Abstract
This cohort study evaluates the risk of postoperative respiratory complications among patients with diabetes undergoing surgery who had vs those who had not a prescription fill for glucagon-like peptide 1 receptor agonists.
In June 2023, the American Society of Anesthesiologists issued guidelines recommending preoperative withholding of glucagon-like peptide 1 receptor agonists (GLP-1 RAs),1 used for type 2 diabetes management and weight loss. These guidelines, which have been questioned,2 arose from reports of delayed gastric emptying and pulmonary aspiration following induction of anesthesia in patients using GLP-1 RAs. Although preoperative medication guidelines can prevent complications, withholding medications can also result in adverse effects.3 Furthermore, associated logistical burdens can result in surgical cancellations, significant care delays, and financial losses from unused operating room time.4,5 These downsides are pronounced for GLP-1 RAs due to recommended withholding periods of up to a week. Given increasing use of GLP-1 RAs, larger-scale evaluations of their perioperative risks are needed; however, a randomized trial of preoperative GLP-1 RA withholding may not be feasible. Therefore, we used a claims database to evaluate the risk of postoperative respiratory complications among patients with diabetes and a prescription fill for GLP-1 RAs who underwent emergency surgery because these patients would be unlikely to withhold their medication in accordance with guidelines.
Methods
We used administrative claims from the Merative MarketScan Commercial Database, a large national database of about 250 million individuals younger than 65 years enrolled in employer-sponsored health insurance plans. We evaluated all patients with type 2 diabetes and a GLP-1 RA prescription fill who had undergone any of 13 emergency surgeries between January 1, 2015, and December 31, 2021 (eTable 1 in Supplement 1). We restricted our sample to patients who had undergone surgery on the same day as an emergency department visit because these patients would be unlikely to have sufficient time to withhold their GLP-1 RA medication. Because patients with diabetes are at higher risk of postoperative respiratory complications,6 our comparison group comprised patients with diabetes and at least 1 fill for a non–GLP-1 RA antidiabetic agent. The outcome was a composite of aspiration pneumonitis, postoperative respiratory failure, and/or admission to the intensive care unit from 0 through 7 postoperative days (eMethods 1, eTable 2 in Supplement 1).
We assessed differences in characteristics between patients with vs without a GLP-1 RA fill using standardized mean differences (SMD), with SMD greater than 0.1 reflecting meaningful differences between groups. To estimate the association between having a fill for a GLP-1 RA preoperatively and postoperative respiratory complications, we used multivariable logistic regression adjusting for demographic characteristics, indicators of diabetic severity including the Diabetes Complications Severity Index, Elixhauser comorbidities, and fixed effects for year and surgery type (eMethods 1 in Supplement 1). Sensitivity analyses included narrowing the outcome to include only pulmonary aspiration and postoperative respiratory failure; reestimating the model using targeted maximum likelihood estimation, a double-robust method that may better approximate causal effects; and restricting the sample to surgeries with low risk of aspiration and postoperative complications (eMethods 2 in Supplement 1). Statistical inferences were based on 2-sided P < .05.
This study was approved by the Stanford University institutional review board with waiver of consent. Analyses were conducted using STATA version 17.0 and R version 4.2.3.
Results
Our final sample included 23 679 patients, 3502 (14.8%) of whom had a GLP-1 RA fill (Table 1). Those with a GLP-1 RA fill were more likely to be male, use more antidiabetic agents, and have diagnoses of obesity. Overall incidence of postoperative respiratory complications was 3.5% for those with a GLP-1 RA fill and 4.0% for those without (odds ratio [OR], 0.85; 95% CI, 0.70-1.04; P = .12). After adjustment, there was no significant difference in the incidence of postoperative respiratory complications between these 2 groups (adjusted OR, 1.03; 95% CI, 0.82-1.29; P = .80). Findings were robust to alternative specifications (Table 2).
Table 1. Characteristics of the Study Population by Use of Glucagon-Like Peptide 1 Receptor Agonist (GLP-1 RA).
