Obesity and osteoarthritis are two common health conditions that affect the aging adult population within the United States. The prevalence of obesity is projected to increase over the next decade with estimates that nearly 70% of patients undergoing primary total joint arthroplasty (TJA) will be obese by 2029. It is well known that obesity is an independent risk factor for the development of osteoarthritis due to the increased loads and joint reactive forces. By contrast, severe functional impairment and pain due to end-stage hip and knee arthritis is often cited as a contributor to poorly controlled weight and limited exercise.
Despite the overwhelming success rates of TJA, obesity is a well-documented risk factor for complications after TJA and has been associated with higher rates of impaired wound healing, infection, deep vein thrombosis, prolonged hospitalization, and revision surgery. Obese patients seeking treatment for their end-stage hip and knee arthritis face a rigorous presurgical clearance process that often includes months to years of dieting, exercise, and medication therapy, with some resorting to bariatric surgery to achieve a body mass index (BMI) within tolerances of their treating surgeon. Similarly, preoperative diabetic control is also critical to minimizing complications after elective procedures. An A1C level of more than 8% is a standard contraindication for elective arthroplasty.
The American Association of Hip and Knee Surgeons has issued guidelines regarding preoperative weight loss. In recent years, it was noted that patients with a BMI >30 kg/m2 were at an increased risk of perioperative complications with further increased risk above BMI >40 kg/m2 and recommended preoperative weight loss below this threshold. However, the most recent guidelines in 2023 state that although patients should be informed of their increased risks when undergoing surgery at their current weight, it is not clear whether postponing TJA for weight reduction improves outcomes and whether the use of absolute BMI thresholds should be discouraged.
Many medications, therapies, and surgical options have targeted the issue of weight loss but either they fail or come with an increased risk of adverse effects. The newest and possibly most promising of these therapies are the glucagon-like peptide-1 (GLP-1) agonists, which were initially intended as a second-line noninsulin therapy for glycemic control among diabetics. This class of medications works by lowering blood glucose levels through multiple mechanisms, including increased insulin secretion, decreased glucagon secretion, and slowing of gastric emptying. The Semaglutide Treatment Effect in People with Obesity Trials, of which eight separate trials have been conducted, have demonstrated substantial weight loss up to 15.2% sustained through a 2-year follow-up period. In a separate randomized controlled trial of nondiabetic patients, semaglutide was found to reduce incidence of death from cardiovascular causes at a mean follow-up of 39 months. The most common reported adverse effects include mild-to-moderate gastrointestinal symptoms, including nausea and vomiting, with one study reporting the incidence as high as 82.2%.
A concern for patients taking GLP-1 agonists is gastroparesis and its potential for inducing perioperative regurgitation and pulmonary aspiration syndrome. Currently, there is no evidence that discontinuation of GLP-1 agonists for a given time (1 to 2 weeks) guarantees normal gastric emptying time in an individual patient. In addition, there is neither evidence nor consensus or societal guidelines on whether discontinuation of GLP-1 agonists before elective surgery will decrease perioperative aspiration risk. In its 2021 consensus statement, the Society for Perioperative Assessment and Quality Improvement advocated continuing GLP-1 agonists to the day of surgery.
There is a paucity of literature involving the safety, outcomes, and complications involving the use of GLP-1 agonists in patients undergoing elective orthopaedic surgery. One study by Magruder et al examined a large national database and demonstrated that among patients undergoing total knee arthroplasty, semaglutide was associated with a decreased risk of sepsis, prosthetic joint infections, and readmissions. However, an increased risk of myocardial infarction, acute kidney injury, pneumonia, and hypoglycemia was observed among 90-day postoperative complications.
As the popularity increases and the number of patients who are prescribed GLP-1 agonists continues to rise, it is increasingly important to understand the implications surrounding their use, especially as it pertains to complications and safety following lower TJA. The purpose of this study was to assess complications, outcomes, and readmissions for total hip arthroplasty (THA) among a cohort of patients taking GLP-1 agonists versus a matched cohort not taking this class of medications. We hypothesize that those patients taking GLP-1 agonists will have an increased perioperative complication rate and readmission rate following elective THA.
Methods
We obtained approval from our institutional review board (IRB #HHC-2023-0010) to perform this study, and no additional funding was obtained. This was a retrospective review of a cohort of THA patients who underwent a primary elective THA between June 2016 and December 2022 at a high-volume orthopaedic hospital. Perioperative details and postoperative clinical outcomes, such as length of stay, 30-day readmission, and return to the emergency department (ED) rates, were collected. Patient demographics, clinical characteristics, medications, and additional surgical history were also collected.
