Preoperative Weight Loss Intervention With Glucagon-Like Peptides 1 Receptor Agonists Is Associated With Lower Complication Rates Following Primary Total Knee Arthroplasty Than Preoperative Bariatric Surgery Authors: George Bcharah, Austin E. Wininger, Zaid Elsabbagh, Collin L. Braithwaite, Paul R. Van Schuyver, Mark J. Spangehl, Henry D. Clarke, Joshua S. Bingham Published: 2026 ABSTRACT BackgroundObesity is a risk factor for adverse outcomes following total knee arthroplasty (TKA). Bariatric surgery (BS) and glucagon-like peptide-1 receptor agonists (GLP-1 RAs) are weight-loss strategies commonly encountered in TKA patients. This study aimed to compare short- and long-term complications among obese patients treated with either BS or GLP-1 RAs prior to TKA.MethodsThe database was queried to identify patients with obesity (body mass index [BMI] ≥35 kg/m2) undergoing primary TKA. Patients were stratified based on interventions within 1 year prior to surgery: BS, GLP-1 RA, and no weight-loss intervention. Propensity score matching was performed to balance baseline characteristics. Subgroup analyses were conducted for BMI 35–40 kg/m2 and >40 kg/m2. Primary outcomes included periprosthetic joint infection (PJI), aseptic loosening, periprosthetic fracture, and revision surgery at 90 days, 2 years, and 5 years.ResultsIn the matched BMI ≥35 kg/m2 cohort (n = 3635 per group), patients with prior BS had significantly higher odds of all surgical complications, including PJI (odds ratio [OR]: 1.70) and all-cause revision (OR: 1.69) compared to the GLP-1 RA cohort. Compared to the no-intervention cohort, BS was associated with increased risks of PJI (OR: 1.33) and periprosthetic fracture (OR: 2.01). Conversely, the GLP-1 RA cohort showed significantly lower risks of surgical complications compared to the no-intervention cohort across most time points and BMI subgroups.ConclusionsPreoperative GLP-1 RA use is associated with lower complication rates than BS and demonstrates a protective safety profile compared with no weight-loss intervention, underscoring their potential role in preoperative obesity patients undergoing TKA. Introduction Obesity, defined as a body mass index (BMI) ≥30 kg/m2, affects nearly 40% of adults in the United States [1]. Elevated BMI is a risk factor for the development of knee osteoarthritis and the need for total knee arthroplasty (TKA) [[2], [3], [4]]. The prevalence of obesity in patients undergoing primary TKA is projected to reach 70% by 2029 [5]. Obese patients who undergo TKA are known to be at an increased risk of postoperative complications, including increased rates of wound complications, periprosthetic joint infection (PJI), mechanical complications (eg, implant instability and/or failure), aseptic loosening, and revision surgery [[6], [7], [8], [9]]. As the prevalence of obesity continues to rise globally, there is a need to better understand the comparative impact of various preoperative weight management strategies on short- and long-term outcomes following primary TKA [10]. Arthroplasty surgeons commonly encounter surgical patients who are actively trying to lose weight or may request that a patient do so prior to surgery [11,12]. Weight reduction strategies generally fall into surgical and pharmacologic categories, each with distinct risk profiles and potential impacts on postoperative outcomes [13]. Bariatric surgery (BS) can provide rapid, substantial, and sustained weight loss and has seen substantial innovation with more laparoscopic techniques [14]. Literature on whether prior BS improves outcomes following TKA is mixed, but some studies suggest a reduction in select 90-day complications, such as venous thromboembolism and stroke [15]. In contrast, glucagon-like peptide-1 receptor agonists (GLP-1 RAs) facilitate weight reduction through central appetite suppression effects, delayed gastric emptying, and improved glycemic control [16,17]. Moreover, GLP-1 RAs exert anti-inflammatory effects through immune cell signaling, reducing pro-inflammatory cytokines, and reducing nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) activation, which may be beneficial for soft tissue healing after TKA [18]. Some evidence also suggest that GLP-1 promotes bone formation and inhibits bone resorption through direct and indirect influences on multiple molecular pathways, including vitamin D regulation and osteoblast/osteoclast expression [19,20]. Although both BS and pharmacologic interventions are widely employed, there is a paucity of literature comparing their impacts on outcomes following TKA, such as PJI, aseptic loosening, and periprosthetic fracture [16,21,22]. Given the differing physiological impacts of BS and pharmacologic weight-loss therapy, a comparative analysis of these interventions is essential to guide clinical decisions and preoperative counseling for obese patients undergoing TKA [23,24]. This study aimed to evaluate the comparative risk of PJI, periprosthetic fracture, instability, aseptic loosening, and all-cause revision surgery in obese patients undergoing primary TKA who had been treated with BS or GLP-1 RAs in the year prior to TKA. A secondary aim was to perform separate analyses to evaluate these same outcomes in obese patients undergoing primary TKA who had been treated with GLP-1 RAs or BS compared to obese patients who received no weight-loss intervention in the year prior to TKA. Material and methods Data source This study was conducted using the TriNetX Research Network (TriNetX LLC, Cambridge, MA). TriNetX is a federated health research platform providing deidentified electronic medical record data from over 100 organizations across 30 countries and more than 140 million patients within the research platform [25]. The TriNetX network used for this analysis included 97 healthcare organizations at the time of query. The database was queried in April 2025 using Current Procedural Terminology (CPT) and International Classification of Diseases Tenth Revision (ICD-10) codes. This database has demonstrated reliable utility to address multiple research topics and has been validated in orthopaedic literature [26,27]. Given that TriNetX provides fully deidentified patient data, this study was exempt from institutional review board approval. Study population A total of 135,739 patients aged 18 years and older with a diagnosis of obesity (BMI ≥35 kg/m2) undergoing primary TKA (CPT 27447) for knee osteoarthritis