Glucagon-like peptide-1 receptor agonists do not increase aspiration during upper endoscopy in patients with diabetes

Abstract

BACKGROUND AND AIMS:

Glucagon-like peptide-1 receptor agonists (GLP1-RA) have been associated with greater retention of gastric contents, however there have been no controlled, population-based studies performed to evaluate the potential adverse effects of GLP1-RA in the periprocedural setting. We aimed to determine if there is increased risk of aspiration and aspiration-related complications after upper endoscopy in patients using GLP1-RA.

METHODS:

We used a nationwide commercial administrative claims database to conduct a retrospective cohort study of patients aged 18-64 with type 2 diabetes who underwent outpatient upper endoscopy from 2005-2021. We identified 6,806,046 unique upper endoscopy procedures. We compared claims for aspiration and associated pulmonary adverse events in the 14 days following upper endoscopy between users of GLP1-RA, dipeptidyl peptidase 4 inhibitors (DPP4i), and chronic opioids. We adjusted for age, sex, Charlson Comorbidity score, underlying respiratory disease, and gastroparesis.

RESULTS:

We found that pulmonary adverse events following upper endoscopy are rare, ranging from 6-25 events per 10,000 procedures. When comparing GLP1-RA to DPP4i, crude relative risks of aspiration (0.67 95%CI 0.25,1.75), aspiration pneumonia (0.95 95%CI 0.40,2.29), pneumonia (1.07 95%CI 0.62,1.86), or respiratory failure (0.75 95%CI 0.38,1.48) were not higher in patients prescribed GLP1-RA. When comparing GLP1-RA to opioids, crude relative risks (95%CI) were 0.42 (0.15,1.16) for aspiration, 0.60 (0.24,1.52) for aspiration pneumonia, 0.30 (0.19,0.49) for pneumonia, and 0.24 (0.13,0.45) for respiratory failure. These results were consistent across several sensitivity analyses.

CONCLUSIONS:

GLP1-RA use is not associated with increased risk of pulmonary complications after upper endoscopy compared to DPP4i use in patients with type 2 diabetes.

Introduction:

Glucagon-like peptide-1 receptor agonist (GLP1-RA) medications are increasingly used for the treatment of type 2 diabetes mellitus (T2DM) and obesity, with emerging benefits in cardiovascular, renal, and neurodegenerative disease.^{1, 2} GLP1-RA have a variety of gastrointestinal side effects including nausea, vomiting, and delayed gastric emptying.³

Delayed gastric emptying associated with GLP1-RA has raised concern for potential adverse effects in the periprocedural setting. Studies of patients taking GLP1-RA have demonstrated increases in gastric residue on esophagogastroduodenoscopy (EGD)^4-7 and gastric ultrasound.^{8, 9} Case reports and single-center case series have documented aspiration events in patients receiving procedural anesthesia on GLP1-RA.^{10, 11} However, few controlled, multi-center studies have investigated whether clinically meaningful outcomes, such as aspiration events or aspiration pneumonia, are increased in patients who undergo procedures on GLP1-RA.¹² Despite the limited evidence in this area, the American Society of Anesthesiology (ASA) has published consensus guidance to hold GLP1-RA administration on the day or week prior to elective procedures (depending on GLP1-RA dose schedule), and if a patient is taking GLP1-RA to treat them with “full stomach” precautions or delay the procedure.¹³ This policy is highly relevant to the practicing endoscopist and anesthesiologist, as EGD is a sedated procedure performed thousands of times per day in the United States with native airway. A joint statement from multiple gastroenterology societies and a clinical practice update called into question these ASA recommendations, and pointed out the significant knowledge gap that exists regarding periprocedural management of GLP1-RA.^{14, 15} Given the lack of controlled, population-based studies evaluating periprocedural complications in patients taking GLP1-RA, we used a large commercial administrative claims database to compare aspiration and aspiration-related complications after upper endoscopy among three medication user groups: (1) GLP1-RA, (2) dipeptidyl peptidase 4 inhibitors (DPP4i), and (3) opioids. DPP4i are a common medication class used for diabetes with a related mechanism of action to GLP1-RA, but negligible effects on gastric emptying¹⁶. Opioids have similar delayed gastric emptying effects to GLP1-RA, but no specific periprocedural management guidelines.