Patients were included in this study if they were between 18 and 89 years and underwent primary elective THA. Exclusion criteria for our study were any nonelective procedure or any revision or conversion THA. A total of 4853 patients were available for review. Patients were sorted by the presence or absence of a GLP-1 agonist on their preoperative medication list. Glucagon-like peptide-1 agonists included dulaglutide, exenatide, liraglutide, and semaglutide.
All patients in our study received spinal anesthesia. In addition, all patients receive a standard multimodal pain management protocol, which includes 15 mg of meloxicam one daily, 15 mg of Ketorolac IV every 6 hours, 975 mg of acetaminophen every six hours, 750 mg of methocarbamol every 6 hours, 5 to 10 mg of oxycodone or 25 to 50 mg of tramadol every 4 hours as needed, and one dose of 10 mg of dexamethasone. No patient in either group had a documented hypersensitivity to opiates. Patients who presented with postoperative nausea and vomiting (PONV) were given 4 mg of ondansetron orally dissolving tablets as needed while in the hospital. If the patient remained nauseous at discharge, they were given a prescription for 4 mg of ondansetron to take at home.
The calipmatch plugin within STATA (Version 17; StataCorp) was used to match each THA patient taking a GLP-1 agonist with up to two THA patients not taking GLP-1 agonists. This matching was performed using suspected confounding variables: aex (exact match), age (exact match), and BMI (within 0.3 points). Primary outcomes were length of stay, return to the ED or readmission within 30 days, inpatient complications (urinary retention, postoperative nausea or vomiting, pulmonary embolism), and postoperative pain levels. Patient charts were also individually assessed for inpatient complications not listed above. In addition, patients completed the Hip Osteoarthritis Outcome Score for Joint Replacement (HOOS-JR) preoperatively and at 12 weeks, 6, and 12 months postoperatively.
Patient demographics and single outcome measures, for example, readmissions or the presence of an inpatient complication, were compared using chi-square tests of proportion. Hip Osteoarthritis Outcome Score for Joint Replacement scores and continuous variables like length of stay were compared using Student t-tests. All analyses were completed using STATA version 17 (StataCorp) or Microsoft Excel 2016 (Microsoft). All results where P < 0.05 were considered to be statistically significant.
Results
A total of 66 patients undergoing unilateral THA and taking a GLP-1 agonist were matched 1:2 with 126 patients who were not taking a GLP-1 agonist. Variations was found in the type and dosage of GLP-1 agonist prescribed with 45% (n = 30) taking semaglutide, 26% (n = 17) prescribed liraglutide, 24% (n = 16) dulaglutide, and 5% (n = 3) exenatide. All GLP-1 agonists prescribed were analyzed together as a single cohort. Patient demographics from both the GLP-1 and no GLP-1 cohorts can be seen in Table 1. No notable differences were observed in race, ethnicity, laterality, length of stay, or 30-day return to ED or readmission rates between the two populations. Inpatient pain levels were also equivalent between the matched cohorts. Inpatient complications were also tabulated. No differences were observed in urinary retention between these two groups. A notable increase was observed in inpatient recorded PONV among the group of patients taking GLP-1 agonists (18.2%) as compared with those patients not taking GLP-1 agonists (6.0%, P = 0.011). In addition, one deep vein thrombosis occurred in the GLP-1 agonist group.
Table 1
Matched Patient Demographics for the Glucagon-Like Peptide-1 and No GLP-1 Cohorts
| Total Hip Arthroplasty Patient Cohort | GLP-1 (n = 66) | Matched (n = 126) | P |
| Male sex | 58% (38) | 58% (73) | 0.961 |
| Age (yrs) | 67.9 ± 8.8 | 66.3 ± 7.0 | 0.978 |
| Race | |||
| American Indian/Alaska native | 0% (0) | 0% (0) | 0.057 |
| Asian | 0% (0) | 0% (0) | |
| Black or African American | 18.2% (12) | 7.9% (10) | |
| Native Hawaiian/Pacific Islander | 0% (0) | 0% (0) | |
| Other | 4.5% (3) | 3.2% (4) | |
| White or Caucasian | 75.8% (50) | 88.9% (112) | |
| Unknown/patient refused | 1.5% (1) | 0% (0) | |
| Ethnicity | |||
| Hispanic or Latino | 4.5% (3) | 2.4% (3) | 0.637 |
| Not Hispanic or Latino | 93.9% (62) | 96.8% (122) | |
| Unknown/patient refused | 1.5% (1) | 0.8% (1) | |
| BMI (kg/m2) | 35.0 ± 4.6 | 34.7 ± 4.5 | 0.702 |
| Laterality | 42.4% Left (28) | 44.4% Left (56) | 0.789 |
| Length of stay (days) | 2.59 ± 0.89 | 2.40 ± 0.80 | 0.144 |
| Return to ED within 30 days | 9.1% (6) | 4.8% (6) | 0.239 |
| Readmission within 30 days | 3.0% (2) | 3.2% (4) | 0.956 |
| Complication (inpatient) | |||
| Urinary retention | 10.6% (7) | 8.0% (10) | 0.536 |
| Emesis | 18.2% (12) | 6.0% (8) | 0.011 |
| DVT/PE | 1.5% (1) | 0% (0) | |
| Pain (inpatient) | |||
| Active | 3.58 ± 1.7 | 3.26 ± 1.6 | 0.198 |
| At rest | 5.05 ± 1.6 | 4.75 ± 1.7 | 0.226 |
Hip Osteoarthritis Outcome Score for Joint Replacement outcome scores were compared between the two cohorts at four time points: preoperatively, at 12 weeks, 6, and 12 months postsurgery. No statistically significant differences was observed between these two groups at any time point (Pre-Op. P = 0.125, 12 weeks, P = 0.773, 6 months, P = 0.170, 12 months, P = 0.138; see Table 2), although the GLP-1 group trended toward lower outcome scores and did not show an improvement beyond 12 weeks.