| Use of GLP-1 RA receptor agonist | Standardized mean differencea | ||
|---|---|---|---|
| Yes (n = 3502) | No (n = 20 177) | ||
| Demographic characteristics | |||
| Age, mean (SD), y | 53.6 (8.1) | 53.9 (8.9) | 0.04 |
| Female, No. (%) | 1720 (49.1) | 10 994 (54.5) | 0.11 |
| Male, No. (%) | 1782 (50.9) | 9183 (45.5) | 0.11 |
| Severity of diabetes | |||
| Preoperative use of insulin, No. (%) | 1131 (32.3) | 5649 (28.0) | 0.09 |
| Total No. of agents used to treat diabetes, mean (SD) | 2.4 (1.0) | 1.5 (0.7) | 1.02 |
| DCSI score, mean (SD)b | 1.3 (1.6) | 1.3 (1.7) | 0.03 |
| Elixhauser comorbidities, No. (%) c | |||
| AIDS/HIV | 14 (0.4) | 85 (0.4) | <0.01 |
| Alcohol misuse | 23 (0.7) | 449 (2.2) | 0.13 |
| Blood loss anemia | 67 (1.9) | 391 (1.9) | <0.01 |
| Other anemia | 257 (7.3) | 1528 (7.6) | <0.01 |
| Cardiac arrhythmias | 586 (16.7) | 3666 (18.2) | 0.04 |
| Congestive heart failure | 222 (6.3) | 1550 (7.7) | 0.05 |
| Coagulopathy | 136 (3.9) | 1004 (5.0) | 0.05 |
| Chronic pulmonary disease | 496 (14.2) | 2819 (14.0) | <0.01 |
| Depression | 623 (17.8) | 3114 (15.4) | 0.06 |
| Illicit drug use | 48 (1.4) | 381 (1.9) | 0.04 |
| Fluid and electrolyte disorders | 637 (18.2) | 4357 (21.6) | 0.09 |
| Hypertension, complicated | 405 (11.6) | 2396 (11.9) | 0.01 |
| Hypertension, uncomplicated | 2793 (79.8) | 15 620 (77.4) | 0.06 |
| Hyperthyroidism | 602 (17.2) | 3292 (16.3) | 0.02 |
| Liver disease | 890 (25.4) | 5026 (24.9) | 0.01 |
| Lymphoma | 20 (0.6) | 145 (0.7) | 0.02 |
| Metastatic cancer | 40 (1.1) | 400 (2.0) | 0.07 |
| Obesity | 1817 (51.9) | 8403 (41.6) | 0.21 |
| Other neurological disorders | 100 (2.9) | 850 (4.2) | 0.07 |
| Paralysis | 20 (0.6) | 218 (1.1) | 0.06 |
| Psychoses | 13 (0.4) | 79 (0.4) | <0.01 |
| Chronic peptic ulcer disease | 64 (1.8) | 481 (2.4) | 0.04 |
| Pulmonary circulation disorder | 74 (2.1) | 546 (2.7) | 0.04 |
| Renal failure | 347 (9.9) | 2276 (11.3) | 0.05 |
| Rheumatoid arthritis | 138 (3.9) | 751 (3.7) | 0.01 |
| Solid tumor | 231 (6.6) | 1525 (7.6) | 0.04 |
| Valvular disease | 180 (5.1) | 1256 (6.2) | 0.05 |
| Weight loss | 80 (2.3) | 722 (3.6) | 0.08 |
| Type of surgery, No. (%) | |||
| Laparoscopic appendectomy | 442 (12.6) | 2128 (10.5) | 0.07 |
| Laparoscopic cholecystectomy | 1185 (33.8) | 6594 (32.7) | 0.03 |
| Colectomy for diverticulitis | 58 (1.7) | 695 (3.4) | 0.11 |
| Operative management | |||
| Traumatic hip fracture | 41 (1.2) | 364 (1.8) | 0.05 |
| Adhesive small bowel obstruction | 21 (0.6) | 126 (0.6) | <0.01 |
| Ovarian torsion | 12 (0.3) | 45 (0.2) | 0.02 |
| Testicular torsiond | <11 | <11 | 0.01 |
| Ectopic pregnancy | 11 (0.3) | 52 (0.3) | 0.01 |
| Incarcerated or strangulated hernia | 394 (11.3) | 2263 (11.2) | <0.01 |
| Laparoscopic or open repair of perforated ulcerd | <11 (<0.2) | 38 (0.2) | <0.01 |
| Transurethral intervention | 1243 (35.5) | 7006 (34.7) | 0.02 |
| Upper endoscopy (bleeding peptic or duodenal ulcer) | 34 (1.0) | 391 (1.9) | 0.08 |
| Upper endoscopy (foreign body removal) | 64 (1.8) | 508 (2.5) | 0.05 |
Abbreviation: DCSI, Diabetes Complications Severity Index.
Standardized mean differences (SMD), with SMD greater than 0.1, reflective of meaningful differences between groups.
A severity score ranging from 0 to 13 and validated for use with International Statistical Classification of Diseases and Related Health Problems, Tenth Revision (ICD-10) diagnosis codes in administrative data. It is based on the sum of individual scores of 0 to 2 for diabetes-related complications in 7 categories (cardiovascular, cerebrovascular and peripheral vascular disease, metabolic disease, nephropathy, retinopathy, and neuropathy).
Twenty-eight of 30 comorbidities were included. Diabetes with and without complications were excluded to avoid collinearity with the Diabetes Complications Severity Index. Each Elixhauser comorbidity was defined as a separate binary variable representing whether the patient did or did not have evidence of that comorbidity in their claims history.
Some cell sizes in this row have been suppressed for patient privacy in accordance with the study’s data use agreement.