Methods:

Study design, data source, and study population

We conducted a retrospective cohort study using Truven Health Analytics MarketScan databases. MarketScan administrative claims databases capture health insurance enrollment information, inpatient, outpatient, and pharmacy data for roughly 20 million private health insurance beneficiaries annually from nearly 100 commercial payers. The study design is outlined in Figure 1. This representation was adopted from a guideline on graphical depiction of healthcare database studies.¹⁷

Figure 1.

Study design

Visual representation adopted from Schneeweiss et. al., 2019.

We included patients aged 18-64 with type 2 diabetes (T2DM) who underwent an outpatient EGD between 2005-2021. For patients who received multiple EGDs during the study period only the first EGD was included, hereafter referred to as the index EGD. At least 12 months of continuous insurance enrollment prior to index EGD was required to assess comorbid conditions and prescribed medications. At least 14 days of continuous insurance enrollment following index EGD was required to assess outcomes. EGDs were identified using Current Procedural Terminology (CPT) codes (Supplemental Table 7).¹⁸ T2DM patients were identified by any 2 (or more) International Classification of Diseases (ICD)9 or ICD10 codes for T2DM (Supplemental Table 8), filed as inpatient or outpatient claims within the 12 months prior to index EGD.¹⁹

Inpatient EGDs were excluded. Patients with type 1 diabetes (T1DM) were excluded, identified by >1 IC9 or ICD10 code for T1DM (Supplemental Table 8), in inpatient or outpatient claims, within the 12 months prior to index EGD.¹⁹ Patients with pre-existing aspiration, aspiration pneumonia, pneumonia, or respiratory failure were excluded, identified by any ICD9 or ICD10 code filed as inpatient or outpatient claims in the 6 months prior to index EGD (Supplemental Table 8).^20-23

We performed a sensitivity analysis excluding patients with known risk factors for aspiration, including a history of dementia, stroke, dysphagia, ear, nose or throat (ENT) cancer, or esophageal cancer.²⁴ These conditions were identified by any ICD9 or ICD10 code (Supplemental Table 8) filed as inpatient or outpatient claims within the 12 months prior to index EGD.

Groups, exposures, and outcomes

Patients who met inclusion and exclusion criteria were divided into 3 groups. The exposure group of patients using GLP1-RA was compared to two active comparator groups of patients using (1) dipeptidyl peptidase 4 inhibitors (DPP4i) or (2) chronic opioids. DPP4i have a related mechanism of action to GLP1-RA and therefore identify a similar group of patients with diabetes receiving prescription pharmacotherapy. However, DPP4i have negligible effects on gastric emptying.¹⁶ Opioids reliably and potently delay gastric emptying. Neither DPP4i nor opioids carry recommendations from professional societies to alter periprocedural administration, pre-anesthesia fasting, or airway management.

We defined GLP1-RA users as patients with 2 or more pharmacy claims for a single GLP1-RA, identified by National Drug Code (NDC) numbers for GLP1-RA (Supplemental Table 9), within the 6 months prior to index EGD.²⁵ We defined DPP4i users as patients with 2 or more DPP4i pharmacy claims, identified by NDC numbers for several common DPP4i (Supplemental Table 9) and no claims for GLP1-RA, in the 6 months prior to index EGD.²⁵ We defined chronic opioid users as patients with 2 or more opioid pharmacy claims, identified by NDC numbers for selected common, potent opioids representative of the medication class (oxycodone and hydromorphone, Supplemental Table 9) and no claims for GLP1-RA, in the 6 months prior to index EGD.²⁶

