GLP-1 and Anesthesia Aspiration: The Evidence (2026)

In 2023 a handful of alarming case reports — patients who had fasted overnight, even up to 18 hours, but still aspirated solid food into their lungs under anesthesia while taking semaglutide — touched off one of the fastest-moving safety debates in perioperative medicine. The American Society of Anesthesiologists told patients to hold the drug; within 18 months the same society and four others walked that back toward a risk-stratified approach. So what does the actual evidence show? The honest answer has two halves that are easy to confuse. What is well-established: GLP-1 drugs leave more food and fluid in the stomach after standard fasting — that part is proven by gastric-ultrasound studies and endoscopy data. What is NOT well-established: that this translates into a measurably higher rate of clinical aspiration. The largest cohorts to date found retained gastric contents went up sharply while actual aspiration pneumonia did not (Chen 2025 ^[10]). This article walks through the case reports, the ultrasound and endoscopy cohorts, and the guideline reversal — and is careful to separate the signal from the proof. For the practical "do I hold my dose" version, see holding a GLP-1 before surgery; for the procedure-specific case, see GLP-1s and colonoscopy/endoscopy prep.

The honest summary

• The case reports are real — and a few were severe. Published 2023–2024 reports describe patients who fasted appropriately yet had large gastric contents, with solid food aspirated into the trachea and bronchi under anesthesia (Klein 2023^[6]; Avraham 2024^[7]; Queiroz 2023^[8]). One patient needed bronchoscopy to remove food from the airway.
• "More food in the stomach" is the part that is proven. In fasted diabetic surgical patients, a GLP-1 raised the odds of a "full stomach" on ultrasound about 11-fold (Sen 2024^[2]); endoscopy studies show roughly 4–5× the odds of retained contents (Silveira 2023^[3]; Baig 2025^[1]).
• Actual aspiration-outcome rates are NOT well established. The leap from "fuller stomach" to "more clinical aspiration" is not supported by the large outcome data so far. In 366,476 surgical patients, GLP-1 use was not associated with more postoperative aspiration pneumonia (Chen 2025^[10]); a JAMA cohort found no rise in postoperative respiratory complications (Dixit 2024^[9]); the endoscopy meta-analysis found no significant increase in aspiration pneumonia (Baig 2025^[1]).
• Case reports cannot give you a rate. They prove the event can happen; they cannot tell you how often, because there is no denominator and aspiration is also under-reported and confounded.
• The guidance reversed. The ASA's June 2023 advice to hold the drug was replaced in October 2024 by multisociety guidance saying most patients can continue it, with risk stratification, a pre-procedure clear-liquid diet, and tools like gastric ultrasound for higher-risk cases (Kindel 2024^[12]; Hashash 2024^[11]).
• Bottom line: the delayed-emptying signal is real and worth managing; a large jump in aspiration deaths is not what the data shows. Don't change your medication on your own — coordinate with your anesthesia team and prescriber.

Where the alarm started: the 2023 case reports

The signal began not with a trial but with individual patients. In the Canadian Journal of Anesthesia, Klein and Hobai described a 42-year-old who had started weekly semaglutide two months earlier and presented for an upper-GI procedure. Despite an 18-hour fast, the stomach held substantial solid content, and food had to be removed from the trachea and bronchi by bronchoscopy after aspiration (Klein 2023^[6]). In Anaesthesia Reports, Avraham and colleagues reported two patients on semaglutide for weight loss who aspirated gastric contents under anesthesia after standard fasting (Avraham 2024^[7]). And in Einstein (São Paulo), Queiroz and colleagues documented aspiration in a fasted semaglutide patient with tomographic (CT) evidence of the aspirated material — a rare instance of imaging confirmation (Queiroz 2023^[8]).

