Feasibility and Utility of Routine Point‐of‐Care Gastric Ultrasonography in Patients Undergoing Upper Gastrointestinal Endoscopy Procedures: A Prospective Cohort Study

Previous studies have indicated that point‐of‐care ultrasonography (POCUS) of the gastric antrum can predict the adequacy of fasting before surgery and anesthesia. The aim of this study was to evaluate the utility of gastric POCUS in patients undergoing upper gastrointestinal (GI) endoscopy procedures.

million endoscopies are performed each year in the United States alone 1 but growing body of evidence indicates that this subset of patients is known to be at a higher risk of aspiration than those having general anesthesia for other procedures. 2,3The aspiration of gastric contents during anesthesia or sedation is a devastating complication that can be associated with significant morbidity and mortality. 4In a recent closed claims analysis of cases with pulmonary aspiration of gastric contents under anesthesia, more than half the number of claims with aspiration had mortality and the rest had permanent or significant injury but more alarmingly, non-compliance with fasting (NPO) guidelines was seen in only 15% of the patients raising the notion that the current NPO guidelines may not sufficiently cover all patients undergoing sedation or anesthesia.Interestingly, patients undergoing endoscopy formed the second biggest population in this report.While, different anesthesia society guidelines recommend similar but un-identical fasting times relevant to the type of meal and the age of patients, these guidelines do not take into consideration about risk factors affecting gastric emptying times or any pre-existing gastrointestinal pathology, thus making the fasting times potentially insufficient for certain subset of patients with risk factors for delayed emptying. 5,6oint-of-care ultrasound (POCUS) is recently being popularized in the perioperative arena with the aim to objectively determine the fasting status and to identify patients with unsafe residual gastric content but evidence on its utility to detect unsafe residual volumes in patients undergoing upper GI endoscopy in routine clinical practice is sparse.More evidence is also needed on the technique as there are studies both supporting [7][8][9][10] or refuting 11,12 a good correlation between residual gastric volumes and antral CSA.The impact of gastric POCUS is probably more relevant in patients presenting for emergency procedures, 13,14 with known non-fasting status or in patients with conditions or medication usage known to prolong gastric emptying but as more literature evidence emerges, even in patients without risk factors for aspiration or delayed gastric emptying, it is clear that adequate fasting status does not necessarily translate into an empty stomach. 15,16Aspiration under anesthesia or procedural sedation is multifactorial in nature and the mere presence of risk factors or even noting high residual gastric content may not necessarily translate into the occurrence of an adverse event but, additional information from gastric POCUS may impact the medical decision making and change the anesthesia and airway management plans as seen in a few reports. 16Perhaps the objective nature of gastric POCUS may better aid practitioners to make informed decisions on sedation and airway management practices during upper GI endoscopy procedures.Given the sensitivity and easy learning curve of gastric POCUS, the technique has been easily incorporated in clinical use by emergency physicians and gastroenterologists. 17 This begs the question as to whether routine gastric ultrasonography has utility to daily clinical practice when caring for patients undergoing upper GI endoscopy procedures.With this background, we performed a pragmatic study to evaluate the feasibility and clinical utility of gastric POCUS in patients undergoing anesthetic care for upper GI endoscopy procedures.We also wanted to know the degree of correlation between the POCUS-based estimate of gastric volume and the endoscopist-measured gastric volumes.