were identified. Patients were then stratified into 3 cohorts: 1) those who had been prescribed GLP-1 RAs within 1 year prior to TKA; 2) those who had BS (ie, laparoscopic and/or open gastric restrictive procedures (with or without bypass), biliopancreatic diversion with duodenal switch, or other documented BS history) within 1 year prior to TKA; and 3) a no-intervention cohort, which was defined as those who never underwent BS and/or were not prescribed GLP-1 RAs within 1 year of the TKA. Patients who had treatment with both BS and GLP-1 RAs were excluded. To address the heterogeneity of risks associated with different classes of obesity, a stratified subanalysis was performed for Class II (BMI 35–40 kg/m2) and Class III (BMI >40 kg/m2) obesity. Outcomes of interest The primary outcomes analyzed were PJI, periprosthetic fractures, knee instability, aseptic loosening, and all-cause revision TKA. These outcomes were evaluated at multiple intervals postoperatively: 90 days, 2 years, 5 years, and 10 years. Secondary medical outcomes evaluated at 90 days postoperatively included deep vein thrombosis, bleeding event, myocardial infarction, stroke, emergency department visit, mortality, pulmonary embolism, acute kidney injury, transfusion, and hospital readmission. Bleeding event was defined by the occurrence of various types of hemorrhagic events, including gastrointestinal hemorrhages, intracranial hemorrhages, hemoptysis and respiratory tract bleeding, hemothorax, hemopericardium, among others. All outcomes were identified through specific ICD-10 diagnostic and CPT codes. Propensity score matching Baseline characteristics including demographics, comorbidities, pertinent medications (eg, corticosteroids, nonsteroidal anti-inflammatory agents), and lab values (eg, albumin) were collected using ICD-10 codes. A 1:1 propensity score matching (PSM) using a greedy nearest-neighbor matching algorithm with a caliper of 0.1 pooled standard deviations was done using 30 baseline characteristics, which were chosen based on the Charlson Comorbidity Index and other relevant comorbidities [28]. Of note, PSM was performed independently for each outcome time point, but only complete data for the largest outcome cohort were included to avoid redundancy. Data analyses Statistical analyses were performed within TriNetX. Baseline characteristics and demographics were summarized using descriptive statistics, with means ± standard deviations for continuous variables and percentages for categorical variables. Cohort differences were assessed using t-tests for continuous variables and chi-square tests for categorical variables. Comparative outcomes between cohorts were analyzed using the TriNetX “Compare Outcomes” tool, calculating odds ratios (ORs) with 95% confidence intervals (CIs). Statistical significance was defined at P < .05. Final cohort A total of 135,739 patients were identified meeting the inclusion criteria. Of these, 4210 patients were in the BS cohort, 7482 patients were in the GLP-1 RA cohort, and 124,047 patients were in the no-intervention cohort. After PSM for the primary comparison (BS vs GLP-1 RAs), 7270 patients (3635 in each group) were included. For the secondary comparisons, 8420 patients (4210 in each group) were matched for the BS vs no-intervention analysis, and 14,956 patients (7478 in each group) were matched for the GLP-1 RA vs no intervention analysis. There were no differences in demographic information, baseline comorbidities, or nutritional status between the matched cohorts (Table 1, Table 2, Table 3).Table 1Patient characteristics in the bariatric surgery and GLP-1 receptor agonist cohorts before and after propensity score matching (BMI >35 kg/m2).VariableBariatric surgery (before) n = 4210GLP-1 receptor agonist (before) n = 7482P valueBariatric surgery (after) n = 3635GLP-1 receptor agonist (after) n = 3635P valueAge, y (mean ± standard deviation)60.3 ± 8.863.3 ± 8.3<.00161.3 ± 8.561.2 ± 8.3.622Female, n (%)3452 (82.0)5001 (66.8)<.0012970 (81.7)2475 (68.1)<.001White, n (%)3217 (76.4)5796 (77.5).1942822 (77.6)2829 (77.8).844Black or African American, n (%)786 (18.7)1090 (14.6)<.001607 (16.7)589 (16.2).569Hispanic or Latino, n (%)227 (5.4)491 (6.6).011203 (5.6)207 (5.7).839Essential hypertension, n (%)1926 (45.8)4862 (65.0)<.0011753 (48.2)1746 (48.0).870Diabetes mellitus, n (%)691 (16.4)3594 (48.0)<.001691 (19.0)685 (18.9).857Chronic kidney disease, n (%)183 (4.4)742 (9.9)<.001176 (4.8)168 (4.6).659Hyperlipidemia, n (%)1060 (25.2)3512 (46.9)<.0011014 (27.9)1025 (28.2).774Diseases of liver, n (%)326 (7.7)1034 (13.8)<.001302 (8.3)295 (8.1).765Heart failure, n (%)202 (4.8)526 (7.0)<.001179 (4.9)195 (5.4).396Atrial fibrillation and flutter, n (%)200 (4.8)440 (5.9).01184 (5.1)149 (4.1).050Nicotine dependence, n (%)252 (6.0)819 (11.0)<.001244 (6.7)237 (6.5).741Chronic obstructive pulmonary disease, n (%)186 (4.4)534 (7.1)<.001172 (4.7)169 (4.7).868Cerebrovascular diseases, n (%)207 (4.9)449 (6.0).015180 (5.0)165 (4.5).408Coagulation defects, n (%)180 (4.3)336 (4.5).587148 (4.1)140 (3.9).631Osteoporosis without fracture, n (%)224 (5.3)302 (4.0).001165 (4.5)166 (4.6).955Gout, n (%)128 (3.0)454 (6.1)<.001121 (3.3)119 (3.3).896Peripheral vascular disease, n (%)97 (2.3)281 (3.8)<.00190 (2.5)93 (2.6).822Sepsis, n (%)104 (2.5)239 (3.2).02680 (2.2)76 (2.1).746Acute myocardial infarction, n (%)86 (2.0)215 (2.9).00776 (2.1)76 (2.1)1.000Rheumatoid arthritis, n (%)144 (3.4)271 (3.6).572128 (3.5)119 (3.3).56Cerebral infarction, n (%)84 (2.0)177 (2.4).19373 (2.0)66 (1.8).549Peptic ulcer, n (%)61 (1.5)69 (0.9).00944 (1.2)47 (1.3).752Systemic lupus erythematosus, n (%)37 (0.9)50 (0.7).20326 (0.7)28 (0.8).785Human immunodeficiency virus, n (%)10 (0.2)18 (0.2).97410 (0.3)10 (0.3)1.000Nonsteroidal anti-inflammatory drugs, n (%)753 (17.9)2105 (28.1)<.001704 (19.4)713 (19.6).790Body mass index (mean ± standard deviation)40.2 ± 7.940.2 ± 5.8.90940.0 ± 7.840.6 ± 5.8.306Albumin (mean ± standard deviation)3.98 ± 0.44.11 ± 0.4<.0014.06 ± 0.44.09 ± 0.4.685Table 2Patient characteristics in the bariatric surgery and no-intervention cohorts before and after propensity score matching (BMI >35 kg/m2).VariableBariatric surgery (before) n = 4210No intervention (before) n = 124,047P valueBariatric surgery (after) n = 4210No intervention (after) n = 4210P valueAge, y (mean ± standard deviation)60.3 ± 8.865.3 ± 9.1<.00160.3 ± 8.860.1 ± 9.2.268Female, n (%)3452 (82.0)79,945 (64.5)<.0013452 (82.0)2772 (65.8)<.001White, n (%)3217 (76.4)98,972 (79.8)<.0013217 (76.4)3206 (76.2).778Black or African American, n (%)786 (18.7)16,726 (13.5)<.001786 (18.7)794 (18.9).823Hispanic or Latino, n (%)227 (5.4)6915 (5.6).611227 (5.4)202 (4.8).215Essential hypertension, n (%)1926 (45.8)55,012 (44.4).0721926 (45.8)1909 (45.3).710Diabetes mellitus, n (%)691 (16.4)19,637 (15.8).308691 (16.4)655 (15.6).284Chronic kidney disease, n (%)186 (4.4)6059 (4.9).971186 (4.4)173 (4.1).098Hyperlipidemia, n (%)1060 (25.2)36,018 (29.0)<.0011060 (25.2)1017 (24.2).277Diseases of liver, n (%)326 (7.7)7753 (6.3)<.001326 (7.7)300 (7.1).280Heart failure, n (%)202 (4.8)4845 (3.9).003202 (4.8)161 (3.8).028Atrial fibrillation and flutter, n (%)200 (4.8)6085 (4.9).647200 (4.8)158 (3.8).023Nicotine dependence, n (%)252 (6.0)8044 (6.5).196252 (6.0)250 (5.9).927Chronic obstructive pulmonary disease, n (%)183 (4.4)5466 (4.4).111183 (4.4)156 (3.7).588Cerebrovascular diseases, n (%)207 (4.9)5852 (4.7).549207 (4.9)192 (4.6).442Coagulation defects, n (%)180 (4.3)3735 (3.0)<.001180 (4.3)140 (3.3).023Osteoporosis without fracture, n (%)224 (5.3)5261 (4.2).001224 (5.3)212 (5.0).555Gout, n (%)128 (3.0)4842 (3.9).004128 (3.0)139 (3.3).494Peripheral vascular disease, n (%)97 (2.3)2996 (2.4).64497 (2.3)88 (2.1).503Sepsis, n (%)104 (2.5)1862 (1.5)<.001104 (2.5)95 (2.3).519Acute myocardial infarction, n (%)86 (2.0)2473 (2.0).82386 (2.0)71 (1.7).227Rheumatoid arthritis, n (%)144 (3.4)3558 (2.9).035144 (3.4)126 (3.0).266Cerebral infarction, n (%)84 (2.0)2237 (1.8).35884 (2.0)77 (1.8).578Peptic ulcer, n (%)61 (1.5)839 (0.7)<.00161 (1.5)57 (1.4).711Systemic lupus erythematosus, n (%)37 (0.9)706 (0.6).00937 (0.9)26 (0.6).164Human immunodeficiency virus, n (%)10 (0.2)202 (0.2).24110 (0.2)10 (0.2)1.000Nonsteroidal anti-inflammatory drugs, n (%)753 (17.9)19,165 (15.5)<.001753 (17.9)740 (17.6).711Body mass index (mean ± standard deviation)40.2 ± 7.936.4 ± 5.4<.00140.2 ± 7.937.7 ± 6.3<.001Albumin (mean ± standard deviation)3.98 ± 0.44.09 ± 0.4<.0014.04 ± 0.44.09 ± 0.4.532Table 3Patient characteristics in the GLP-1 RAs and no-intervention cohorts before and after propensity score matching (BMI >35 kg/m2).VariableGLP-1 RA (before) n = 7482No intervention (before) n = 124,047P valueGLP-1 RA (after) n = 7478No intervention (after) n = 7478P valueAge, y (mean ± standard deviation)63.3 ± 8.365.3 ± 9.1<.00163.3 ± 8.363.2 ± 8.8.652Female, n (%)5001 (66.8)79,945 (64.5)<.0015000 (66.9)4803 (64.2).001White, n (%)5796 (77.5)98,972 (79.8)<.0015794 (77.5)5777 (77.3).740Black or African American, n (%)1090 (14.6)16,726 (13.5).0081090 (14.6)1056 (14.1).428Hispanic or Latino, n (%)491 (6.6)6915 (5.6)<.001490 (6.6)495 (6.6).869Essential hypertension, n (%)4862 (65.0)55,012 (44.4)<.0014858 (65.0)4858 (65.0)1.000Diabetes mellitus, n (%)3594 (48.0)19,637 (15.8)<.0013590 (48.0)3586 (48.0).948Chronic kidney disease, n (%)742 (9.9)6059 (4.9)<.001741 (9.9)666 (8.9).036Hyperlipidemia, n (%)3512 (46.9)36,018 (29.0)<.0013509 (46.9)3496 (46.8).831Diseases of liver, n (%)1034 (13.8)7753 (6.3)<.0011031 (13.8)1014 (13.6).686Heart failure, n (%)526 (7.0)4845 (3.9)<.001524 (7.0)461 (6.2).038Atrial fibrillation and flutter, n (%)440 (5.9)6085 (4.9)<.001440 (5.9)471 (6.3).289Nicotine dependence, n (%)819 (11.0)8044 (6.5)<.001817 (10.9)774 (10.4).254Chronic obstructive pulmonary disease, n (%)534 (7.1)5466 (4.4)<.001534 (7.1)474 (6.3).050Cerebrovascular diseases, n (%)449 (6.0)5852 (4.7)<.001449 (6.0)381 (5.1).015Coagulation defects, n (%)336 (4.5)3735 (3.0)<.001336 (4.5)311 (4.2).315Osteoporosis without fracture, n (%)302 (4.0)5261 (4.2).393302 (4.0)271 (3.6).187Gout, n (%)454 (6.1)4842 (3.9)<.001454 (6.1)478 (6.4).417Peripheral vascular disease, n (%)281 (3.8)2996 (2.4)<.001281 (3.8)256 (3.4).272Sepsis, n (%)239 (3.2)1862 (1.5)<.001238 (3.2)213 (2.9).232Acute myocardial infarction, n (%)215 (2.9)2473 (2.0)<.001215 (2.9)190 (2.5).208Rheumatoid arthritis, n (%)271 (3.6)3558 (2.9)<.001271 (3.6)249 (3.3).326Cerebral infarction, n (%)177 (2.4)2237 (1.8)<.001177 (2.4)168 (2.3).624Peptic ulcer, n (%)69 (0.9)839 (0.7).01369 (0.9)68 (0.9).932Systemic lupus erythematosus, n (%)50 (0.7)706 (0.6).27150 (0.7)42 (0.6).403Human immunodeficiency virus, n (%)18 (0.2)202 (0.2).1118 (0.2)11 (0.2).193Nonsteroidal anti-inflammatory drugs, n (%)2105 (28.1)19,165 (15.5)<.0012102 (28.1)2105 (28.2).957Body mass index (mean ± standard deviation)40.2 ± 5.836.4 ± 5.4<.00140.2 ± 5.837.4 ± 6.1<.001Albumin (mean ± standard deviation)4.11 ± 0.44.09 ± 0.4<.0014.11 ± 0.44.12 ± 0.4.668 Table 1Patient characteristics in the bariatric surgery and GLP-1 receptor agonist cohorts before and after propensity score matching (BMI >35 kg/m2).VariableBariatric surgery (before) n = 4210GLP-1 receptor agonist (before) n = 7482P valueBariatric surgery (after) n = 3635GLP-1 receptor agonist (after) n = 3635P valueAge, y (mean ± standard deviation)60.3 ± 8.863.3 ± 8.3<.00161.3 ± 8.561.2 ± 8.3.622Female, n (%)3452 (82.0)5001 (66.8)<.0012970 (81.7)2475 (68.1)<.001White, n (%)3217 (76.4)5796 (77.5).1942822 (77.6)2829 (77.8).844Black or African American, n (%)786 (18.7)1090 (14.6)<.001607 (16.7)589 (16.2).569Hispanic or Latino, n (%)227 (5.4)491 (6.6).011203 (5.6)207 (5.7).839Essential hypertension, n (%)1926 (45.8)4862 (65.0)<.0011753 (48.2)1746 (48.0).870Diabetes mellitus, n (%)691 (16.4)3594 (48.0)<.001691 (19.0)685 (18.9).857Chronic kidney disease, n (%)183 (4.4)742 (9.9)<.001176 (4.8)168 (4.6).659Hyperlipidemia, n (%)1060 (25.2)3512 (46.9)<.0011014 (27.9)1025 (28.2).774Diseases of liver, n (%)326 (7.7)1034 (13.8)<.001302 (8.3)295 (8.1).765Heart failure, n (%)202 (4.8)526 (7.0)<.001179 (4.9)195 (5.4).396Atrial fibrillation and flutter, n (%)200 (4.8)440 (5.9).01184 (5.1)149 (4.1).050Nicotine dependence, n (%)252 (6.0)819 (11.0)<.001244 (6.7)237 (6.5).741Chronic obstructive pulmonary disease, n (%)186 (4.4)534 (7.1)<.001172 (4.7)169 (4.7).868Cerebrovascular diseases, n (%)207 (4.9)449 (6.0).015180 (5.0)165 (4.5).408Coagulation defects, n (%)180 (4.3)336 (4.5).587148 (4.1)140 (3.9).631Osteoporosis without fracture, n (%)224 (5.3)302 (4.0).001165 (4.5)166 (4.6).955Gout, n (%)128 (3.0)454 (6.1)<.001121 (3.3)119 (3.3).896Peripheral vascular disease, n (%)97 (2.3)281 (3.8)<.00190 (2.5)93 (2.6).822Sepsis, n (%)104 (2.5)239 (3.2).02680 (2.2)76 (2.1).746Acute myocardial infarction, n (%)86 (2.0)215 (2.9).00776 (2.1)76 (2.1)1.000Rheumatoid arthritis, n (%)144 (3.4)271 (3.6).572128 (3.5)119 (3.3).56Cerebral infarction, n (%)84 (2.0)177 (2.4).19373 (2.0)66 (1.8).549Peptic ulcer, n (%)61 (1.5)69 (0.9).00944 (1.2)47 (1.3).752Systemic lupus erythematosus, n (%)37 (0.9)50 (0.7).20326 (0.7)28 (0.8).785Human immunodeficiency virus, n (%)10 (0.2)18 (0.2).97410 (0.3)10 (0.3)1.000Nonsteroidal anti-inflammatory drugs, n (%)753 (17.9)2105 (28.1)<.001704 (19.4)713 (19.6).790Body mass index (mean ± standard deviation)40.2 ± 7.940.2 ± 5.8.90940.0 ± 7.840.6 ± 5.8.306Albumin (mean ± standard deviation)3.98 ± 0.44.11 ± 0.4<.0014.06 ± 0.44.09 ± 0.4.685 Patient characteristics in the bariatric surgery and GLP-1 receptor agonist cohorts before and after propensity score matching (BMI >35 kg/m2). Table 2Patient characteristics in the bariatric surgery and no-intervention cohorts before and after propensity score matching (BMI >35 kg/m2).VariableBariatric surgery (before) n = 4210No intervention (before) n = 124,047P valueBariatric surgery (after) n = 4210No intervention (after) n = 4210P valueAge, y (mean ± standard deviation)60.3 ± 8.865.3 ± 9.1<.00160.3 ± 8.860.1 ± 9.2.268Female, n (%)3452 (82.0)79,945 (64.5)<.0013452 (82.0)2772 (65.8)<.001White, n (%)3217 (76.4)98,972 (79.8)<.0013217 (76.4)3206 (76.2).778Black or African American, n (%)786 (18.7)16,726 (13.5)<.001786 (18.7)794 (18.9).823Hispanic or Latino, n (%)227 (5.4)6915 (5.6).611227 (5.4)202 (4.8).215Essential hypertension, n (%)1926 (45.8)55,012 (44.4).0721926 (45.8)1909 (45.3).710Diabetes mellitus, n (%)691 (16.4)19,637 (15.8).308691 (16.4)655 (15.6).284Chronic kidney disease, n (%)186 (4.4)6059 (4.9).971186 (4.4)173 (4.1).098Hyperlipidemia, n (%)1060 (25.2)36,018 (29.0)<.0011060 (25.2)1017 (24.2).277Diseases of liver, n (%)326 (7.7)7753 (6.3)<.001326 (7.7)300 (7.1).280Heart failure, n (%)202 (4.8)4845 (3.9).003202 (4.8)161 (3.8).028Atrial fibrillation and flutter, n (%)200 (4.8)6085 (4.9).647200 (4.8)158 (3.8).023Nicotine dependence, n (%)252 (6.0)8044 (6.5).196252 (6.0)250 (5.9).927Chronic obstructive pulmonary disease, n (%)183 (4.4)5466 (4.4).111183 (4.4)156 (3.7).588Cerebrovascular diseases, n (%)207 (4.9)5852 (4.7).549207 (4.9)192 (4.6).442Coagulation defects, n (%)180 (4.3)3735 (3.0)<.001180 (4.3)140 (3.3).023Osteoporosis without fracture, n (%)224 (5.3)5261 (4.2).001224 (5.3)212 (5.0).555Gout, n (%)128 (3.0)4842 (3.9).004128 (3.0)139 (3.3).494Peripheral vascular disease, n (%)97 (2.3)2996 (2.4).64497 (2.3)88 (2.1).503Sepsis, n (%)104 (2.5)1862 (1.5)<.001104 (2.5)95 (2.3).519Acute myocardial infarction, n (%)86 (2.0)2473 (2.0).82386 (2.0)71 (1.7).227Rheumatoid arthritis, n (%)144 (3.4)3558 (2.9).035144 (3.4)126 (3.0).266Cerebral infarction, n (%)84 (2.0)2237 (1.8).35884 (2.0)77 (1.8).578Peptic ulcer, n (%)61 (1.5)839 (0.7)<.00161 (1.5)57 (1.4).711Systemic lupus erythematosus, n (%)37 (0.9)706 (0.6).00937 (0.9)26 (0.6).164Human immunodeficiency virus, n (%)10 (0.2)202 (0.2).24110 (0.2)10 (0.2)1.000Nonsteroidal anti-inflammatory drugs, n (%)753 (17.9)19,165 (15.5)<.001753 (17.9)740 (17.6).711Body mass index (mean ± standard deviation)40.2 ± 7.936.4 ± 5.4<.00140.2 ± 7.937.7 ± 6.3<.001Albumin (mean ± standard deviation)3.98 ± 0.44.09 ± 0.4<.0014.04 ± 0.44.09 ± 0.4.532 Patient characteristics in the bariatric surgery and no-intervention cohorts before and after propensity score matching (BMI >35 kg/m2). Table 3Patient characteristics in the GLP-1 RAs and no-intervention cohorts before and after propensity score matching (BMI >35 kg/m2).VariableGLP-1 RA (before) n = 7482No intervention (before) n = 124,047P valueGLP-1 RA (after) n = 7478No intervention (after) n = 7478P valueAge, y (mean ± standard deviation)63.3 ± 8.365.3 ± 9.1<.00163.3 ± 8.363.2 ± 8.8.652Female, n (%)5001 (66.8)79,945 (64.5)<.0015000 (66.9)4803 (64.2).001White, n (%)5796 (77.5)98,972 (79.8)<.0015794 (77.5)5777 (77.3).740Black or African American, n (%)1090 (14.6)16,726 (13.5).0081090 (14.6)1056 (14.1).428Hispanic or Latino, n (%)491 (6.6)6915 (5.6)<.001490 (6.6)495 (6.6).869Essential hypertension, n (%)4862 (65.0)55,012 (44.4)<.0014858 (65.0)4858 (65.0)1.000Diabetes mellitus, n (%)3594 (48.0)19,637 (15.8)<.0013590 (48.0)3586 (48.0).948Chronic kidney disease, n (%)742 (9.9)6059 (4.9)<.001741 (9.9)666 (8.9).036Hyperlipidemia, n (%)3512 (46.9)36,018 (29.0)<.0013509 (46.9)3496 (46.8).831Diseases of liver, n (%)1034 (13.8)7753 (6.3)<.0011031 (13.8)1014 (13.6).686Heart failure, n (%)526 (7.0)4845 (3.9)<.001524 (7.0)461 (6.2).038Atrial fibrillation and flutter, n (%)440 (5.9)6085 (4.9)<.001440 (5.9)471 (6.3).289Nicotine dependence, n (%)819 (11.0)8044 (6.5)<.001817 (10.9)774 (10.4).254Chronic obstructive pulmonary disease, n (%)534 (7.1)5466 (4.4)<.001534 (7.1)474 (6.3).050Cerebrovascular diseases, n (%)449 (6.0)5852 (4.7)<.001449 (6.0)381 (5.1).015Coagulation defects, n (%)336 (4.5)3735 (3.0)<.001336 (4.5)311 (4.2).315Osteoporosis without fracture, n (%)302 (4.0)5261 (4.2).393302 (4.0)271 (3.6).187Gout, n (%)454 (6.1)4842 (3.9)<.001454 (6.1)478 (6.4).417Peripheral vascular disease, n (%)281 (3.8)2996 (2.4)<.001281 (3.8)256 (3.4).272Sepsis, n (%)239 (3.2)1862 (1.5)<.001238 (3.2)213 (2.9).232Acute myocardial infarction, n (%)215 (2.9)2473 (2.0)<.001215 (2.9)190 (2.5).208Rheumatoid arthritis, n (%)271 (3.6)3558 (2.9)<.001271 (3.6)249 (3.3).326Cerebral infarction, n (%)177 (2.4)2237 (1.8)<.001177 (2.4)168 (2.3).624Peptic ulcer, n (%)69 (0.9)839 (0.7).01369 (0.9)68 (0.9).932Systemic lupus erythematosus, n (%)50 (0.7)706 (0.6).27150 (0.7)42 (0.6).403Human immunodeficiency virus, n (%)18 (0.2)202 (0.2).1118 (0.2)11 (0.2).193Nonsteroidal anti-inflammatory drugs, n (%)2105 (28.1)19,165 (15.5)<.0012102 (28.1)2105 (28.2).957Body mass index (mean ± standard deviation)40.2 ± 5.836.4 ± 5.4<.00140.2 ± 5.837.4 ± 6.1<.001Albumin (mean ± standard deviation)4.11 ± 0.44.09 ± 0.4<.0014.11 ± 0.44.12 ± 0.4.668 Patient characteristics in the GLP-1 RAs and no-intervention cohorts before and after propensity score matching (BMI >35 kg/m2). Results Bariatric surgery vs GLP-1 receptor agonists (BMI >35 kg/m2) At 90 days postoperatively, the risk of PJI (OR: 1.88; 95% CI: 1.43–2.49), all-cause revision surgery (OR: 1.89; 95% CI: 1.43–2.49), aseptic loosening (OR: 3.14; 95% CI: 1.85–5.36), periprosthetic fracture (OR: 3.35; 95% CI: 1.91–5.87), deep vein thrombosis (OR: 1.85; 95% CI: 1.49–2.30), bleeding (OR: 2.24; 95% CI: 1.95–2.57), myocardial infarction (OR: 1.56; 95% CI: 1.19–2.04), and stroke (OR: 1.32; 95% CI: 1.02–1.71) were significantly greater in the BS cohort