These reports were genuinely important: they established that a patient who follows the rules — overnight or longer fast, no food, no obvious red flags — can still have a stomach full enough to aspirate. That is the kind of event anesthesiologists are trained to prevent, and it had not been on the routine radar for an outpatient weight-loss or diabetes drug. The reaction was swift and reasonable. But case reports have a structural limit that is easy to forget in the headlines: they are a numerator with no denominator.

What's proven: delayed emptying and retained gastric contents

The single most reproducible finding is mechanistic and direct: people on a GLP-1 have more in their stomach after fasting. The cleanest perioperative demonstration is Sen 2024^[2] in JAMA Surgery — a prospective study using point-of-care gastric ultrasound in appropriately fasted diabetic patients before elective surgery. GLP-1 users had a markedly higher median gastric volume than non-users and roughly 11× the odds of a "full stomach" on ultrasound (odds ratio ~11). Crucially, this was a surgical population assessed by imaging, not an endoscopy case series — so it isolates the gastric-content effect cleanly.

Endoscopy data point the same direction. Silveira 2023^[3] retrospectively studied 404 upper-endoscopy patients: those who had taken semaglutide within 30 days had retained gastric contents about 24% of the time versus 5% in non-users (adjusted odds ratio ~5), even though most had stopped roughly 10 days earlier. A prospective gastric-ultrasound study in volunteers recently started on semaglutide found the same direction of effect (Sherwin 2023^[4]). Pooling 23 studies and 262,018 patients, Baig 2025^[1] found GLP-1 users had about 4.5× the odds of retained gastric contents and of having the endoscopy aborted early.

There is also a dose-timing dimension. In Santos 2024^[5], among semaglutide users having upper endoscopy, stopping the drug for fewer than 8 days still left roughly a 10-fold higher chance of retained contents, and 8–14 days about a 4.6-fold higher chance; only stopping for more than 14 days (in people without GI symptoms) brought the risk in line with non-users. Sen 2024 similarly found GLP-1 use within 7 days associated with higher gastric volumes than holding longer^[2]. So the delayed-emptying effect is robust, dose-related in time, and not reliably erased by a one-week hold.

What's NOT proven: a measurable jump in aspiration outcomes

Here is the crux, and the part most often lost in coverage. A fuller stomach is a risk factor for aspiration — but the outcome studies have not shown that GLP-1 users actually aspirate, or develop aspiration pneumonia, at a meaningfully higher rate. The largest direct test is Chen 2025^[10] in JAMA Network Open: a cohort of 366,476 adults undergoing common surgical procedures found no significant difference in postoperative aspiration pneumonia between GLP-1 users and non-users after adjustment for demographics, comorbidities, and surgical factors. The authors explicitly suggested it may be worth reassessing blanket pre-procedure withholding.

That is consistent with the other outcome data. Dixit 2024^[9] in JAMA used a claims database of patients with diabetes undergoing emergency surgery and found no increase in postoperative respiratory complications with preoperative GLP-1 use — notable because emergency patients can't hold the drug or fast on schedule, making them a natural stress test. And the Baig 2025^[1] endoscopy meta-analysis, despite finding ~4.5× more retained contents, found no significant increase in aspiration pneumonia (pooled odds ratio ~0.96). Three independent datasets, the same disconnect: contents up, clinical aspiration not clearly up.

Why might "more food in the stomach" not produce a flood of aspiration events? Several plausible reasons, none yet proven: clinical aspiration is rare at baseline, so even a relative increase may be small in absolute terms; modern airway management (rapid-sequence induction, suctioning, recognizing a "full stomach") may catch many at-risk cases; and the most severe events may be concentrated in patients with ongoing GI symptoms rather than across all users. The honest framing is that the numerator (events) stayed low while the risk factor (retained contents) went up — and the field doesn't yet have the granular, prospective outcome data to fully resolve why.