Table 2. Primary Adjusted Model and Sensitivity Analyses.
| Outcome in glucagon-like peptide 1 receptor agonist exposure, % (95% CI) | Estimate (95% CI) | P value | ||
|---|---|---|---|---|
| Yes | No | |||
| Primary modela | 4.0 (3.3 to 4.7) | 3.9 (3.7 to 4.2) | OR, 1.03 (0.82 to 1.29) | .80 |
| Sensitivity analyses b | ||||
| 1 (prespecified outcome variable)c | 2.2 (1.6 to 2.7) | 2.5 (2.3 to 2.7) | OR, 0.85 (0.64 to 1.14) | .30 |
| 2d,e | 4.2 (4.1 to 4.3) | 3.9 (3.8 to 4.0) | Average treatment effect, %, 0.3 (−1.0 to 1.6) | |
| 3f | 3.0 (2.3 to 3.7) | 2.9 (2.6 to 3.2) | OR, 04 (0.79 to 1.36) | .80 |
Estimated using multivariable logistic regression.
Sensitivity analyses are further described in eMethods 2 in Supplement 1.
Replicated the primary analysis but redefined the outcome as those with a diagnosis of aspiration and/or postoperative respiratory failure (excluding intensive care unit admission).
Reestimated the model using targeted maximum likelihood estimation, a double-robust estimation method.
P value was not applicable for the estimates calculated using targeted maximum likelihood estimation because the 95% CI was calculated using bootstrap resamples.
Replicated the primary analysis but restricted the analytic sample to those who had undergone surgeries considered lower risk of aspiration and/or postoperative complications (laparoscopic appendectomy or cholecystectomy and transurethral interventions; n = 18 585).
Discussion
Preoperative use of GLP-1 RAs in patients undergoing emergency surgery was not associated with a higher risk of postoperative respiratory complications compared with patients not using GLP-1 RAs. Study limitations include being restricted to commercially insured patients, being unable to measure preoperative duration of GLP-1 RA therapy,3 and the lack of information about patient adherence to the medication. In addition, the study did not examine fills of GLP-1 RAs for weight loss alone. Results of this study suggest that liberalizing the withholding guidelines for GLP-1 RAs preoperatively should be considered.
Section Editors: Kristin Walter, MD, and Jody W. Zylke, MD, Deputy Editors; Karen Lasser, MD, Senior Editor.
eTable 1. Diagnostic and Procedural Codes Used to Define Emergency Surgeries Included in Study
eMethods 1. Sample Creation, Definitions of Variables, and Statistical Approach
eTable 2. Diagnostic Codes Used to Define Postoperative Respiratory Complications
eMethods 2. Description of Sensitivity Analyses
eReferences
Data Sharing Statement
References
- 1.Joshi GP, Abdelmalak BB, Weigel WA, et al. American Society of Anesthesiologists consensus-based guidance on preoperative management of patients (adults and children) on glucagon-like peptide-1 (GLP-1) receptor agonists. American Society of Anesthesiologists. June 29, 2023. Accessed October 15, 2023. https://www.asahq.org/about-asa/newsroom/news-releases/2023/06/american-society-of-anesthesiologists-consensus-based-guidance-on-preoperative
- 2.van Zuylen ML, Siegelaar SE, Plummer MP, Deane AM, Hermanides J, Hulst AH. Perioperative management of long-acting glucagon-like peptide-1 (GLP-1) receptor agonists. Br J Anaesth. 2024;132(4):644-648. doi: 10.1016/j.bja.2024.01.001 [DOI] [PubMed] [Google Scholar]
- 3.Bloomgarden Z. Should glucagon-like peptide-1 receptor agonist treatment be withheld in preoperative management? J Diabetes. 2023;15(9):712-713. doi: 10.1111/1753-0407.13469 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Tewfik GL, Rodriguez-Aponte C, Zhang K, Ezzat B, Suri P, Chaudhry F. Outcomes and disposition of patients after case cancellation on day of surgery for reasons attributed to medical or anesthetic care. Anesth Analg. 2022;135(4):845-854. doi: 10.1213/ANE.0000000000006156 [DOI] [PubMed] [Google Scholar]
- 5.Childers CP, Maggard-Gibbons M. Understanding costs of care in the operating room. JAMA Surg. 2018;153(4):e176233. doi: 10.1001/jamasurg.2017.6233 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Ramachandran SK, Nafiu OO, Ghaferi A, Tremper KK, Shanks A, Kheterpal S. Independent predictors and outcomes of unanticipated early postoperative tracheal intubation after nonemergent, noncardiac surgery. Anesthesiology. 2011;115(1):44-53. doi: 10.1097/ALN.0b013e31821cf6de [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eTable 1. Diagnostic and Procedural Codes Used to Define Emergency Surgeries Included in Study
eMethods 1. Sample Creation, Definitions of Variables, and Statistical Approach
eTable 2. Diagnostic Codes Used to Define Postoperative Respiratory Complications
eMethods 2. Description of Sensitivity Analyses
eReferences
Data Sharing Statement