnetic JI predicted mean (95% CI)Treatment difference(Δ)One-sided P-value*Baseline effect (slope)Non-inferior?Weight91.3 kg (86.2, 96.5)93.3 kg (88.2, 98.5)2.0 kg0.0910.75 kgper 1 kgbaselineNoBody mass index31.0 kg/m2 (28.9, 33.1)31.3 kg/m2 (29.7, 32.9)0.3 kg/m20.0170.75 kg/m2per 1 kg/m2 baselineYesHbA1C56.2 mmol/mol (42.2, 70.2)/7.3% (6.0, 8.6)41.5 mmol/mol (36.2, 46.7)/5.9% (5.5, 6.4) − 14.7 mmol/mol/ − 3.5% < 0.001 − 0.01 mmol/mol per 1 mmol/mol baselineYesFasting plasma glucose8.6 mmol/L (5.7, 11.5)/155.0 mg/dL (103.0, 208.0)5.4 mmol/L (3.8, 6.9)/97.0 mg/dL (69.0, 125.0) − 3.2 mmol/L/ − 58.0 mg/dL < 0.001 − 0.51 mmol/L per 1 mmol/L baselineYesP-value for one-sided non-inferiority t-test Non-inferiority analysis of jejuno-ileostomy (JI) efficacy outcomes 0.75 kg per 1 kg baseline 0.75 kg/m2 per 1 kg/m2 baseline P-value for one-sided non-inferiority t-test Corresponding model-adjusted predictions (95% CIs) for stapled JI vs MagJI were as follows: BMI 31.0 kg/m2 (28.9, 33.1) vs 31.3 kg/m2 (29.7, 32.9); HbA1C 56.2 mmol/mol [7.3%] (42.2, 70.2) vs 41.5 mmol/mol [5.9%] (36.2, 46.7); FPG 8.6 mmol/L [155 mg/dL] (5.7, 11.5) vs 5.4 mmol/L [97 mg/dL] (3.8, 6.9) (Table 2). Figure 6 provides a visual representation of these non-inferiority analyses, illustrating the relationship between predicted outcomes and non-inferiority margins for each endpoint. Non-inferiority was most strongly supported for HbA1C and FPG (both p < 0.001).Fig. 6One-year outcomes for an average pre-treatment patient receiving either stapled or magnetic jejuno-ileostomy (JI) including 95% CIs. The vertical red dotted line indicates the non-inferiority margin: the upper threshold (U) below which the magnetic JI treatment must fall to be considered non-inferior. All outcome measures except weight demonstrated statistically significant non-inferiority (Color figure online) One-year outcomes for an average pre-treatment patient receiving either stapled or magnetic jejuno-ileostomy (JI) including 95% CIs. The vertical red dotted line indicates the non-inferiority margin: the upper threshold (U) below which the magnetic JI treatment must fall to be considered non-inferior. All outcome measures except weight demonstrated statistically significant non-inferiority (Color figure online) Baseline values significantly influenced 1-year outcomes. For weight and BMI, higher baseline measures predicted higher follow-up values: each 1 kg above the baseline mean corresponded to 0.75 kg higher weight at 1 year (p < 0.001), and each 1 kg/m2 above the baseline mean predicted 0.75 kg/m2 higher BMI at 1 year (p < 0.001). For metabolic parameters, the pattern differed. Higher baseline FPG was associated with greater improvement, with each 1 mmol/L increase at baseline predicting a 0.51 mmol/L greater reduction at one year (p < 0.05). Baseline HbA1C had minimal predictive value; however, as each 1 mmol/mol higher baseline HbA1C corresponded to only a 0.01 mmol/mol decrease at 1 year (p > 0.05). Discussion The recent first-in-human study of side-to-side magnetic JI anastomosis bipartition demonstrated initial feasibility and safety of the procedure and BMAS device at 6-month follow-up. Patients who were comfortable with swallowing both BMAS magnets (33.3%) obviated two endoscopic procedures; 66.7% swallowed one magnet, averting one endoscopy prior to laparoscopic apposition of the intraluminal magnets [23]. The current study extends these findings with the first 1-year non-inferiority comparison of magnetic vs stapled JI. MagJI demonstrated non-inferior outcomes for BMI and metabolic parameters. Importantly, baseline differences between groups would be expected to favor the stapled cohort, suggesting that the observed non-inferiority of MagJI may represent a conservative estimate of effect. While both techniques achieved significant improvements, regression-adjusted HbA1C and FPG values favored MagJI (41.5 vs 56.2 mmol/mol and 5.4 vs 8.6 mmol/L, respectively), suggesting magnetic anastomosis may achieve metabolic benefits at least equivalent to, and possibly greater than the established technique. The current study’s 1-year follow-up precludes definitive conclusions regarding the durability of HbA1C improvement; however, existing literature on stapled JI bipartition without SG suggests that metabolic improvements are maintained at longer follow-up [21, 22], and longer-term follow-up of the MagJI cohort is ongoing. Non-significant interaction effects indicated that baseline values predicted outcomes similarly in both groups, validating the regression-based comparison methodology. These findings align with previous preclinical and clinical studies demonstrating that magnetic compression anastomosis produces physiologic effects comparable to stapled anastomoses while reducing procedural invasiveness [25–27]. The absence of retained staples or sutures in MCA eliminates a recognized source of chronic anastomotic inflammation, ulceration, and adhesion formation that has been documented with conventional stapled and sutured techniques [28–30]. Comparison of JI to other published studies of JI-centric procedures (e.g., JI bipartition, jejuno-ileal bypass, partial jejunal diversion, JI transit bipartition) is not straightforward as the procedure and nomenclature are not standardized. However, most studies of side-to-side JI describe variations of a dual path created to convey a portion of chyme rapidly through a newly established, more direct connection of jejunum to ileum, while maintaining essential nutrient absorption by moving chyme through the full native intestine. The metabolic effects of weight loss and T2D improvement are accomplished in part by early stimulation of incretin release, including GLP-1 and PYY [31, 32]. Several studies in the MBS literature, such as those of Liang et al. [33] and Santoro et al. [34], describe jejunal and/or ileal transit bipartition