compared to the GLP-1 RA cohort (Table 4).Table 4Rate of postoperative outcomes in the bariatric surgery and GLP-1 RA cohorts at 90 days, 2 years, and 5 years (BMI >35 kg/m2).OutcomesBariatric surgery (n = 3635)GLP-1 receptor agonists (n = 3635)Odds ratio (95% confidence interval)P value90 d Revision TKA148 (4.1)80 (2.2)1.89 (1.43-2.49)<.001 Periprosthetic joint infection146 (4.0)79 (2.2)1.88 (1.43-2.49)<.001 Aseptic loosening56 (1.5)18 (0.5)3.14 (1.85-5.36)<.001 Periprosthetic fracture53 (1.5)16 (0.4)3.35 (1.91-5.87)<.001 Deep vein thrombosis (DVT)235 (6.5)131 (3.6)1.85 (1.49-2.30)<.001 Bleeding706 (19.4)353 (9.7)2.24 (1.95-2.57)<.001 Myocardial infarction140 (3.9)91 (2.5)1.56 (1.19-2.04).001 Stroke140 (3.9)107 (2.9)1.32 (1.02-1.71).0332 y Revision TKA100 (2.8)62 (1.7)1.63 (1.18-2.25).003 Periprosthetic joint infection96 (2.6)62 (1.7)1.56 (1.13-2.16).006 Aseptic loosening33 (0.9)a<103.32 (1.63-6.75)<.001 Periprosthetic fracture29 (0.8)13 (0.4)2.24 (1.16-4.32).0135 y Revision TKA128 (3.5)77 (2.1)1.69 (1.27-2.25)<.001 Periprosthetic joint infection126 (3.5)75 (2.1)1.70 (1.28-2.28)<.001 Aseptic loosening52 (1.4)14 (0.4)3.75 (2.08-6.79)<.001 Periprosthetic fracture41 (1.1)15 (0.4)2.75 (1.52-4.98).001Bolded values indicate significance for the P-value less than 0.05.aTriNetX does not report counts of <10. Table 4Rate of postoperative outcomes in the bariatric surgery and GLP-1 RA cohorts at 90 days, 2 years, and 5 years (BMI >35 kg/m2).OutcomesBariatric surgery (n = 3635)GLP-1 receptor agonists (n = 3635)Odds ratio (95% confidence interval)P value90 d Revision TKA148 (4.1)80 (2.2)1.89 (1.43-2.49)<.001 Periprosthetic joint infection146 (4.0)79 (2.2)1.88 (1.43-2.49)<.001 Aseptic loosening56 (1.5)18 (0.5)3.14 (1.85-5.36)<.001 Periprosthetic fracture53 (1.5)16 (0.4)3.35 (1.91-5.87)<.001 Deep vein thrombosis (DVT)235 (6.5)131 (3.6)1.85 (1.49-2.30)<.001 Bleeding706 (19.4)353 (9.7)2.24 (1.95-2.57)<.001 Myocardial infarction140 (3.9)91 (2.5)1.56 (1.19-2.04).001 Stroke140 (3.9)107 (2.9)1.32 (1.02-1.71).0332 y Revision TKA100 (2.8)62 (1.7)1.63 (1.18-2.25).003 Periprosthetic joint infection96 (2.6)62 (1.7)1.56 (1.13-2.16).006 Aseptic loosening33 (0.9)a<103.32 (1.63-6.75)<.001 Periprosthetic fracture29 (0.8)13 (0.4)2.24 (1.16-4.32).0135 y Revision TKA128 (3.5)77 (2.1)1.69 (1.27-2.25)<.001 Periprosthetic joint infection126 (3.5)75 (2.1)1.70 (1.28-2.28)<.001 Aseptic loosening52 (1.4)14 (0.4)3.75 (2.08-6.79)<.001 Periprosthetic fracture41 (1.1)15 (0.4)2.75 (1.52-4.98).001Bolded values indicate significance for the P-value less than 0.05.aTriNetX does not report counts of <10. Rate of postoperative outcomes in the bariatric surgery and GLP-1 RA cohorts at 90 days, 2 years, and 5 years (BMI >35 kg/m2). Bolded values indicate significance for the P-value less than 0.05. TriNetX does not report counts of <10. At 2 years postoperatively, the risk of PJI (OR: 1.56; 95% CI: 1.13–2.16), all-cause revision surgery (OR: 1.63; 95% CI: 1.18–2.25), aseptic loosening (OR: 3.32; 95% CI: 1.63–6.75), and periprosthetic fracture (OR: 2.24; 95% CI: 1.16–4.32) were significantly greater in the BS cohort compared to the GLP-1 RA cohort. At 5 years postoperatively, patients in the BS group had significantly greater odds for developing PJI (OR: 1.70; 95% CI: 1.28–2.28), all-cause revision surgery (OR: 1.69; 95% CI: 1.27–2.25), aseptic loosening (OR: 3.75; 95% CI: 2.08–6.79), and periprosthetic fracture (OR: 2.75; 95% CI: 1.52–4.98) when compared to their counterparts in the GLP-1 RA cohort. BMI subgroup analysis: bariatric surgery vs GLP-1 receptor agonists In the BMI 35–40 kg/m2 subgroup, similar trends were observed. At 5 years, the BS cohort demonstrated significantly increased odds of PJI (OR: 2.15; 95% CI: 1.59–2.92), all-cause revision (OR: 1.92; 95% CI: 1.43–2.57), aseptic loosening (OR: 4.34; 95% CI: 2.42–7.78), and periprosthetic fracture (OR: 3.61; 95% CI: 1.99–6.53) compared to the GLP-1 RA cohort (Table 5).Table 5Rate of postoperative outcomes in the bariatric surgery and GLP-1 RA cohorts at 90 days, 2 years, and 5 years (BMI 35-40 kg/m2).OutcomesBariatric surgery (n = 3812)GLP-1 receptor agonists (n = 3812)Odds ratio (95% confidence interval)P value90 d Revision TKA28 (0.7)25 (0.7)1.12 (0.65-1.93).679 Periprosthetic joint infection61 (1.6)31 (0.8)1.98 (1.28-3.06).002 Aseptic looseninga<10a<10—— Periprosthetic fracturea<10a<10—— Deep vein thrombosis (DVT)81 (2.1)66 (1.7)1.23 (0.89-1.71).212 Bleeding228 (6.0)132 (3.5)1.77 (1.42-2.21)<.001 Myocardial infarction26 (0.7)20 (0.5)1.30 (0.73-2.34).375 Stroke19 (0.5)32 (0.8)0.59 (0.34-1.05).0682 y Revision TKA95 (2.5)56 (1.5)1.71 (1.23-2.39).001 Periprosthetic joint infection105 (2.8)54 (1.4)1.97 (1.42-2.75)<.001 Aseptic looseninga<10a<10—— Periprosthetic fracturea<10a<10——5 y Revision TKA132 (3.5)70 (1.8)1.92 (1.43-2.57)<.001 Periprosthetic joint infection131 (3.4)62 (1.6)2.15 (1.59-2.92)<.001 Aseptic loosening60 (1.6)14 (0.4)4.34 (2.42-7.78)<.001 Periprosthetic fracture50 (1.3)14 (0.4)3.61 (1.99-6.53)<.001Bolded values indicate significance for the P-value less than 0.05.aTriNetX does not report counts of <10. Table 5Rate of postoperative outcomes in the bariatric surgery and GLP-1 RA cohorts at 90 days, 2 years, and 5 years (BMI 35-40 kg/m2).OutcomesBariatric surgery (n = 3812)GLP-1 receptor agonists (n = 3812)Odds ratio (95% confidence interval)P value90 d Revision TKA28 (0.7)25 (0.7)1.12 (0.65-1.93).679 Periprosthetic joint infection61 (1.6)31 (0.8)1.98 (1.28-3.06).002 Aseptic looseninga<10a<10—— Periprosthetic fracturea<10a<10—— Deep vein thrombosis (DVT)81 (2.1)66 (1.7)1.23 (0.89-1.71).212 Bleeding228 (6.0)132 (3.5)1.77 (1.42-2.21)<.001 Myocardial infarction26 (0.7)20 (0.5)1.30 (0.73-2.34).375 Stroke19 (0.5)32 (0.8)0.59 (0.34-1.05).0682 y Revision TKA95 (2.5)56 (1.5)1.71 (1.23-2.39).001 Periprosthetic joint infection105 (2.8)54 (1.4)1.97 (1.42-2.75)<.001 Aseptic looseninga<10a<10—— Periprosthetic fracturea<10a<10——5 y Revision TKA132 (3.5)70 (1.8)1.92 (1.43-2.57)<.001 Periprosthetic joint infection131 (3.4)62 (1.6)2.15 (1.59-2.92)<.001 Aseptic loosening60 (1.6)14 (0.4)4.34 (2.42-7.78)<.001 Periprosthetic fracture50 (1.3)14 (0.4)3.61 (1.99-6.53)<.001Bolded values indicate significance for the P-value less than 0.05.aTriNetX does not report counts of <10. Rate of postoperative outcomes in the bariatric surgery and GLP-1 RA cohorts at 90 days, 2 years, and 5 years (BMI 35-40 kg/m2). Bolded values indicate significance for the P-value