Perioperative Management of Patients Taking Glucagon-like Peptide-1 Receptor Agonists: Applying Evidence to Clinical Practice

The clinical benefits of GLP-1 receptor agonists are derived partly through their propensity to slow down gastrointestinal motility. In the outpatient setting, reported gastrointestinal adverse events related to GLP-1 receptor agonists have risen as their utilization has surged.²  Most commonly reported side effects are abdominal pain (57.6%), nausea and vomiting (23.4%), constipation (30.4%), and diarrhea (32.7%).³  Gastrointestinal side effects occur for both short- and long-acting preparations, and for subcutaneously and orally administered preparations. Typically, these adverse effects are not serious, are self-limited, and are more pronounced at the initiation of the drug or during dose escalation phases. They often subside when maintenance doses are reached.

GLP-1 receptor agonists’ ability to slow gastric emptying to a gastroparesis level has translated into concerns for perioperative aspiration.^4,5  Multiple case reports, case series, and studies have been published since 2022.^6–21  These concerns led the American Society of Anesthesiologists (ASA; Schaumburg, Illinois) to publish a consensus-based guidance suggesting that patients discontinue GLP-1 receptor agonists for 1 day for short-acting preparations and 1 week for long-acting preparations, while observing the current ASA fasting guidelines without any modifications.²²  However, the ASA recommendations have been questioned because of the lack of data supporting them and new data that are rapidly emerging.

There are several GLP-1 receptor agonists currently available (table 1). Short-acting preparations include lixisenatide and exenatide. The lixisenatide-alone formulation was withdrawn from the United States market in January 2023 but remains available in Europe. Lixisenatide is currently available only in combination with long-acting insulin in the United States. Exenatide is available as a short-acting twice-daily subcutaneous preparation or as a long-acting weekly formulation with the drug slowly released from a matrix. Other long-acting GLP-1 receptor agonists include liraglutide, dulaglutide, and semaglutide. Liraglutide is mostly albumin-bound, thus having a reservoir for prolonged release, which accounts for its long half-life.²⁶  Dulaglutide is a GLP-1 molecule coupled with an immunoglobulin Fc fragment leading to slow degradation. Semaglutide is available in two forms: a weekly long-acting preparation, which is albumin-bound, and an oral form with low bioavailability after ingestion leading to the need for daily dosing.²⁶ 

Tirzepatide is a dual GLP-1 receptor agonist and glucose-dependent insulinotropic polypeptide (GIP) dosed weekly and approved for treatment of diabetes mellitus and obesity. The various GLP-1 receptor agonists and their pharmacologic properties are presented in table 1. Orforglipron, a daily oral GLP-1 receptor agonist, and retatrutide, a triple GLP-1, GIP, and glucagon receptor agonist administered subcutaneously, are in late phases of development.^27–29 

General principles guiding perioperative medication management include a thoughtful assessment of the risks and benefits of continuing or discontinuing a medication based on the indication, pharmacokinetics, and potential interactions with anesthetic agents. In certain circumstances, the planned procedure and the anesthetic are important considerations. Medications are withheld perioperatively in select circumstances when risks are perceived to outweigh benefits and no alternative measures can mitigate the periprocedural risks.

In patients with diabetes mellitus, GLP-1 receptor agonists reduce glycemia through multiple mechanisms such as slowing gastric emptying, increasing pancreatic insulin secretion, decreasing glucagon release, and decreasing appetite.^30,31  In the outpatient setting, cardiovascular outcome trials and metanalyses demonstrate that GLP-1 receptor agonists reduce major adverse cardiovascular events, including acute myocardial infarction, stroke, and mortality in individuals with multiple cardiovascular risk factors and either type 2 diabetes or obesity.^32–34  The American Diabetes Association (Arlington, Virginia) and the American Association of Clinical Endocrinology (Jacksonville, Florida) recommend GLP-1 receptor agonists as first-line drugs along with metformin for patients with type 2 diabetes with an established or high risk of atherosclerotic cardiovascular disease, stroke, transient ischemic attacks, or chronic kidney disease.^35,36 

There are emerging data of the benefits of continuing GLP-1 receptor agonists during the perioperative period specifically in patients with diabetes mellitus.