Methods
Patients scheduled for advanced upper GI endoscopy procedures in the ambulatory suite at University of Alabama Hospital were consented to be included in the study.Institutional ethics approval (UAB institutional review board approval number: 300003835.Approved on: September 29, 2019) was obtained before the recruitment of patients for the study.Patients presenting for elective advanced upper endoscopic procedures were eligible for study recruitment.Any patients with altered gastric anatomy, previous history of gastric surgery, history of gastric outlet obstruction, history of medications or predictors for delayed gastric emptying (chronic narcotics, cannabinoids, EtOH abuse, anticholinergics), emergent cases; indication for upper GI endoscopy due to a GI bleed, pregnant patients were excluded from the study.All ultrasound exams were performed by two experienced anesthesiologists who had at least performed 50 scans before the beginning of study.Safe and unsafe ultrasound scans in these adequately fasted patients were determined both qualitatively and quantitatively.A qualitative gastric ultrasonography was performed on enrolled patients before the beginning of the endoscopy procedure in supine (Figure 1A) and right lateral decubitus (Figure 1B) to determine the nature of content (empty/minimal fluid; fluid contents; solid content/heterogenous content). 8Representative gastric ultrasound images showing empty stomach, clear liquids, and solid contents are shown in Figure 2, A-C, respectively.A curvilinear low-frequency transducer (2-5 MHz; Sonosite™ EPIQ, Fujifilm, Bothell, WA) was used with standard abdominal settings to identify the relevant anatomic landmarks.With the patient in the supine and right lateral decubitus position, the antrum was imaged in a parasagittal plane in the epigastric area using the left lobe of the liver, the inferior vena cava/aorta, and the superior mesenteric artery/ vein as anatomic landmarks.A cross-sectional view of the gastric antrum was identified adjacent to the left lobe of liver in that plane.A free-tracing tool function was utilized to mark the full thickness of the gastric wall, from serosa to serosa, in between peristaltic contractions to obtain the gastric antral cross-sectional area (CSA).A similar measurement was carried out with the patient in right lateral decubitus positions.Utilizing CSA measurements and age, gastric volumes were calculated using both the nomogram and the formula developed by Perlas et al 10 (Gastric volume [mL] = 27.0 + 14.6 Â right-lateral CSA -1.28 Â age) before the beginning of the endoscopy.A normal threshold was defined as liquid gastric volumes <1.5 mL/kg.Subsequently, patients received monitored anesthesia care or general anesthesia at the discretion of the anesthesia provider assigned to the case.Anesthesia providers had information about the gastric ultrasonography data so that it could alter their management.The gastroenterologist blinded to ultrasound data performed upper GI endoscopy and suctioned gastric contents were collected and measured.While feasibility of gastric POCUS was the main objective of the study, this was assessed using the percentage of patients with unsafe gastric content.Additionally, we wanted to assess the degree of correlation between directly measured gastric volumes and calculated gastric volumes via the nomogram and the formula.The decision-making after utilizing gastric POCUS is shown in Figure 3.

Statistical Analysis
Descriptive statistics were summarized as mean (AEstandard deviation [SD]) for continuous variables and as sample proportions for categorical variables.To measure agreement between methods producing above/below threshold decisions, we reported the Kappa statistic for agreement.To measure strength and direction of linear association for continuous measurements, Pearson's correlation coefficient was calculated.For all correlation coefficients, we examined scatterplots to visually inspect the data for outlier and influential observations.When outliers or influential observations were observed, a sensitivity analysis removing the outliers was conducted to observe the magnitude by which the observed statistics depended upon that subset of observations.A similar sensitivity  analysis was performed by including and excluding patients with high BMI (>35).Analyses were conducted using SAS 9.4 with each analysis using a Type I error rate of 0.05.No a priori sample size was calculated for this study design.