We performed sensitivity analyses that further increased the stringency of the definition of a medication user. First, for all three drug groups (GLP1-RA, DPP4i, or opioids), we changed the definition of a user from 2 or more pharmacy claims to 3 or more pharmacy claims. Second, we used a medication possession ratio (MPR) cutoff to define “adherent” users. The MPR equals (sum of total days supply of all drug claims) / (total number of days from first drug claim during the 6 month period prior to the index endoscopy to the index endoscopy). An MPR > 0.8 defines an “adherent user” in the literature.²⁷ Third, we evaluated incident medication users as opposed to all other analyses which evaluated prevalent medication users. This analysis was performed to address prevalent user bias. Incident medication users were defined as patients who had 2 or more pharmacy claims for a drug (GLP1-RA, DPP4i, or opioid) in the 6 months prior to index EGD, with a washout period of no claims for the same drug class in the 6-12 months prior to index EGD. Fourth, in our primary analysis, patients who had claims for both opioids and DPP4i were included in both the chronic opioid user group and the DPP4i user group. Additionally, patients taking a GLP1-RA and a DPP4i or opioid were included in the GLP1-RA group. To assess whether outcome results could be driven by patients using multiple medications, we performed a sensitivity analysis that excluded patients with any claim for DPP4i or opioids in the 12 months prior to index EGD in the GLP1-RA group, excluded patients with any claim for DPP4i or GLP1-RA in the 12 months prior to index EGD in the opioid group, and excluded patients with any claim for GLP1-RA or opioids in the 12 months prior to index EGD in the DPP4i group.

The four primary outcomes were defined by any claim for aspiration pneumonia, aspiration, pneumonia, or respiratory failure within the 14 days following index EGD. We additionally assessed a composite primary outcome measure, hereafter referred to as composite pulmonary adverse events (AE), defined as any claim for aspiration pneumonia, aspiration, pneumonia, or respiratory failure within the 14 days following index EGD. We defined these outcomes using ICD9 and ICD10 codes, filed as inpatient or outpatient claims (Supplemental Table 8).^20-23

Secondary outcomes included emergency room (ER) visit or hospitalization for any cause within the 14 days following index EGD. Hospitalization was defined by any claim associated with an inpatient admission. ER visit was defined by any claim with a service type corresponding to emergency room visit.

Covariate data

We collected the covariate data available in the MarketScan databases, which includes age and sex. We collected data on relevant co-morbidities including chronic obstructive pulmonary disease (COPD), asthma, and gastroparesis, defined by ICD9 or ICD10 codes filed as inpatient or outpatient claims within 12 months prior to index EGD (Supplemental Table 8).^{22, 28} To assess other pre-existing health conditions we used the Charlson Comorbidity Index (CCI), with diabetes and COPD removed, using inpatient and outpatient claims filed within 12 months prior to index EGD.²⁹

We conducted sensitivity analyses using propensity score weighting to balance baseline covariates between groups. Propensity scores were estimated separately for the two comparisons (GLP1-RA vs DPP4i, GLP1-RA vs opioids) with age, sex, Charlson score, history of COPD, asthma, and gastroparesis as predictors of treatment group. We then used inverse probability of treatment weighting (IPTW) to balance covariates across treatment groups compared. Covariate balance in the weighted samples was assessed based on absolute standardized mean differences (SMD).

Statistical analysis

Descriptive statistics were used to compare covariate distribution between medication user groups. To determine whether GLP1-RA users compared to either active comparator had increased risk of primary or secondary outcomes, we estimated relative risks using Poisson regression with robust error variance.³⁰ Multivariable models were adjusted for age, sex, Charlson score, and history of COPD, asthma, or gastroparesis (any of the three). We did not estimate adjusted relative risks when the number of outcomes was 10 or fewer to avoid overfitting. This method was used to estimate relative risks and adjusted relative risks for secondary outcomes and sensitivity analyses.