• Studies in cardiac surgery show glycemia benefit of GLP-1 receptor agonists (liraglutide and exenatide) when compared to insulin, and delayed initiation and lower doses of insulin needed for glycemic control^37–42  (table 2).

• Benefits of GLP-1 receptor agonist such as better glycemia control and fewer prosthetic joint infections have been described in patients having noncardiac surgeries. These effects are possibly explained by the immunomodulatory and anti-inflammatory effects of GLP-1 receptor agonists^43,44  (table 2).

• There is a lack of data on the benefits of continuing GLP-1 receptor agonists in patients with obesity, because all the studies to date have been performed in patients with diabetes mellitus.

The perioperative risks of continuing GLP-1 receptor agonists pertain to their propensity to delay gastric emptying, and concerns for perioperative aspiration of gastric contents. Self-reported symptomatology such as nausea and vomiting suggests delayed gastric emptying.⁴⁵  However, lack of symptoms correlate poorly with gastric emptying. Data show that even when patients report no symptoms, they can still have significantly altered gastric emptying times.⁴⁵  Evidence of delayed gastric emptying from GLP-1 receptor agonist comes from case reports and case series (table 3) and from outpatient gastric emptying studies (table 4).^{6–21,46–88} 

A summary of studies and case reports of aspiration events and residual gastric contents assessed by ultrasonography or endoscopy is shown in table 3.

• It is notable that in these reports, the patients followed current recommendations for fasting (i.e., 8 h fasting time for solids and 2 h for liquids).⁸⁹ 

• Although the number of aspiration events is small, the vast majority occurred in patients receiving monitored anesthesia care. In a retrospective database review of patients having esophagogastroduodenoscopy, the only two aspiration events occurred in patients receiving monitored anesthesia care, and aspiration rates were comparable to those seen in the general population of patients receiving anesthesia.¹⁹ 

• The risk of aspiration does not appear to be increased in patients having emergency surgeries who have not had time to stop their GLP-1 receptor agonist, which may be partly due to the use of rapid sequence induction and intubation.²⁰  Also, general endotracheal anesthesia seems to protect against the risk of aspiration.⁴⁸ 

While the risk of aspiration is present and should not be underappreciated, the evidence for delayed gastric emptying primarily comes from outpatient data. The data on gastric emptying in patients on GLP-1 receptor agonist are heterogeneous in nonprocedural and nonsurgical settings (table 4), and interpretation is nuanced.

• These data are mixed with respect to the method used to assess gastric emptying; the studied population, which included patients with obesity, diabetes, or healthy volunteers; the reported results; and perioperative relevance.

• The most accurate assessment of gastric emptying is the nuclear scintigraphic method, which evaluates solid food retention in the stomach. The gastric motility breath test is another infrequently used study to measure solid content transit times. However, few studies have assessed gastric emptying in patients taking GLP-1 receptor agonist using these methods. A recent metanalysis of studies looking at GLP-1 receptor agonist effects on gastric emptying reported prolonged times required for 50% of ingested gastric contents to leave the stomach. Patients taking GLP-1 receptor agonist had a time required for 50% of ingested gastric contents to leave the stomach of 138.4 min (95% CI, 74.5 to 202.3) versus 95.0 min (95% CI, 54.9 to 135.0) for the placebo group, with a pooled mean difference of 36.0 min (P < 0.01).⁸⁸  This evidence supports a delayed gastric emptying of solid foods being decreased in patients taking GLP-1 receptor agonists.

• The majority of studies have used an acetaminophen absorption test, which assesses gastric emptying of liquids rather than solids. Consequently, many experts consider data obtained from acetaminophen absorption studies inadequate to base clinical decisions ofor patients on GLP-1 receptor agonists.⁸⁸  A metanalysis of 10 studies with 411 individuals using an acetaminophen test showed no significant delay in gastric emptying at 1, 4, and 5 h in patients taking GLP-1 receptor agonist.⁹⁰  These data point toward normal emptying of liquids in these patients.