Results
A total of 124 patients were screened and approached for the study of which 83 patients were included while the remainder refused to participate (n = 37) or qualified for exclusion (n = 4; hiatal hernia: 2; peptic ulcer disease: 1, history of gastric surgery: 1).Of the 83 patients undergoing gastric ultrasonography, 2 patients had unsafe amount of liquid contents determined and 3 had presence of solid food contents thus giving an incidence of around 6% (n = 5/83) of unsafe qualitative US scans.One of the patients with solid content on scanning had a history of diabetes and liver transplantation and although we collected data on patient comorbidities, given that a majority of patients with or without comorbidities had low residual volumes, we did not explore the influence of comorbidities such as diabetes, renal failure, neuromuscular disorders, or liver disease on residual volume in our study.After these exclusions, a total of 78 patients were included for correlation of US-determined volumes and actual measured volumes.Patient characteristics, indication for endoscopy, and patient positioning for the procedure are provided in Table 1.A majority of patients underwent esophagogastroduodenoscopy (EGD) AE therapy (37.17%) followed by endoscopic ultrasound (EUS; 33.33%), Endoscopic retrograde cholangiopancreatography (ERCP; 21.79%), and upper double-balloon enteroscopy (DBE; 6.41%).Monitored anesthesia care (MAC) was utilized in a majority of patients (77% n = 60) while the rest of the study patients received general anesthesia (GA; 23%; n = 18).MAC practices mainly included propofol-based sedation with natural airway and in a few cases, based on procedural discomfort, fentanyl or midazolam was used in conjunction while in patients receiving GA, endotracheal tube was placed to secure airway along with inhaled anesthesia maintenance and muscle relaxation as necessary.In the two cases with ultrasound estimates with unsafe thresholds (outlier measurements), the cases were conducted under general anesthesia with rapid sequence induction similar for those patients who had solid contents evident in the US scans.Apart from using rapid sequence induction in these above cases, no deviation was found from the planned anesthetic option (MAC vs GA) in any of the cases.There were no intra or post-procedural airway related adverse events.Patient fasting characteristics and gastric ultrasound measurement summary measures are provided in Table 2. Qualitative assessments showed unsafe liquid volumes or solid contents in 5 out of 83 cases but was accurate in 4 cases (one case of unsafe amount of liquid content >450 mL and 3 cases of solid contents) but false in one case.The falsely determined outlier measurement was noted in one instance where the calculated volumes were falsely overestimated in comparison to the measured volume (Case number 83 where the volume was estimated to be around 240 mL by either calculations but actual measurement was 5 mL) and we believe this could be a sonographer error in identifying gastric sonoanatomy.Our study showed a moderate correlation between the formula based gastric ultrasound volume estimates and the measured gastric volume by endoscopy (r = .38,95% CI: 0.17, 0.55; P = .0004);as well as that between nomogrambased estimation of gastric volume and measured gastric volumes (r = .40,95% CI: 0.20, 0.57; P = .0002;Table 3 and Figure 4).This was further confirmed on observations of simple Kappa coefficient which again showed a moderate agreement between calculated measurement using formula and measured volumes (κ = 0.66 [95% CI: 0.038, 1.00]) as well as that between nomogram-based calculation and measured volumes (κ = 0.66 [95% CI: 0.0396, 1.00]).A sensitivity analysis for BMI > 35 (n = 15 patients) on residuals did not reveal any significant influence on the kappa estimates.