less than 0.05. TriNetX does not report counts of <10. In the BMI >40 kg/m2 subgroup, BS was also associated with higher 5-year risks of PJI (OR: 1.85; 95% CI: 1.37–2.51), all-cause revision (OR: 1.64; 95% CI: 1.22–2.22), aseptic loosening (OR: 2.86; 95% CI: 1.67–4.91), and periprosthetic fracture (OR: 2.35; 95% CI: 1.28–4.31) (Table 6).Table 6Rate of postoperative outcomes in the bariatric surgery and GLP-1 RA cohorts at 90 days, 2 years, and 5 years (BMI >40 kg/m2).OutcomesBariatric surgery (n = 3189)GLP-1 receptor agonists (n = 3189)Odds ratio (95% confidence interval)P value90 d Revision TKA31 (1.0)25 (0.8)1.24 (0.73-2.11).421 Periprosthetic joint infection20 (0.6)23 (0.7)0.87 (0.48-1.59).646 Aseptic looseninga<10a<10—— Periprosthetic fracturea<10a<10—— Deep vein thrombosis (DVT)77 (2.4)46 (1.4)1.69 (1.17-2.44).005 Bleeding199 (6.2)131 (4.1)1.55 (1.24-1.95)<.001 Myocardial infarction17 (0.5)18 (0.6)0.94 (0.49-1.84).865 Stroke17 (0.5)33 (1.0)0.51 (0.29-0.92).0232 y Revision TKA92 (2.9)63 (2.0)1.47 (1.07-2.04).018 Periprosthetic joint infection99 (3.1)57 (1.8)1.76 (1.27-2.45).001 Aseptic loosening33 (1.0)15 (0.5)2.21 (1.20-4.08).009 Periprosthetic fracture23 (0.7)14 (0.4)1.65 (0.85-3.21).1385 y Revision TKA115 (3.6)71 (2.2)1.64 (1.22-2.22).001 Periprosthetic joint infection120 (3.8)66 (2.1)1.85 (1.37-2.51)<.001 Aseptic loosening51 (1.6)18 (0.6)2.86 (1.67-4.91)<.001 Periprosthetic fracture35 (1.1)15 (0.5)2.35 (1.28-4.31).005Bolded values indicate significance for the P-value less than 0.05.aTriNetX does not report counts of <10. Table 6Rate of postoperative outcomes in the bariatric surgery and GLP-1 RA cohorts at 90 days, 2 years, and 5 years (BMI >40 kg/m2).OutcomesBariatric surgery (n = 3189)GLP-1 receptor agonists (n = 3189)Odds ratio (95% confidence interval)P value90 d Revision TKA31 (1.0)25 (0.8)1.24 (0.73-2.11).421 Periprosthetic joint infection20 (0.6)23 (0.7)0.87 (0.48-1.59).646 Aseptic looseninga<10a<10—— Periprosthetic fracturea<10a<10—— Deep vein thrombosis (DVT)77 (2.4)46 (1.4)1.69 (1.17-2.44).005 Bleeding199 (6.2)131 (4.1)1.55 (1.24-1.95)<.001 Myocardial infarction17 (0.5)18 (0.6)0.94 (0.49-1.84).865 Stroke17 (0.5)33 (1.0)0.51 (0.29-0.92).0232 y Revision TKA92 (2.9)63 (2.0)1.47 (1.07-2.04).018 Periprosthetic joint infection99 (3.1)57 (1.8)1.76 (1.27-2.45).001 Aseptic loosening33 (1.0)15 (0.5)2.21 (1.20-4.08).009 Periprosthetic fracture23 (0.7)14 (0.4)1.65 (0.85-3.21).1385 y Revision TKA115 (3.6)71 (2.2)1.64 (1.22-2.22).001 Periprosthetic joint infection120 (3.8)66 (2.1)1.85 (1.37-2.51)<.001 Aseptic loosening51 (1.6)18 (0.6)2.86 (1.67-4.91)<.001 Periprosthetic fracture35 (1.1)15 (0.5)2.35 (1.28-4.31).005Bolded values indicate significance for the P-value less than 0.05.aTriNetX does not report counts of <10. Rate of postoperative outcomes in the bariatric surgery and GLP-1 RA cohorts at 90 days, 2 years, and 5 years (BMI >40 kg/m2). Bolded values indicate significance for the P-value less than 0.05. TriNetX does not report counts of <10. Comparison with no weight-loss intervention Patients prescribed GLP-1 RAs generally showed lower risks of surgical outcomes compared to those with no weight-loss intervention across BMI categories at 5 years (Table 7). In the BMI >35 kg/m2 group, GLP-1 RA patients had significantly lower odds of PJI (OR: 0.67; 95% CI: 0.54–0.83), all-cause revision (OR: 0.66; 95% CI: 0.54–0.82), aseptic loosening (OR 0.41; 95% CI: 0.28–0.62), and periprosthetic fracture (OR: 0.65; 95% CI: 0.43–1.00).Table 7Rate of postoperative surgical outcomes in the GLP-1 RA and no-intervention cohorts across BMI ranges at 5 years.OutcomesGLP-1 receptor agonists (n, %)No intervention (n, %)Odds ratio (95% confidence interval)P value(BMI >35 kg/m2) 5 y (n = 7478) Revision TKA148 (2.0)222 (3.0)0.66 (0.54-0.82)<.001 Periprosthetic joint infection144 (1.9)212 (2.8)0.67 (0.54-0.83)<.001 Aseptic loosening34 (0.5)82 (1.1)0.41 (0.28-0.62)<.001 Periprosthetic fracture36 (0.5)55 (0.7)0.65 (0.43-1.00).046(BMI 35-40 kg/m2) 5 y (n = 6321) Revision TKA107 (1.7)169 (2.7)0.63 (0.50-0.81)<.001 Periprosthetic joint infection101 (1.6)162 (2.6)0.62 (0.48-0.79)<.001 Aseptic loosening29 (0.5)64 (1.0)0.45 (0.29-0.70)<.001 Periprosthetic fracture28 (0.4)56 (0.9)0.50 (0.32-0.79).002(BMI >40 kg/m2) 5 y (n = 4858) Revision TKA104 (2.1)136 (2.8)0.76 (0.59-0.98).037 Periprosthetic joint infection103 (2.1)151 (3.1)0.68 (0.52-0.87).002 Aseptic loosening23 (0.5)70 (1.4)0.33 (0.20-0.52)<.001 Periprosthetic fracture28 (0.6)37 (0.8)0.76 (0.46-1.24).263Bolded values indicate significance for the P-value less than 0.05. Table 7Rate of postoperative surgical outcomes in the GLP-1 RA and no-intervention cohorts across BMI ranges at 5 years.OutcomesGLP-1 receptor agonists (n, %)No intervention (n, %)Odds ratio (95% confidence interval)P value(BMI >35 kg/m2) 5 y (n = 7478) Revision TKA148 (2.0)222 (3.0)0.66 (0.54-0.82)<.001 Periprosthetic joint infection144 (1.9)212 (2.8)0.67 (0.54-0.83)<.001 Aseptic loosening34 (0.5)82 (1.1)0.41 (0.28-0.62)<.001 Periprosthetic fracture36 (0.5)55 (0.7)0.65 (0.43-1.00).046(BMI 35-40 kg/m2) 5 y (n = 6321) Revision TKA107 (1.7)169 (2.7)0.63 (0.50-0.81)<.001 Periprosthetic joint infection101 (1.6)162 (2.6)0.62 (0.48-0.79)<.001 Aseptic loosening29 (0.5)64 (1.0)0.45 (0.29-0.70)<.001 Periprosthetic fracture28 (0.4)56 (0.9)0.50 (0.32-0.79).002(BMI >40 kg/m2) 5 y (n = 4858) Revision TKA104 (2.1)136 (2.8)0.76 (0.59-0.98).037 Periprosthetic joint infection103 (2.1)151 (3.1)0.68 (0.52-0.87).002 Aseptic loosening23 (0.5)70 (1.4)0.33 (0.20-0.52)<.001 Periprosthetic fracture28 (0.6)37 (0.8)0.76 (0.46-1.24).263Bolded values indicate significance for the P-value less than 0.05. Rate of postoperative surgical outcomes in the GLP-1 RA and no-intervention cohorts across BMI ranges at 5 years. Bolded values indicate significance for the P-value less than 0.05. In contrast, BS was associated with elevated or similar surgical risks compared to the no-intervention cohort (Table 8). In the BMI >35 kg/m2 group at 5 years, BS patients had significantly higher odds of PJI (OR: 1.33; 95% CI: 1.04–1.70) and periprosthetic fracture (OR: 2.01; 95% CI: 1.24–3.26), but no significant difference was observed for all-cause revision (OR: 1.03; 95% CI: 0.82–1.29). In the BMI >40 kg/m2 subgroup, BS was associated with significantly higher odds of all-cause revision (OR: 1.50; 95% CI: 1.16–1.95) and periprosthetic fracture (OR: 1.92; 95% CI: 1.16–3.19).Table 8Rate of postoperative