• There is a difference between short- and long-acting drugs and their effects on gastric emptying. The plasma concentration of short-acting drugs peaks and decreases rapidly. This intermittent stimulation of the GLP-1 receptor may impact gastric emptying more than occurs with long-acting drugs, which continuously stimulate receptors, resulting in tachyphylaxis. It may take weeks to months for patients on short-acting medications to develop tachyphylaxis compared to 4 to 5 weeks for long-acting preparations.⁷³ 

• Data on tachyphylaxis for long-acting preparations are mixed as well, with some studies showing that even long-acting preparations (liraglutide, weekly semaglutide, long-acting exenatide) significantly delay gastric emptying.^62,65,67  These data are corroborated by a metanalysis reporting that the type of GLP-1 receptor agonist did not correlate with the gastric emptying effects regardless of the type of assessment (scintigraphy or acetaminophen absorption).⁹⁰ 

• Gastric emptying studies have not assessed effects beyond 4 to 6 h. The limited evidence from emptying studies using scintigraphy have detected a prolongation (138 vs. 95 min) of gastric emptying measured at 4 h, which is much less than the recommended fasting time for solids, but longer than that recommended for liquids.

• Gastric emptying effects are more pronounced in the first few weeks after initiating GLP-1 receptor agonists or escalating doses. During this time, gastric emptying for liquids can be prolonged as assessed by acetaminophen tests.^13,87 

Overall, while the current gastric emptying studies may be difficult to interpret and extrapolate to the perioperative setting, the data point toward a prolonged gastric emptying time for solids. In addition, the presence of residual gastric contents directly observed via endoscopy in patients fasting for solids between 8 and 18 h provides irrefutable evidence that gastric emptying time for solids is affected, regardless of the duration of therapy or patients being in a maintenance phase.^12,21  While effects on gastric emptying for solids are more pronounced with short-acting preparations and long-acting GLP-1 receptor agonists during therapy initiation or dose escalation, some studies show delays in gastric emptying even during the maintenance phase. Emptying time for liquids may not be as significantly impacted with long-acting preparations during steady state due to tachyphylaxis. However, liquid emptying remains slightly prolonged with short-acting preparations, and during start of therapy or dose escalation periods with long-acting GLP-1 receptor agonists.^13,86 

The potential benefit of stopping GLP-1 receptor agonists for only one pharmacologic half-life has come under scrutiny. There are several considerations that counter the benefit of stopping the GLP-1 receptor agonist, including the following:

• Pharmacokinetic data demonstrate that there is residual circulating drug after missing one dose of both short- and long-acting GLP-1 receptor agonists.

  • Pharmacokinetic modeling of liraglutide shows that at the end of one half-life of elimination, the blood concentration is higher in patients taking 1.8 mg compared to the peak concentration of liraglutide 0.6 mg²⁴  (fig. 1).

  • The simulated semaglutide concentration dosed at 1 mg showed that when missing one dose, there was a 48% decrease in minimum concentration expected before the next planned dose 7 days later. Delaying the dose 5 days causes minimum semaglutide concentrations to be 37% lower and the maximum concentration after the next dose to be 14% higher compared with the same subject at a normal weekly steady state. In both cases, the nadir of the semaglutide concentrations is higher than the mean concentration at steady state of a similar patient taking 0.5 mg weekly, and closer to regular steady state after 3 weeks⁹¹  (fig. 2).

  • Similarly, if a dose of dulaglutide or tirzepatide is missed, there is residual circulating GLP-1 receptor agonist. Moreover, if a dose of dulaglutide is not taken within 3 days or if tirzepatide is held for 4 days, there are transient 20% higher concentrations after the subsequent doses of the GLP-1 receptor agonist^25,92  (figs. 3 and 4).

  • Package inserts advise skipping the missed dose and administering the GLP-1 receptor agonist on the regularly scheduled day if fewer than 3 days remain until the next dose for exenatide, dulaglutide, or tirzepatide, or fewer than 2 days for semaglutide.⁹³ 

• More importantly, clinical data confirm that even when a GLP-1 receptor agonist dose is omitted, some patients have residual gastric contents on the day of procedure while following standard fasting guidelines.⁸⁹ 

  • A prospective study of 124 patients presenting for surgeries found ultrasonographic confirmation of increased gastric contents defined by the presence of solids, thick liquids, or more than 1.5 ml/kg clear liquids in patients taking GLP-1 receptor agonists. Interestingly, there was no association between the duration of drug interruption and the prevalence of increased gastric contents, including in some patients who had stopped their drugs more than 7 days before surgery.²¹ 