Results:

Overview of the population

Our study population included 6,806,046 unique upper endoscopy procedures between 2005-2021. We identified 274,211 outpatient EGD procedures in patients with type 2 diabetes who met inclusion criteria. We identified 15,119 GLP1-RA users, 14,407 DPP4i users, and 7,257 chronic opioid users (Figure 2). Baseline clinical characteristics of the medication user groups are shown in Table 1. Patients in each group had similar age and sex distributions. We observed differences between groups in a composite measure of comorbidities. While patients in the GLP1-RA group and DPP4i group were similar, patients in the chronic opioid group were more likely to have a CCI of 2+, indicating a higher burden of comorbidities. When looking specifically at the prevalence of COPD and asthma, while patients prescribed GLP1-RA and DPP4i were similar, patients prescribed chronic opioids more commonly had these comorbid conditions. Gastroparesis was equally rare in all three groups.

Figure 2.

Study population

N= number of patients remaining after each selection or exclusion

*For patients with multiple EGD procedures during the study time period, only the first EGD was included

Table 1.

Demographic characteristics and comorbidities among type 2 diabetic patients prescribed GLP1-RA, DPP4i, and opioids who underwent upper endoscopy between 2005-2021 (N=36,783)

	GLP1-RA N = 15,119		DPP4i N= 14,407		Opioids N=7,257
	N	%	N	%	N	%
Age * , years
<40	977	6%	473	3%	446	6%
40-49	3264	22%	2055	14%	1361	19%
50-59	6958	46%	6665	46%	3358	46%
60-64	3920	26%	5214	36%	2092	29%
Median (IQR)	55 (49-60)		57 (52-61)		56 (50-60)
Sex
Male	6101	40%	6959	48%	3424	47%
Female	9018	60%	7448	52%	3833	53%
Charlson score †
0	8425	56%	7706	53%	2777	38%
1	3794	25%	3172	22%	1794	25%
≥2	2900	19%	3529	24%	2686	37%
COPD ‡
No	14600	97%	13955	97%	6660	92%
Yes	519	3%	452	3%	597	8%
Asthma ‡
No	13582	90%	13199	92%	6226	86%
Yes	1537	10%	1208	8%	1037	14%
Gastroparesis ‡
No	14896	99%	14322	99%	7180	99%
Yes	223	1%	85	1%	77	1%

^*At time of index EGD

^†Does not include diabetes, diabetes complications, or COPD

^‡Any diagnosis code in the year prior to index EGD

Primary analysis

Overall, pulmonary adverse events following upper endoscopy were rare (Table 2). In our cohort of patients with type 2 diabetes who underwent EGD and were prescribed GLP1-RA, DPP4i, or chronic opioids, per 10,000 EGDs we observed 6.80 aspirations, 7.61 aspiration pneumonias, and 25.56 cases of respiratory failure.

Table 2:

Risk of primary and secondary outcomes among users of GLP1-RA (N=15,119), DPP4i (N=14,407) and opioids (N=7,257)

	GLP1-RA		DPP4i		Opioids		GLP1-RA vs DPP4i (ref)		GLP1-RA vs Opioids (ref)
	N	%	N	%	N	%	Unadjusted RR (95% CI)	Adjusted RR (95% CI)*	Unadjusted RR (95% CI)	Adjusted RR (95% CI)*
Aspiration	7	0.05%	10	0.07%	8	0.11%	0.67 (0.25, 1.75)	--	0.42 (0.15, 1.16)	--
Aspiration pneumonia	10	0.07%	10	0.07%	8	0.11%	0.95 (0.40, 2.29)	--	0.60 (0.24, 1.52)	--
Pneumonia	27	0.18%	24	0.17%	43	0.59%	1.07 (0.62, 1.86)	1.15 (0.66, 2.00)	0.30 (0.19, 0.49)	0.38 (0.23, 0.63)
Respiratory failure	15	0.10%	19	0.13%	30	0.41%	0.75 (0.38, 1.48)	0.82 (0.41, 1.62)	0.24 (0.13, 0.45)	0.34 (0.18, 0.64)
Composite pulmonary AE †	47	0.31%	54	0.37%	73	1.01%	0.83 (0.56, 1.23)	0.90 (0.60, 1.34)	0.31 (0.21, 0.45)	0.38 (0.26, 0.55)
Hospitalization	335	2.22%	335	2.33%	358	4.93%	0.73 (0.62, 0.86)	0.76 (0.64, 0.89)	0.34 (0.29, 0.40)	0.41 (0.35, 0.48)
ER visit	628	4.15%	711	4.94%	776	10.69%	0.84 (0.76, 0.93)	0.84 (0.76, 0.94)	0.39 (0.35, 0.43)	0.44 (0.39, 0.49)