surgical outcomes in the bariatric surgery and no-intervention cohorts across BMI ranges at 5 years.OutcomesBariatric surgery (n, %)No intervention (n, %)Odds ratio (95% confidence interval)P value(BMI >35 kg/m2) 5 y (n = 4210) Revision TKA155 (3.7)151 (3.6)1.03 (0.82-1.29).816 Periprosthetic joint infection151 (3.6)115 (2.7)1.33 (1.04-1.70).025 Aseptic loosening60 (1.4)53 (1.3)1.13 (0.78-1.65).507 Periprosthetic fracture50 (1.2)25 (0.6)2.01 (1.24-3.26).004(BMI 35-40 kg/m2) 5 y (n = 4750) Revision TKA171 (3.6)149 (3.1)1.15 (0.92-1.44).211 Periprosthetic joint infection161 (3.4)122 (2.6)1.33 (1.05-1.69).019 Aseptic loosening74 (1.6)51 (1.1)1.46 (1.02-2.09).038 Periprosthetic fracture62 (1.3)29 (0.6)2.15 (1.38-3.35).001(BMI >40 kg/m2) 5 y (n = 3927) Revision TKA144 (3.7)97 (2.5)1.50 (1.16-1.95).002 Periprosthetic joint infection141 (3.6)127 (3.2)1.11 (0.87-1.42).384 Aseptic loosening61 (1.6)44 (1.1)1.39 (0.94-2.06).095 Periprosthetic fracture44 (1.1)23 (0.6)1.92 (1.16-3.19).010Bolded values indicate significance for the P-value less than 0.05. Table 8Rate of postoperative surgical outcomes in the bariatric surgery and no-intervention cohorts across BMI ranges at 5 years.OutcomesBariatric surgery (n, %)No intervention (n, %)Odds ratio (95% confidence interval)P value(BMI >35 kg/m2) 5 y (n = 4210) Revision TKA155 (3.7)151 (3.6)1.03 (0.82-1.29).816 Periprosthetic joint infection151 (3.6)115 (2.7)1.33 (1.04-1.70).025 Aseptic loosening60 (1.4)53 (1.3)1.13 (0.78-1.65).507 Periprosthetic fracture50 (1.2)25 (0.6)2.01 (1.24-3.26).004(BMI 35-40 kg/m2) 5 y (n = 4750) Revision TKA171 (3.6)149 (3.1)1.15 (0.92-1.44).211 Periprosthetic joint infection161 (3.4)122 (2.6)1.33 (1.05-1.69).019 Aseptic loosening74 (1.6)51 (1.1)1.46 (1.02-2.09).038 Periprosthetic fracture62 (1.3)29 (0.6)2.15 (1.38-3.35).001(BMI >40 kg/m2) 5 y (n = 3927) Revision TKA144 (3.7)97 (2.5)1.50 (1.16-1.95).002 Periprosthetic joint infection141 (3.6)127 (3.2)1.11 (0.87-1.42).384 Aseptic loosening61 (1.6)44 (1.1)1.39 (0.94-2.06).095 Periprosthetic fracture44 (1.1)23 (0.6)1.92 (1.16-3.19).010Bolded values indicate significance for the P-value less than 0.05. Rate of postoperative surgical outcomes in the bariatric surgery and no-intervention cohorts across BMI ranges at 5 years. Bolded values indicate significance for the P-value less than 0.05. Discussion This study demonstrated that obese patients undergoing primary TKA who were prescribed GLP-1 RAs within 1 year prior to surgery experienced a lower risk of PJI, all-cause revision, aseptic loosening, and periprosthetic fracture at short- and long-term follow-up compared to those managed with BS. Secondary analysis revealed that patients prescribed GLP-1 RAs also had lower rates of aseptic loosening, periprosthetic fracture, PJI, and all-cause revision at 5 years compared to patients who received no weight loss-intervention. These findings were consistent across BMI cohorts. The results reinforce the clinical value of pharmacologic weight loss in the perioperative period and underscore the need for individualized risk stratification and preoperative counseling [29,30]. Obesity is now recognized as a chronic, systemic inflammatory condition that contributes to tissue degeneration, poor wound healing, and metabolic dysfunction [31,32]. These pro-inflammatory effects are mediated through cytokine release, oxidative stress, and dysregulated immune cell activity [33]. Our findings suggest that GLP-1 RAs may help counteract this inflammatory burden through mechanisms beyond weight loss, including reduced NF-κB signaling, improved glycemic control, and attenuation of cytokine-mediated tissue injury [18,34]. These effects likely contribute to the observed reductions in PJI and implant-related complications seen in this study. The no-intervention cohort, lacking both targeted metabolic correction and inflammation modulation, exhibited persistently elevated risks for major complications, further supporting the physiologic benefit of GLP-1 RAs as a preoperative optimization tool [35]. These findings align with results in nondiabetic arthroplasty patients, where GLP-1 RAs were associated with reduced transfusion requirements and no increase in surgical or medical complication rates [36]. Several mechanisms may explain the higher revision and complication rates in the BS cohort. While BS achieves substantial weight loss, the resulting nutritional and metabolic changes can compromise musculoskeletal health and soft tissue healing [37]. Deficiencies in calcium, vitamin D, and protein following BS impair bone remodeling and collagen synthesis, key elements of implant stability and wound healing [38]. Rapid weight loss can also promote bone resorption, increasing the risk of aseptic loosening and periprosthetic fracture [[39], [40], [41], [42]]. These structural changes likely contribute to the higher revision rates observed. Furthermore, protein-calorie malnutrition may lead to impaired immune function, increasing the risk of infection, soft tissue laxity, and delayed healing [43]. These effects align with our observed higher rates of revision, PJI, and periprosthetic fracture in the BS group compared to GLP-1 RA. Although patients referred for BS may present with higher initial BMI than those prescribed GLP-1 RAs, BMI was included as a continuous variable in the PSM, and post-matching BMI was comparable between cohorts (40.0 ± 7.8 vs 40.6 ± 5.8, P = .306). The consistency of findings across BMI subgroups (35–40 kg/m2 and >40 kg/m2) further suggests that baseline BMI severity alone does not account for the observed differences in outcomes. The elevated PJI risk in the BS cohort may also stem from impaired immune response secondary to hypoalbuminemia and nutrient depletion [44,45]. Although albumin was included in the matching, it is important to note that albumin alone may not be the only indicator of malnutrition, and the optimal marker for nutritional status remains unclear [46,47]. Both groups had levels above 3.5 g/dL, suggesting minimal clinical difference [48]. BS may also impact systemic immunity via other mechanisms, including gut microbiome alterations [49]. These factors likely contribute to the nearly twofold increase in PJI risk observed in the BS group at 90 days and underscore the importance of nutritional and immunologic optimization in this population. In contrast, GLP-1 RAs were associated with lower complication rates at most time points. These