  • Another study of 404 patients found no relationship between the interval interruption of semaglutide and the presence of residual gastric contents in patients having endoscopies. The interruption intervals of semaglutide in patients without retained gastric contents were 10 days (6 to 15) compared to 11 days (7.75 to 12.5) in those with delayed emptying (P = 0.67).¹² 

Considerations of risks associated with stopping GLP-1 receptor agonists are the following:

• In patients with diabetes mellitus prescribed GLP-1 receptor agonists, no glycemic data are available on the effects of stopping GLP-1 receptor agonists. It is likely that glycemia control will deteriorate, especially in those with poor control, requiring multidrug regimens, taking insulin, or on high doses of antihyperglycemics. Moreover, acute hyperglycemia slows gastric emptying, so it is possible that this will negate the benefit of stopping the drugs.⁹⁴ 

• Patients taking high doses of GLP-1 receptor agonists (more common in obese patients, but also in patients with diabetes) and those experiencing severe symptoms at initiation or during escalation of treatment are at high risk of adverse gastrointestinal effects when doses are missed and then restarted.⁹³  These patients often need to be started on a lower dose of drug to mitigate side effects. However, this is logistically difficult since some of these medications (e.g., tirzepatide, dulaglutide) are packaged in prefilled syringes, and patients are provided with precise doses monthly. In contrast, semaglutide pens allow adjustment of doses.⁹³ 

Both pharmacokinetic and clinical data indicate that short interruptions of long-acting GLP-1 receptor agonist (i.e., one half-life) are not sufficient to ensure complete clearance of the drug. While no data are available on gastric emptying from residual GLP-1 receptor agonist concentrations, it is likely that there is no complete dissipation of the effect. The longer duration of interruption needed (i.e., 4 or 5 half-lives) is impractical, potentially harmful, and not patient-centric.

Moreover, there are data suggesting safety of continuation of GLP-1 receptor agonists periprocedurally. A study in patients having endoscopy while continuing their medications and following a modified fasting regimen reported no increased gastric volumes and no aspiration events.⁵³ 

Symptomatic patients describing severe gastrointestinal symptoms may represent a special category where the medication regimen can be interrupted or dose de-escalated. However, even in these patients, current guidelines recommend dietary interventions such as small frequent meals and consumption of bland foods as a first-line intervention.^95,96  Medication adjustments are recommended only when symptoms do not resolve.^95,96 

Current guidance suggests that the standard fasting time is sufficient to mitigate aspiration risk in patients taking GLP-1 receptor agonists.²²  However, this approach is questionable based on clinical data (table 3).

Considerations for solid fasting times are the following:

• Many patients found to have residual gastric contents on the day of their procedures have observed the standard ASA fasting times.^12,21  In these studies, despite interrupting GLP-1 receptor agonist for more than 7 days, patients had retained gastric contents.^12,21 

• Data from a case control study showed lower rates of persistent stomach contents in patients having afternoon procedures with prolonged fasting compared to those having procedures in the morning (4% vs. 11%; P < 0.01).¹⁷ 

• Several studies reveal that patients having colonoscopies (with or without upper endoscopies) have a lower incidence of retained gastric contents.^12,16,17,47  This is most likely due to the fact that patients having colonoscopies typically follow a clear liquid diet the day before the procedure, and less likely due to the concurrent bowel preparation.⁹⁷ 

The evidence suggests that an extended fasting time for solids should be recommended for patients taking GLP-1 receptor agonists. Perhaps the best data in support of this approach come from a study of 57 patients having endoscopies where GLP-1 receptor agonists were not stopped. All of the patients followed a liquid diet for 24 h or more and were nil per os (nothing by mouth; NPO) for more than 12 h before their endoscopies. No patients had detectable residual gastric contents determined by endoscopy, and no aspiration events occurred.⁵³  Overall, there is clear evidence of delayed gastric emptying for solids irrespective of the type of GLP 1 receptor agonist preparation or duration of treatment, suggesting that a clear liquid diet for 24 h or more may be necessary for these patients.

At this time, the best duration of fasting from liquids is unclear because data are scarce and heterogeneous.