Discussion
Our prospective study in the cohort of patients undergoing advanced upper gastrointestinal endoscopy procedures showed that routine gastric POCUS  assessments are feasible in daily practice.Around 6% (n = 5/83) of the scans showed unsafe contents which changes the airway and anesthesia management plans for these patients.The utility of gastric POCUS scans was not only in the qualitative determination of gastric content type but also in quantitative ultrasound measurements to predict volumes that were above or below the safe threshold value for each patient.In contrast to validation studies on quantification of contents, our study showed only a moderate correlation with actual measurement of the contents despite low residual volumes in most subjects in the study.While no cases had underestimation of unsafe volumes, one case of false overestimation of unsafe thresholds was evident in our study highlighting the subjective component of ultrasonography.
Aspiration pneumonia has known to be directly causative or contributory to patient mortality in about 57,000 deaths in the United States and a majority of these happen in the context of a hospital admission. 18hile aspiration during planned elective procedures receiving anesthetic care is rare, 19 interestingly, it is being noted that there is an increase in the rate of aspiration events and associated adverse outcomes over the last decade and have a significant association with the increasing use of sedative techniques apart from inherent patient risk factors. 18Patients undergoing upper GI endoscopy procedures thus present with the combination of having upper gastrointestinal disorder where sedative techniques are employed, a unique scenario where both these high-risk factors for aspiration are encountered in the context where a definitive airway is not always employed.While gastric aspiration during upper GI endoscopies is still fortunately rare (4.6/10,000 cases), 3 it is difficult to ignore the fact that it is more common than in patients undergoing elective general anesthesia in the operating room environment.Generous use of deep sedation has along with an unprotected airway is not uncommon in this patient population and use of excessive sedation has a well-known association in the closed claims analysis on the aspiration cases during anesthetic care where patients undergoing MAC was the second most common anesthetic type associated with aspiration. 4 Aspiration of gastric contents in the context of anesthetic care is multifactorial including patient risk factors (difficult airway, respiratory, neurological, gastrointestinal diseases, medications, conditions that slow gastric motility like autonomic neuropathy, etc), procedural factors, fasting duration, and airway management choices among others.Several strategies that exist to decrease the risk of aspiration exist but are not fool proof.First and foremost, an important consideration pre-procedurally has been the adherence to the fasting guidelines 20 but, this does not necessarily guarantee an empty stomach as evident from studies showing variation in gastric emptying time between fasted patients and, in emergency scenarios, anesthesiologists often misdiagnose full stomach conditions. 5,21,22Additionally, from pediatric studies, it is evident that prolongation of fasting duration does not necessarily result in lower gastric residual volumes. 15,23Hence it is clear that fasting duration does not guarantee an empty stomach in all patients and perhaps an adjunct method to fasting duration to ensure an empty stomach by objective confirmation with gastric POCUS may seem like an attractive strategy.
Performing routine gastric POCUS exams seem tedious in a busy clinical practice and there is an argument to be made for performing Gastric POCUS exams only in high-risk individuals.Risk factor identification to stratify and perform gastric POCUS only in patients at a higher risk of either harboring unsafe residual gastric volumes or for aspiration events is an attractive strategy logistically but, such risk factor identification in regular practice is often flawed.Risk factors are commonly discounted at the time of pre-procedural assessment or airway management as evident from the NAP-4 project where 93% of patients having anesthesia-related aspiration events had risk factors for aspiration while the anesthesia caregivers estimated such risk factors in only 10% of these patients. 19Additionally, the occurrence of high gastric residuals may still commonly occur despite the absence of risk factors despite adequate fasting. 16,21Finally, the clinical scenario is complicated by the fact that even when high residual gastric volumes are identified, this does not necessarily lead to aspiration events. 16,21,23he point of care ultrasound has gained popularity in the perioperative anesthesia setting, 5,9,21,24 and has shown steep learning curve and utility among nonanesthesia providers such as emergency physicians 17 but the adoption of gastric POCUS into routine clinical practice has several challenges.First, gastric POCUS by itself is not necessarily confirmatory of an empty stomach as many conditions such as gastric bands or previous gastric surgery can spuriously show a low residual volume and hence its utility is probably more in cases where the qualitative scans reveal an unsafe volume or solid content.Hence, rather than being a replacement for fasting guidelines, an important tool to standardize care, the utility of gastric ultrasonography is more complementary to existing guidelines.A simplified example can be that the mere sonographic appearance of an empty stomach on gastric POCUS exam cannot rule out the risk of aspiration but, a documentation of solid or unsafe liquid content definitely rules out fasting state (confirms increased aspiration risk).