agents promote gradual weight loss through appetite suppression, delayed gastric emptying, and preserved intestinal absorptive function [50]. Studies have also shown improvements in insulin sensitivity, muscle composition, and reduced ectopic fat accumulation in patients using GLP-1 RAs [51,52]. Anti-inflammatory and pro-reparative effects may further enhance soft tissue healing and immune response [18]. By mitigating chronic low-grade inflammation associated with obesity and preserving muscle integrity relative to BS, GLP-1 RAs may help protect against infection and implant failure, which contributes to the lower rates of revision and PJI seen in this cohort [53]. Additionally, GLP-1 RAs may reduce periarticular adipose tissue, improving joint mechanics and decreasing stress on the implant [35,54]. Unlike BS, which compromises nutrient absorption and hormonal homeostasis, GLP-1 RAs preserve physiological function while achieving metabolic improvement, offering a safer preoperative strategy [55]. Several limitations need to be considered. The retrospective design limits the ability to establish causality, and residual confounding may exist despite robust PSM. Specific data regarding the timing of weight-loss interventions and adherence to pharmacologic therapy, as well as the types, approaches, and risks of different bariatric procedures performed, were not available and could influence the outcomes observed. Surgical technique variations, implant selection, and rehabilitation protocols were also not captured in the database, which are factors that may impact long-term revision risks. The database also lacks information on patient-reported functional outcomes, which could provide further insight into the overall success of different weight-loss strategies on outcome after TKA. While we analyzed instability and dislocation, the ICD-10 codes utilized did not differentiate between global or isolated flexion instability. Additionally, while we utilized CPT codes to identify bariatric procedures, we could not distinguish between specific surgical subtypes (eg, sleeve gastrectomy vs Roux-en-Y gastric bypass) due to sample size limitations, though metabolic impacts may vary between procedures. Despite these limitations, this study benefits from a large, well-matched patient cohort, reducing selection bias and strengthening the validity of the findings. The use of multi-institutional data enhances generalizability, and the comprehensive analysis of multiple complication endpoints provides a thorough comparison of BS vs pharmacologic weight-loss therapy in the context of long-term TKA outcomes. Further research is warranted to optimize preoperative weight-loss strategies for TKA patients. Prospective studies should investigate the ideal timing of weight-loss interventions, the long-term effects of different pharmacologic agents on implant survival, and targeted perioperative strategies to mitigate risks associated with malnutrition and metabolic changes. The TriNetX database does not reliably capture serial BMI measurements or quantitative weight-loss trajectories, and therefore we cannot confirm the magnitude, timing relative to surgery, or durability of weight loss achieved by either intervention. Future prospective studies should evaluate the dose-response relationship between preoperative weight loss and postoperative outcomes. Importantly, the current study cannot differentiate whether the favorable outcomes in the GLP-1 RA cohort are attributable to weight loss itself or to the pharmacologic effects of these agents, including anti-inflammatory, glycemic, and potential osteoprotective properties. It is plausible that medically supervised, nutritionally supported weight-loss programs could yield similar outcomes, especially if comparable metabolic improvements are achieved without the absorptive compromise associated with BS. Future studies should include structured lifestyle intervention cohorts to isolate the independent contribution of weight loss from the pharmacologic effects of GLP-1 RAs. Conclusions BS within 1 year prior to TKA increased the risk of aseptic loosening, periprosthetic fracture, PJI, and all-cause revision surgery at 2 and 5 years after TKA compared to weight-loss therapy with GLP-1 RAs. Our secondary analysis demonstrated that patients on GLP-1 RAs also had lower rates of aseptic loosening, periprosthetic fracture, PJI, and all-cause revision at 5 years compared to patients who received no weight-loss intervention. These findings suggest that for obese patients requiring TKA, GLP-1 RAs may offer a safer preoperative optimization strategy than BS, particularly when joint replacement is planned within the near term. The study underscores the necessity of careful patient selection and individualized preoperative counseling. CRediT authorship contribution statement George Bcharah: Writing – review & editing, Writing – original draft, Project administration, Methodology, Conceptualization. Austin E. Wininger: Writing – review & editing, Writing – original draft, Resources, Data curation, Conceptualization. Zaid Elsabbagh: Writing – review & editing, Writing – original draft, Validation, Methodology, Investigation. Collin L. Braithwaite: Writing – review & editing, Methodology, Investigation, Conceptualization. Paul R. Van Schuyver: Resources, Methodology, Investigation, Conceptualization. Mark J. Spangehl: Writing – review & editing, Validation, Supervision. Henry D. Clarke: Writing – review & editing, Writing – original draft, Supervision, Project administration. Joshua S. Bingham: Writing – review & editing, Supervision. Conflicts of interest M. Spangehl owns stocks or stock options in Sonoran Biosciences; received research support from Stryker and DePuy Synthes as a Principal Investigator; and is a member of the medical/orthopaedic publications editorial/governing board at Journal of Arthroplasty. H.D. Clarke received royalties from Restore 3D/ConforMIS, Zimmer Biomet, and Aspire-ISE/Optimus; is a paid consultant for Restore 3D/ConforMIS and Zimmer Biomet; is an unpaid consultant for OSSO VR; owns stocks or stock options in Aspire-ISE/Optimus; received royalties, financial, or material support from Elsevier; and is a board member/committee appointments for AAOS; Knee Society, AAHKS, and AAOS; all other authors declare no potential conflicts of interest. For full disclosure statements refer to https://doi.org/10.1016/j.artd.2026.102033.