• Various liquids empty from the stomach at different rates. High-caloric liquids (those containing fat or greater than 10% glucose) empty more slowly than water or low caloric clear liquids (10% glucose or less) in patients with normal gastric emptying.⁹⁸ 

• Patients on short-acting preparations have delayed gastric emptying for liquids after weeks of treatments. Significant liquid intragastric content has been described at 4 h after 8 weeks of treatment with lixisenatide with consumption of high-caloric glucose–containing liquids.^73–76 

• In individual studies, patients on long-acting preparations in a maintenance phase at least 4 weeks from initiation seem to have slightly prolonged delays in gastric emptying for liquids in the first 1 to 2 h that normalize by 5 h.^84–86  However, in the majority of these studies, acetaminophen elixir was given after a solid meal, which likely confounds the data and may overestimate the delay (table 4).

• Patients who recently started or dose-escalated long-acting GLP-1 receptor agonist within the previous 4 weeks do not exhibit tachyphylaxis and may need more prolonged NPO fasting times. A study of 20 patients detected that 30% of the semaglutide group (most of whom had been on therapy for less than 4 weeks) compared to 90% of the controls were considered to have empty stomachs 2 h after drinking 12 oz of clear liquids in a supine position.¹³  These findings corroborate other data on gastric emptying assessed by acetaminophen absorption.⁸⁷ 

• A recently published meta-analysis did not show any significant delay in gastric emptying for liquids regardless of the GLP-1 receptor agonist. However, it did not include all of the studies, and there was heterogenicity in the reported measurements and in the nature of the meals coadministered with the acetaminophen.⁹⁰ 

There are concerns that prolonged fasting for liquids is not patient friendly and counters the principles of enhanced recovery. High-caloric carbohydrate drinks have been shown to empty slower than water even in patients not taking GLP-1 receptor agonists. They are generally not recommended in patients with diabetes where the risks of hyperglycemia may outweigh their benefits.⁹⁹  They may be restricted for longer than 4 h in patients on GLP-1 receptor agonists for obesity, because their emptying is particularly delayed, especially by short-acting GLP-1 receptor agonist.^73–76 

Extended restriction periods for other non- or low-caloric clear liquids may not be advantageous or necessary for the majority of patients on maintenance doses of long-acting GLP-1 receptor agonists for more than 4 weeks. However, not all patients exhibit tachyphylaxis.⁶¹  The evidence suggests fasting for non- or low-caloric clear liquids for longer than current recommendations is necessary for patients taking long-acting preparations during initiation and dose escalation periods, and those on short-acting GLP-1 receptor agonists.

Point-of-care gastric ultrasonography has been used to identify patients taking GLP-1 receptor agonists who have residual gastric content preprocedurally.^{10,13,21,51,99,100 101}  Training and achieving proficiency in gastric ultrasonography followed by wide adoption among anesthesiology clinicians will take some time. Most importantly, gastric ultrasonography only identifies those with retained gastric contents after they arrive for their procedures. It does not allow for timely interventions to avoid the risk of perioperative aspiration other than a change in the planned anesthetic or postponement of the procedure. A better approach is needed to ensure a majority of patients taking GLP-1 receptor agonists present without retained gastric contents preprocedurally and can avoid unnecessary anesthesia procedures such as rapid sequence induction and intubation, or aspiration on emergence from anesthesia. Gastric ultrasonography may be best suited for patients who are symptomatic or are asymptomatic without having observed prolonged fasting times.

Emerging data point to the limitations of the current ASA consensus-based guidance for patients taking GLP-1 receptor agonists. It is becoming more obvious that interrupting GLP-1 receptor agonists as currently suggested may not offer much benefit, since a longer washout time is needed to resolve the effects on gastric emptying. Moreover, the current approach gives clinicians a false sense of security when proceeding with anesthesia, especially when a monitored anesthesia care or a standard general anesthesia induction technique is planned, as residual gastric contents may be present. Furthermore, the glycemia benefit and the perioperative effects of mitigating major adverse cardiovascular events may be lost with drug interruptions in patients with diabetes mellitus. Additionally, there is the possibility of renewed side effects when reinstating a high-dose GLP-1 receptor agonist, and de-escalating may not be feasible due to fixed-dose prefilled pen preparations.