Table 2
Hip Osteoarthritis Outcome Scores for the Glucagon-Like Peptide-1 and No GLP-1 Cohorts
| HOOS | Pre-Op | 12 Weeks | 6 Months | 12 Months |
| GLP1 | 48.07 ± 10.6 | 77.68 ± 15.3 | 76.88 ± 17.5 | 76.83 ± 18.5 |
| No GLP1 | 52.51 ± 13.7 | 76.67 ± 14.1 | 82.52 ± 14.2 | 83.49 ± 16.5 |
Return to ED and readmissions were further assessed in the GLP-1 and non-GLP-1 matched cohorts to determine the reason for return. Of six patients who returned to the ED in the GLP-1 agonist group, three presented with the chief report of nausea and vomiting, which required admission in two patients. The other three patients returned for leg swelling, shortness of breath, and a superficial wound infection. In the group not taking GLP-1 agonists, patients returned to the ED for a variety of reasons: two hip fractures, one periprosthetic joint infection, one spinal headache, one bleeding gastric ulcer, and one case of anemia. The increased occurrence of nausea and vomiting in the GLP-1 cohort causing return to the ED and readmission was not statistically significant, due to the small sample size (fisher exact = 0.182).
Conclusion
This is one of the first studies to report on postoperative complications and outcomes of GLP-1 agonists among patients undergoing THA. Our study found that among patients taking GLP-1 agonists who undergo primary THA, there is a notable risk of early PONV, which occurred before discharge. A post hoc power analysis showed 73% power with an α of 0.05, even with the small sample size. Although not statistically significant, a higher rate of return visits to the ED within 30 days was observed among the GLP-1 cohort, of which half were due to nausea or vomiting. Meanwhile, none of the patients in the matched non-GLP-1 cohort visited the ED due to nausea or vomiting. No difference was observed in the length of stay, readmission rates, or other complications, including urinary retention and venous thromboembolism among the two cohorts.
All patients undergoing primary THA completed the HOOS-JR outcome scores preoperatively and at 12 weeks, 6 months, and 1 year postoperatively. Although the non–GLP-1 cohort reported slightly higher outcome scores at the 6 month and 1 year period, this difference was not notable. A post hoc power analysis showed 69% power with α = 0.05 at 1 year; 80% power would require 91 patients in the GLP-1 group and 182 patients in the non–GLP-1 cohort. As GLP-1 agonists are being more commonly prescribed, subsequent studies can include larger numbers of patients to determine whether the trends seen in HOOS-JR scores reach statistical significance at any time point.
This study has a few important limitations. It is a retrospective cohort study conducted at a single institution with a relatively small sample size. None of the patients in our cohort were discharged on the same day; thus, we are unable to state whether the PONV, which was higher in the GLP-1 cohort, would have influenced potential same-day discharge. At the time these data were obtained, no standardized prescribing practices or anesthetic considerations were found for patients taking GLP-1 agonists before surgery; as a result, it is not known exactly when patients stopped their medications before surgery. Information extracted from the perioperative nursing visit tells us that nearly all patients were instructed to “take [their GLP-1 agonist] as directed” up until the day of surgery. This has important clinical consequences because it has now been determined that there may be an increased risk of aspiration or gastrointestinal side effects if these medications are not stopped within the appropriate window. Another limitation of this study is that the authors do not know for what indication, the patient was prescribed their GLP-1 agonist, whether for diabetic control or weight loss, and the duration they had been taking this medication before surgery. In addition, our study combines all GLP-1 agonist medications into a single cohort, thus limiting information that may be pertinent to the dosage, route, and specific mechanism of action of each drug. It is well described in the literature that the gastrointestinal side effects are predominantly associated with early use and are typically dose related. As such, it is not possible to determine whether the PONV is directly related to the dosing and timing of these medications. Unfortunately, postoperative nausea and vomiting is a fairly common report with multiple potential sources that may include factors such as pain control, anesthetic use, perioperative glucocorticoid use, preoperative hemoglobin, or patient-specific risk factors, such as smoking history or a history of motion sickness, migraines, or previous documentation of PONV. The retrospective nature of our study limited our ability to identify and control for these other potential causes of PONV. Although no notable effect was observed on readmissions and outcomes, GLP-1 use was associated with a markedly increased risk of early PONV among our study cohort. Our study cohort included patients undergoing inpatient THA, with an average length of stay of 2.59 and 2.40 days. Although GLP-1 agonist use among our study cohort did not affect the length of stay or clinical outcomes, future prospective studies should assess the rate of PONV especially among patients scheduled for same-day discharge because this may represent a potential barrier to discharge. In addition, future studies should assess complications and outcomes of stopping weekly dosed GLP-1 agonists one week before surgery and examine complications associated with specific GLP-1 agonists and dosages.