^*Adjusted for age, sex, Charlson score, and history of COPD, asthma, and gastroparesis (any of the 3)

^†Composite pulmonary adverse events defined as any claim for aspiration, aspiration pneumonia, pneumonia, or respiratory failure in the 14 days following index EGD

The crude and adjusted risk of individual pulmonary adverse events such as aspiration, aspiration pneumonia, pneumonia, or respiratory failure was not higher in GLP1-RA users compared to DPP4i users or to opioid users (Table 2). This held true for the composite measure as well. Crude and adjusted risks of pneumonia, respiratory failure, and composite pulmonary adverse events were higher in the chronic opioid group compared to the GLP1-RA group. Regarding secondary outcomes, the adjusted risks of post-EGD hospitalization or ER visit were higher in opioid and DPP4i users compared to GLP1-RA users.

Sensitivity analyses

When we removed patients with baseline risk factors for pulmonary complications from our primary analysis (N=4,358), the results were largely unchanged (Supplemental Table 2).

We performed four separate analyses that made the definitions of a GLP1-RA, DPP4i, and opioid user more stringent than those used in our primary analysis. First, we increased the definition of a medication user from ≥2 claims per drug in the 6 months prior to index EGD to ≥3 claims per drug (Supplemental Table 3). Second, instead of a claim number-based definition for medication users, we selected medication users based on an MPR > 0.80, meaning that >80% of the days in the 6 months prior to the index EGD the patient had an “active” prescription for the medication (Supplemental Table 4). Third, we assessed incident medication users instead of prevalent medication users (Supplemental Table 5). Fourth, we excluded any patients who were taking multiple medications included in our study, which included 686 GLP1-RA users also using DPP4i, 417 also using opioids, and 21 using both DPP4i and opioids, as well as 382 DPP4i users also using opioids (Table 3). Results from these analyses yielded results similar to our primary analysis.

Table 3.

Risk of primary and secondary outcomes after sensitivity analysis performed excluding multiple drug users among users of GLP1-RA (N=12,212), DPP4i (N=13,009) and opioids (N=6,689)

	GLP1-RA		DPP4i		Opioids		GLP1-RA vs DPP4i (ref)		GLP1-RA vs Opioids (ref)
	N	%	N	%	N	%	Unadjusted RR (95% CI)	Adjusted RR (95% CI)*	Unadjusted RR (95% CI)	Adjusted RR (95% CI)*
Aspiration	6	0.05%	10	0.08%	7	0.10%	0.64 (0.23, 1.77)	--	0.47 (0.16, 1.40)	--
Aspiration pneumonia	8	0.07%	8	0.06%	8	0.12%	1.07 (0.40, 2.85)	--	0.55 (0.21, 1.47)	--
Pneumonia	22	0.18%	21	0.16%	39	0.58%	1.12 (0.62, 2.04)	1.17 (0.64, 2.13)	0.31 (0.18, 0.52)	0.40 (0.23, 0.68)
Respiratory failure	12	0.10%	16	0.12%	26	0.39%	0.80 (0.38, 1.70)	0.94 (0.44, 2.02)	0.25 (0.13, 0.50)	0.39 (0.19, 0.77)
Composite pulmonary AE †	38	0.31%	48	0.37%	66	0.99%	0.85 (0.55, 1.30)	0.93 (0.60, 1.43)	0.32 (0.21, 0.47)	0.40 (0.27, 0.61)
Hospitalization	203	1.67%	291	2.24%	330	4.94%	0.75 (0.63, 0.89)	0.77 (0.65, 0.93)	0.34 (0.29, 0.40)	0.41 (0.34, 0.49)
ER visit	471	3.88%	602	4.63%	720	10.77%	0.84 (0.74, 0.94)	0.85 (0.75, 0.95)	0.36 (0.32, 0.40)	0.41 (0.36, 0.46)