We did not aim to decrease the incidence of aspiration events as a result of gastric POCUS findings despite the fact that this subset of patients is at a higher risk of aspiration events compared with other patients receiving anesthetic care.We merely sought to examine the feasibility and utility of gastric POCUS before patients received anesthetic care for advanced upper GI endoscopy procedures and further identify if the calculated volumes truly reflected measured volumes.Several studies have evaluated the utility of gastric POCUS in the context of anesthesia care, a majority of them have utilized designs where fasting volunteers drank a premeasured quantity of liquids with subsequent measurement of gastric antral CSA. 8,10While the formula and the nomogram showed good correlation between estimated and measured volume in the first validation study of the technique, some studies have shown poor correlation of ultrasound measurements between observers 12,25 or between the estimated and measured gastric volume 11,14 and could be due to operator experience or the differences in the technique of delineating the cross-sectional area.While we found a strong correlation between nomogram and formula-based measurements, only a moderate degree of correlation was found between estimated volumes (nomogram or formula based) and measured volumes despite a majority of patients having low residual, and is an important finding of our study.Our study members had adequate expertise in gastric POCUS but more importantly, unlike other studies estimating after ingesting known volume of fluids, we measured actual residual volume by aspiration of gastric contents via the endoscope.While we did not have strong but only a moderate degree of correlation between estimated and measured volumes, our study findings are still in agreement with studies by Bouvet et al, 9 Kruisselbrink et al 24 and the validation study by Perlas et al 10 all of which utilized comparable study designs of aspirating residual gastric volumes to compare with US-based estimates, albeit in different patient populations.More importantly, our study demonstrated the utility of qualitative scans in patients undergoing upper gastrointestinal endoscopy procedures as two patients had unsafe liquid residuals and three patients had solid content despite adequate fasting status.We did not look at the impact of individual predictors of delayed gastric emptying given the pragmatic nature of our study design but some evidence seems to indicate that gastric POCUS may be more useful in patients with known risk factors of delayed gastric emptying.Garg et al in a study of residual gastric volumes in diabetics compared with non-diabetics after overnight fasting noted that while none of the study patients had unsafe residual volumes, some evidence of delayed gastric emptying in diabetic patients was noted in the form of greater residual gastric volume compared with non-diabetic patients.Another study 5 noted a higher incidence of significant residual despite 6-10 hours fasting in patients with underlying comorbidities like diabetes, renal failure, or those with higher BMI.While our study did not explore such association due to the incidence of low residual volumes in a majority of patients, perhaps future studies should consider evaluating the impact of individual predictors of delayed gastric emptying.Additionally, our study did not find the information of gastric ultrasound in change in the anesthetic plan except in opting for rapid sequence induction in those with unsafe content who were already scheduled for general anesthesia.This could reflect the institutional culture of opting for definitive airway in patients with risk factors or those undergoing procedures with risk of aspiration and perhaps the influence of gastric POCUS on determining anesthetic plan is needs to be studied.
There are several limitations to our study.The study design lends itself to performer and interpreter bias but the sonographic data was blinded to the gastroenterologist performing endoscopy.Perhaps the false estimation of an unsafe gastric volume in one of our patients may represent the impact of cognitive bias in image interpretation. 26While our overall data contained patients with various co-morbidities, the impact of individual comorbidities were not explored.
Finally, since the majority of the patients followed fasting guidelines, the dataset had measurements mainly in lower thresholds and, while ethically unfeasible, the data would have been much stronger if there were more data points with measurements above the normal threshold.

Conclusions
The gastric POCUS is a feasible tool to determine unsafe residual gastric contents in patients undergoing advanced upper GI endoscopy procedures.Qualitative scans were more useful in determining unsafe contents while volume assessments had moderate correlation with measured volumes.Routine gastric ultrasonography can guide airway management choices and improve safety in patients undergoing advanced upper GI endoscopic procedures.

Figure 1 .
Figure 1.Representative probe position in supine and right lateral decubitus positions.

Figure 2 .
Figure 2. A, Ultrasound images representing-A, empty stomach; B, liquid gastric content; C, early solid phase of gastric content.

Figure 3 .
Figure 3. Study algorithm for decision making after gastric POCUS.

Figure 4 .
Figure 4. Scatter plots comparing outcome measures-A, formula-based calculation vs measured volume; B, nomogram-based calculation vs measured volume; and C, formula-based calculation vs nomogram-based calculation of gastric residual volume.

Table 2 .
Patient Fasting Characteristics and Gastric POCUS Measurements

Table 3 .
Pearson's Correlation Coefficient for the Strength and Direction of Relationship Between Measured and Calculated Gastric Volumes