GLP-1 receptor agonists may need to be interrupted in patients who have recently started these medications or are in dose escalation phases, and have significant symptoms of fullness, nausea, and vomiting. Stopping the drugs or de-escalating doses, when possible, may be beneficial to decrease risks of aspiration in these patients who have not reached a steady state or developed tachyphylaxis to the gastrointestinal side effects. These patients may be referred to their prescribing physician for diet and medication modification before proceeding with elective procedures requiring anesthesia.

However, the first and most important intervention appears to be changing the duration of fasting for solids and possibly for liquids. Prolonged fasting time for solid foods beyond current recommendations is likely more effective than drug interruption to prevent patients from presenting on the day of procedures with residual gastric contents.⁵³  Data from the “real world” and gastric emptying studies support recommending a clear liquid diet for 24 h before anesthesia.⁵³  The fasting duration for liquids is unclear at this time, but it appears that patients who have recently started a long-acting GLP-1 receptor agonist or are in the dose escalation period and those on short-acting preparations may need a more prolonged NPO time than currently recommended. The caloric content of the liquids is important, with high-caloric clear liquids likely needing a longer fasting time than the low- or noncaloric clear liquids (8 h vs. 4 h). While these timeframes for liquids may appear arbitrary, data suggest that complete NPO times ranging between 4 and 12 h are more likely to result in lower gastric volumes, thus balancing the risks of aspiration with enhanced recovery principles.

Patients not tolerating the GLP-1 receptor agonist should be advised to follow a bland, low-fat, and low-fiber diet and to eat small meals. Medication changes (i.e., dose de-escalation) are secondarily considered. Elective surgeries should be postponed in patients with significant symptoms refractory to these interventions. Patients who are planning dose escalation should be advised not to do so in the weeks leading up to elective procedures. Collaboration with the prescribing clinicians is important when making decisions.

Proceeding with the planned anesthetic appears safe for asymptomatic patients presenting for elective procedures who are in maintenance phases, and have followed clear liquid diets for 24 h. For symptomatic patients orthose who have not been on clear liquids for 24 h, a gastric ultrasound may help guide decisions.

Since the publication of the ASA consensus-based guidance, a number of studies have been published on the gastric emptying effects of GLP-1 receptor agonists. Therefore, we propose updated recommendations for decision-making during the preoperative period and on the day of surgery (figs. 5 and 6).

Data on the perioperative management of patients taking GLP-1 receptor agonists are sparse, but emerging. There is heterogenicity as to the subjects studied (healthy controls, patients with obesity or diabetes) and the tests and measurements used, which may limit the ability to extrapolate findings across populations. Perioperative clinicians need to consider indications for medication, dosing schedule, concurrent symptoms, and the logistics of access to the medications in specific doses. Dietary recommendations, anesthetic options, and the relative urgency of the proposed procedure are also important factors to review. Indeed, a multidisciplinary, patient-centric approach is needed when caring for these patients.

The authors of this article served on the Society for Perioperative Assessment and Quality Improvement (SPAQI; Boston, Massachussetts) taskforce on GLP-1 RAs, with Dr. Oprea in the role of Chair. There was timing overlap between the work of the SPAQI taskforce and the last revision of the manuscript, and the authors acknowledge the taskforce for the discussion on liquid fasting times. The authors also want to acknowledge the ongoing work of the SPAQI taskforce that may result in future publication of recommendations for patients on GLP-1 RAs. Finally, the authors would like to acknowledge Michael Camilleri, M.D., D.Sc., Division of Gastroenterology and Hepatology, Mayo Clinic, Rochester, Minnesota, for his contribution to their understanding of the gastric emptying studies.

Support was provided solely from institutional and/or departmental sources.

Dr. Oprea reports honoraria from Westchester Medical Center (Valhalla, New York) and Brigham and Women’s Hospital (Boston, Massachusetts). Dr. Umpierrez received grant support from Novo Nordisk (Bagsværd, Denmark), Merck (Rahway, New Jersey), Insulcloud (Madrid, Spain), Dexcom (San Diego, California), Abbott (Green Oaks, Illinois), AstraZeneca (Cambridge, United Kingdom), and Bayer (Leverkusen, Germany). Dr. Sweitzer received funding from UpToDate (Boston, Massachussetts), International Anesthesia Research Society (San Francisco, California), and Medtronic (Minneapolis, Minnesota). Dr. Hepner received funding from Pharmacosmos (Holbæk, Denmark) and UpToDate.