^*Adjusted for age, sex, Charlson score, and history of COPD, asthma, and gastroparesis (any of the 3)

^†Composite pulmonary adverse events defined as any claim for aspiration, aspiration pneumonia, pneumonia, or respiratory failure in the 14 days following index EGD

Note: Removed N=686 GLP1-RA users also prescribed DPP4i, N=417 GLP1-RA users also prescribed opioids, N=21 GLP1-RA users also prescribed DPP4i and opioids. N=382 users of both DPP4i and opioids were removed from the DPP4i and opioid group.

Lastly, we performed sensitivity analyses using inverse probability of treatment weighting to balance baseline covariates between treatment groups, accounting for patient differences between drug user groups by a different method less sensitive to small numbers of outcomes (Supplemental Table 6, Table 4). Again, we observed no increased risk of primary or secondary outcomes in the GLP1-RA group when compared to the DPP4i or opioid group.

Table 4.

Inverse probability of treatment weighted estimates of relative risk (RR) of primary and secondary outcomes among users of GLP1-RA compared to DPP4i and opioids

	GLP1-RA vs DPP4i (ref) RR (95% CI)	GLP1-RA vs Opioids (ref) RR (95% CI)
Aspiration	0.68 (0.26, 1.79)	0.42 (0.15, 1.18)
Aspiration pneumonia	1.21 (0.50, 2.94)	1.02 (0.35, 2.92)
Pneumonia	1.19 (0.68, 2.08)	0.49 (0.28, 0.87)
Respiratory failure	0.85 (0.43, 1.71)	0.41 (0.21, 0.80)
Composite pulmonary AE *	0.93 (0.62, 1.38)	0.45 (0.30, 0.69)
Hospitalization	0.77 (0.65, 0.91)	0.44 (0.37, 0.52)
ER visit	0.86 (0.77, 0.95)	0.45 (0.41, 0.51)

^*Composite pulmonary adverse events defined as any claim for aspiration, aspiration pneumonia, pneumonia, or respiratory failure in the 14 days following index EGD

Discussion:

This study presents the first controlled nationwide evaluation of post-endoscopic clinical outcomes in patients using GLP1-RA. Our study found that pulmonary adverse events after EGD are rare, on the order of 6-25 events per 10,000 EGDs, in a population of patients with type 2 diabetes. This finding is similar to previous estimates,³¹ including a recent single-center study of patients using GLP1-RA.¹¹ Importantly, we did not observe an increased risk of pulmonary adverse events after EGD in patients with type 2 diabetes prescribed GLP1-RA compared to patients prescribed DPP4i or opioids.

These findings are somewhat surprising given our knowledge of GLP1-RA physiology and the existing studies on GLP1-RA in the periprocedural setting. A recently published research letter¹² using a large health record database found a statistically significant increase in aspiration following endoscopy in GLP1-RA users compared to non-users. Whether this is a clinically significant increase remains unclear. Other literature has focused on various assessments of gastric contents in patients using GLP1-RA. While there is an intuitive and established connection between gastric contents and periprocedural adverse events,³² our investigation suggests that the surrogate outcome of gastric contents may not equate to clinical outcomes.

The findings of this national insurance claims study are highly generalizable. The use of opioids as an active comparator, a class of medications which similarly to GLP1-RA result in pharmacologic gastroparesis, yet have no periprocedural management guidelines, provides important context. These findings do not support the recent ASA guidance to manage GLP1-RA differently than other medications, such as opioids or DPP4i, in the periprocedural setting. Concerns based on GLP1-RA pharmacology and observational studies of surrogate outcomes are legitimate. However, given the well-documented salutary effects of GLP1-RA in patients with metabolic disease, these concerns must be weighed against the potential harms of holding these medications, delaying procedures, or performing prophylactic intubation, which could include additional complexity in pre-procedural instructions, worsened glycemic control, complications and costs from intubation, and costs of rescheduling procedures. These concerns should motivate high-quality, prospective studies of protocols of periprocedural GLP1-RA management, rather than overly restricting the use of highly efficacious medications.