How the guidance moved: from blanket hold to risk-stratified

In June 2023, with the case reports fresh, the American Society of Anesthesiologists issued consensus-based guidance to hold GLP-1s before elective procedures — skip the dose on the day for daily agents, and stop weekly agents (Ozempic/Wegovy semaglutide, Mounjaro/Zepbound tirzepatide) about a week ahead. It was explicitly consensus-based, acknowledging limited evidence; in practice it led some teams to cancel procedures when patients hadn't held the drug.

Gastroenterologists pushed back quickly. The American Gastroenterological Association's rapid clinical practice update argued for an individualized approach rather than reflexive discontinuation, reasoning that stopping a GLP-1 can worsen diabetes control, a single held weekly dose may not reliably empty the stomach anyway, and cancelling needed procedures carries its own harm (Hashash 2024^[11]).

In October 2024 the camps converged. Multisociety guidance — developed with anesthesiology, gastroenterology, and bariatric-surgery groups — concluded that most patients can continue their GLP-1 before elective surgery, with decisions based on shared decision-making and individual aspiration risk (Kindel 2024^[12]). Its practical middle path: for patients where delayed emptying is a concern, a clear-liquid diet for at least 24 hours before the procedure (as already done for colonoscopy and bariatric surgery); for higher-risk cases, options like point-of-care gastric ultrasound on the day, treating the patient as "full stomach" with a rapid-sequence anesthesia technique, or, rarely, rescheduling. A strict drug hold became one option among several rather than the default. This shift mirrors the evidence above: the retained-contents signal is real (so manage it), but the absence of a clear outcome signal made a blanket hold hard to justify.

What the evidence does and doesn't let you conclude

• You CAN conclude: a GLP-1 meaningfully raises the chance your stomach isn't empty after standard fasting — proven by ultrasound and endoscopy (Sen 2024^[2]; Baig 2025^[1]; Silveira 2023^[3]).
• You CAN conclude: aspiration can happen in a fasted GLP-1 patient — the case reports prove the event is possible and occasionally serious (Klein 2023^[6]; Avraham 2024^[7]; Queiroz 2023^[8]).
• You CANNOT conclude: that GLP-1 users aspirate at a clearly higher rate — the large outcome cohorts did not show a significant increase in aspiration pneumonia or respiratory complications (Chen 2025^[10]; Dixit 2024^[9]; Baig 2025^[1]).
• You CANNOT conclude: that a one-week hold guarantees an empty stomach — under-8-day and 8–14-day holds still left elevated retained-content odds (Santos 2024^[5]).
• Biggest evidence gap: there is no large, prospective study linking measured gastric volume to actual clinical aspiration outcomes in GLP-1 users — so the contents-to-outcome bridge remains inferred, not measured.

Bottom line

The GLP-1 aspiration story is a case study in separating a risk factor from an outcome. The risk factor is solidly established: these drugs leave more in the stomach after fasting — about 11× the odds of a full stomach on surgical ultrasound (Sen 2024^[2]) and ~4.5× retained contents at endoscopy (Baig 2025^[1]), and vivid case reports prove aspiration can occur even after an 18-hour fast (Klein 2023^[6]). The outcome is not: the largest cohorts (≈390,000 patients combined) found no clear increase in aspiration pneumonia or respiratory complications (Chen 2025^[10]; Dixit 2024^[9]). That disconnect is exactly why the guidance moved from a 2023 blanket hold to 2024 risk-stratified management — continue the drug for most, use a clear-liquid diet and gastric ultrasound for the higher-risk, and reserve a hold for selected cases (Kindel 2024^[12]). The practical takeaway hasn't changed: don't change your medication on your own. Tell your anesthesia team and prescriber exactly which GLP-1 you take, your dose, and when you last took it, and follow your facility's plan. See the practical ASA-hold guide and the colonoscopy/endoscopy prep article for the procedure-specific steps.

This article is educational and is not medical advice. Every study, case report, society update, and multisociety consensus cited above is indexed in PubMed and was verified against the live PubMed database before publication. Perioperative GLP-1 decisions should be made with your anesthesia team and prescriber.