Given the increasing popularity among patients and a surgeon's desire to optimize diabetes and weight before elective joint arthroplasty, it is expected to see a greater proportion of elective TJA patients taking GLP-1 agonists in the near future. It is imperative to understand the complications and risks associated with the use of GLP-1 agonists as this will guide clinical decision making and enable surgeons to better counsel their patients preoperatively.
Acknowledgments
The authors thank Angie Alban, BS, for her assistance in editing this manuscript.
References
1.
Foreman CW, Callaghan JJ, Brown TS, Elkins JM, Otero JE: Total joint arthroplasty in the morbidly obese: How body mass index ≥40 influences patient retention, treatment decisions, and treatment outcomes. J Arthroplasty 2020;35:39-44.2.
Mayfield CK, Mont MA, Lieberman JR, Heckmann ND: Medical weight optimization for arthroplasty patients: A primer of emerging therapies for the joint arthroplasty surgeon. J Arthroplasty 2024;39:38-43.3.
Samson AJ, Mercer GE, Campbell DG: Total knee replacement in the morbidly obese: A literature review. ANZ J Surg 2010;80:595-599.4.
Amanatullah DF, Ohanisian L, Bailony R: Medications available for weight reduction in elective total joint arthroplasty. JBJS Rev 2020;8:e0123.5.
Malinzak RA, Ritter MA, Berend ME, Meding JB, Olberding EM, Davis KE: Morbidly obese, diabetic, younger, and unilateral joint arthroplasty patients have elevated total joint arthroplasty infection rates. J Arthroplasty 2009;24(6 suppl):84-88.6.
Hannon CP, Goodman SM, Austin MS, et al.: 2023 American College of Rheumatology and American Association of Hip and Knee Surgeons Clinical Practice Guideline for the optimal timing of elective hip or knee arthroplasty for patients with symptomatic moderate-to-severe osteoarthritis or advanced symptomatic osteonecrosis with secondary arthritis for whom nonoperative therapy is ineffective. Arthritis Care Res (Hoboken) 2023;75:2227-2238.7.
Jacofsky DJ, Springer BD, Mont MA, Ushakumari DS, Sladen RN. The impact of glucagon-like peptide-1 agonists on hip and knee arthroplasty and perioperative considerations. J Arthroplasty. 2023 8:S0883-S5403.8.
Garvey WT, Batterham RL, Bhatta M, et al.; STEP 5 Study Group: Two-year effects of semaglutide in adults with overweight or obesity: The STEP 5 trial. Nat Med 2022;28:2083-2091.9.
Lincoff AM, Brown-Frandsen K, Colhoun HM, et al.; SELECT Trial Investigators: Semaglutide and cardiovascular outcomes in obesity without diabetes. N Engl J Med 2023;389:2221-2232.10.
Gulak MA, Murphy P: Regurgitation under anesthesia in a fasted patient prescribed semaglutide for weight loss: A case report. Can J Anaesth 2023;70:1397-1400, English.11.
Silveira SQ, da Silva LM, de Campos Vieira Abib A, et al.: Relationship between perioperative semaglutide use and residual gastric content: A retrospective analysis of patients undergoing elective upper endoscopy. J Clin Anesth 2023;87:111091.12.
Pfeifer KJ, Selzer A, Mendez CE, et al.: Preoperative management of endocrine, hormonal, and urologic medications: Society for Perioperative Assessment and Quality Improvement (SPAQI) consensus statement. Mayo Clin Proc 2021;96:1655-1669.13.
Magruder ML, Yao VJH, Rodriguez AN, Ng MK, Sasson V, Erez O: Does semaglutide use decrease complications and costs following total knee arthroplasty? J Arthroplasty 2023;38:2311-2315.e1.