This study has important limitations. Inherent to insurance claims-based studies we have limited covariate data, claims are subject to misclassification, and we are not able to use medical records to validate our diagnostic code definitions or findings. Exposures were defined by dispensed prescriptions, and we cannot verify whether patients were using these medications in the time before or on the day of the procedure. For the opioid comparison group, we selected commonly prescribed opioids rather than include every possible opioid user, which relies on the assumption of a class effect. We could not determine how many aspiration events may have been avoided by actions of the involved providers, such as aborted, truncated, or prophylactically intubated procedures. Accordingly, extrapolation of our findings to other procedures requiring anesthesia where gastric contents are not immediately visualized, such as most surgical procedures, should be done with caution.

A further limitation is that widespread use of newer, more potent GLP1-RA was not included in this analysis due to MarketScan’s data extending only to 2021. We also excluded GLP1-RA use exclusively for weight loss to minimize between-group variability by focusing on patients with type 2 diabetes. These limitations may impact the precision with which our estimates reflect the true impact of GLP1-RA on post-EGD pulmonary adverse events, which is likely low based upon the low total numbers of primary and secondary outcome events. However, even if these threats to precision limit the confidence with which we can exclude a slightly increased relative risk of pulmonary adverse events in patients prescribed GLP1-RA, they should not differentially bias the results. Further, even in this improbable scenario, the absolute risk would be quite small. Overall, consistent results across sensitivity analyses in 36,783 EGDs suggest that our findings are robust.

Lastly, our primary analysis evaluated prevalent medication users, as opposed to incident users. This raises the possibility of prevalent user bias, which is important considering the possible tachyphylaxis of long-acting GLP1-RA-mediated delayed gastric emptying.³³ In order to address this important potential bias, we performed an analysis of incident users within 6 months prior to index EGD with a washout period, who might be most impacted by medication-induced gastroparesis. While outcome numbers became lower than previous analyses, the overall findings were similar in direction and magnitude to our primary analysis, suggesting that even in this subgroup, pulmonary adverse events are rare and not more frequent in GLP1-RA users.

In conclusion, using a national insurance claims database, we found that patients with diabetes prescribed GLP1-RA did not have an increased risk of post-EGD aspiration events compared to patients prescribed DPP4i or chronic opioids. Given the rarity of post-EGD pulmonary events, and the lack of increased risk in this group of patients, measures to further decrease the risk of aspiration in GLP1-RA users may not be merited. These findings should be replicated in other populations and be considered when developing clinical guidelines for the periprocedural management of GLP1-RA medications.

What You Need to Know:

BACKGROUND:

Glucagon-like peptide-1 receptor agonist (GLP1-RA) medication use is associated with retained gastric contents and possible aspiration during procedures requiring anesthesia, including upper endoscopy.

FINDINGS:

Aspiration and pulmonary adverse events following upper endoscopy are rare. GLP1-RA users did not have an increased risk of aspiration or pulmonary complications following upper endoscopy compared to dipeptidyl peptidase 4 inhibitor (DPP4i) or opioid users.

IMPLICATIONS FOR PATIENT CARE:

Extensive measures to reduce the risk of postendoscopy pulmonary complications in GLP1-RA users may not be warranted. These findings should be considered when determining periprocedural management guidelines for GLP1-RA.

Abbreviations:

AE

Adverse events

ASA

American Society of Anesthesiology

CCI

Charlson Comorbidity Index

CI

Confidence interval

COPD

Chronic obstructive pulmonary disease

CPT

Current Procedural Terminology

DPP4i

dipeptidyl peptidase 4 inhibitors

EGD

esophagogastroduodenoscopy

ENT

Ear, nose, and throat

ER

Emergency room

GLP1-RA

glucagon-like peptide-1 receptor agonist

ICD

International Classification of Diseases

MPR

Medication possession ratio

RR

Relative risk

T1DM

type 1 diabetes mellitus

T2DM

type 2 diabetes mellitus