with SG; however, these refer to jejunojejunal, duodeno-jejunal, gastro-jejunal, gastro-ileal, or ileo-ileal anastomoses rather than to JI bipartitions. While Hassn et al. [35] and Tosta et al. [36] report a modified JI bypass with SG, the studies do not compare JI outcomes with and without SG to establish the contribution of JI alone. The few investigations of stapled JI bipartition without SG have shown it to be feasible and metabolically effective for BMI and T2D reduction in both preclinical and clinical settings. An early study (2003) by Woods et al. noted that following partial jejunal diversion in rodents fed a high-fat diet who were fasted to induce metabolic dysfunction, or, rodents who underwent a mixed-meal tolerance test, GLP-1 was elevated significantly and insulin levels were reduced [17]. Similar results were published by Evans et al. in diet-induced insulin-resistant rhesus macaques, where side-to-side JI anastomosis was achieved by magnetic compression. These authors reported HOMA-IR improvement and linear weight decrease over the 6-week procedure follow-up with a significant increase in adiponectin and a decrease in leptin levels [37]. In our institution’s clinical series of side-to-side stapled JI patients without SG (n = 158) from which the current study controls were selected, indirect evidence of the efficacy of JI to reduce T2D independently was observed. Notably, one patient who attained rapid T2D resolution following JI later experienced complete T2D recurrence within three months after anastomosis reversal, demonstrating the procedure’s direct metabolic effect. Li et al. reported outcomes of side-to-side JI without SG in a cohort of 57 diabetic patients of Chinese ethnicity (BMI 24.0–32.0 kg/m2) with all patients achieving either significant T2D improvement (n = 23, 40.0%) or complete resolution (n = 34, 60.0%) at 1 year [38]. Wang et al. recently reported retrospective evaluation of metabolic disease indicators in 78 patients with overweight or mild obesity and T2D who underwent side-to-side JI without SG. At 6-month follow-up, weight, HbA1C, FPG, total and LDL cholesterol, and triglycerides were significantly reduced and fasting insulin, 2-h insulin, and 2-h C-peptide indices were all significantly higher. Although results were limited to 6-month follow-up, JI meaningfully reduced weight and improved T2D [39]. These findings are in agreement with the preliminary reports of effective T2D improvement after JI without SG by Melissas et al. [21] and Fried et al. [22], and in accord with the current study’s non-inferiority comparison. Collectively, existing evidence supports the proposition that magnetic JI without SG represents a safe and effective primary therapy for selected patients with mild obesity and/or T2D, with potential metabolic advantages over stapled techniques. Relative to the published JI literature, the current study’s regression-adjusted 1-year HbA1C of 41.5 mmol/mol (5.9%) in the MagJI group, which falls below the ADA diagnostic threshold for T2D, compares favorably to reported outcomes after stapled JI without SG [21, 22, 38, 39]. This suggests that MagJI delivered metabolic outcomes at least equivalent to those of the established technique. In terms of medication reduction, the current study did not systematically collect antidiabetic medication data, a limitation that will be addressed in the planned larger-scale follow-up study. Magnetic JI may also be a feasible revisional addition to a complicated SG, or to boost waning SG BMI and comorbid disease reduction [40, 41] similar to that observed in series of magnetic DI after SG [15, 16]. Revisional MBS accounted for approximately 11.4% of MBS operations in the United States and 10.4% worldwide in 2024 [42], making safe, reliable augmentations or modifications of primary operations increasingly needed. MagJI may also be appropriate to consider in the context of hybrid and staged medical and surgical therapies. Despite the well-established efficacy of procedures such as Roux-en-Y gastric bypass, biliopancreatic diversion, sleeve gastrectomy, SADI-S, and one-anastomosis gastric bypass across a broad range of BMI and comorbidity profiles [43–47], a significant proportion of patients with mild obesity and/or short-duration T2D do not access or elect these procedures due to personal preference, fear of surgical risk, or healthcare system barriers [5–10]. For this stratum of patients, a minimally invasive, anatomy-preserving option with a favorable safety profile and meaningful metabolic efficacy may lower the threshold for surgical intervention and broaden access to effective metabolic therapy [48]. The BMAS device used in MagJI, and other MBS procedures, is an innovative minimally invasive tool that may aid in providing these benefits to both surgical teams and patients [49]. Longer-term head-to-head comparison studies of MagJI with other minimally invasive therapies are needed to determine whether incisionless, staple/suture-free MagJI is such a procedure. Limitations The small sample size and single-center design limit statistical power and generalizability. The non-randomized design with historical controls introduces potential selection bias and residual confounding. Strengths include prospective MagJI enrollment and use of covariate-adjusted analyses. Larger, multicenter randomized studies with longer follow-up are needed to confirm these findings and assess durability and metabolic outcomes. Conclusions This non-randomized comparative analysis of two small MBS cohorts provided preliminary evidence that magnetic JI resulted in statistically significant BMI, HbA1C, and FPG reduction not inferior to that of stapled JI over 1 year follow-up. The minimally invasive MagJI procedure is a safe, effective, anatomy sparing technique that may offer a patient-accessible MBS option for treatment of mild obesity and T2D.