Association of enteral feeding with microaspiration in critically ill adults

Abstract

Aim:

This study explored relationships between enteral feeding and tracheal pepsin A.

Background:

Mechanically ventilated (MV) patients receiving enteral feeding are at risk for microaspiration. Tracheal pepsin A, an enzyme specific to gastric cells, was a proxy for microaspiration of gastric secretions.

Methods:

Secondary analysis of RCT data from critically ill, MV adults was conducted. Microaspiration prevention included elevated head of bed, endotracheal tube cuff pressure management, and regular oral care. Tracheal secretions for pepsin A were collected every 12 h. Microaspiration was defined as pepsin A ≥ 6.25 ng/mL. Positive pepsin A in >30 % of individual tracheal samples was defined as abundant microaspiration (frequent aspirator). Chi-squared, Fisher’s Exact test, and generalized linear model (GLM) were used.

Results:

Tracheal pepsin A was present in 111/283 (39 %) mechanically ventilated patients and 48 (17 %) had abundant microaspiration. Enteral feeding was associated with tracheal pepsin A, which occurred within 24 h of enteral feeding. Of the patients who aspirated, the majority received some enteral feeding 96/111 (86 %), compared to only 15/111 (14 %) who received no feeding. A greater number of positive pepsin A events occurred with post-pyloric feeding tube location (55.6 %) vs. gastric (48.6 %), although significant only at the event-level. Frequent aspirators (abundant pepsin A) had higher pepsin A levels compared to infrequent aspirators.

Conclusions:

Our findings confirmed the stomach as the microaspiration source. Contrary to other studies, distal feeding tube location did not mitigate microaspiration. Timing for first positive pepsin A should be studied for possible association with enteral feeding intolerance.

1. Introduction

Aspiration of oral and gastric contents is common in patients receiving mechanical ventilation (MV) and is associated with many complications: ventilator-associated events (VAE), pneumonia (VAP), acute respiratory distress syndrome (ARDS), pneumonitis, and other aspiration syndromes (Nseir et al., 2011; Son et al., 2017). Complications result in prolonged ventilation and hospitalization, and higher mortality (Zilberberg et al., 2020).

Despite preventative measures, microaspiration of gastric contents into the lower airway occurs in approximately half of MV patients (Jaillette et al., 2015; Jaillette et al., 2017; Metheny et al., 2006). The endotracheal tube which is required for ventilation holds the glottis in an open position, thereby creating a tract for leakage of gastric and/or oral secretions past the endotracheal tube (ETT) cuff (Rouźe et al., 2018). Multimodal standards of care to mitigate aspiration risk in MV patients include head of bed (HOB) elevation to 30–45°, regular oropharyngeal suctioning, and ETT cuff pressure management (20–30 cmH₂O) to prevent secretion leakage around the cuff (Taylor et al., 2016). Oral care, such as use of chlorhexidine, reduces bacterial colonization of oropharyngeal secretions that may leak into the patient’s trachea (Klompas et al., 2014).

Measurement of pepsin in oral and tracheal secretions is performed to assess gastric reflux and microaspiration respectively (Jaillette et al., 2015; Jaillette et al., 2017; Metheny et al., 2008; Talbert et al., 2021) Pepsin A, a subclass of pepsin, is a specific measure originating from the stomach. Stomach chief cells produce pepsinogen, which is activated by gastric acid and converted into its active enzymatic form, pepsin A (Hallal et al., 2015). Other pepsin subclasses, such as pepsin C, may originate from cells in the mouth and other locations (Hallal et al., 2015). Pepsin causes inflammation in direct contact with extra-esophageal epithelial cells (Bardhan et al., 2012). Minor reflux of gastric contents into the pulmonary system are typically neutralized and removed; however, these normal defense mechanisms are inefficient during episodes of critical illness (Klimara et al., 2020). Few studies have reported on the relationship between tracheal pepsin and enteral feeding (EF) in adult, critically ill patients (Metheny et al., 2006; Metheny et al., 2008; Schallom et al., 2013).

This paper reports on secondary data analysis from the NO-ASPIRATE study which focused on pulmonary outcomes (NCT02284178). The parent study examined the effect of a deep oropharyngeal suctioning intervention on microaspiration of oral secretions (Sole et al., 2019; Talbert et al., 2021). Tracheal amylase was the primary endpoint, and tracheal pepsin A was added partway through the study as a proxy measure for gastric microaspiration in a subset of patients (n = 297). This secondary analysis assessed the relationship of tracheal pepsin (microaspiration) with the primary outcome of EF and secondary outcome of distal feeding tube location. The university and study hospital institutional review boards approved the study and informed consent was obtained.

2. Methods

The parent study was performed in three intensive care units (ICUs) in a large academic medical center in Central Florida (Sole et al., 2019). Data including pepsin A samples were collected every 12 h for a maximum of 14 consecutive days.

2.1. Subjects

In the parent study, eligible patients were enrolled within 24 h of intubation and MV. Patients were excluded for: aspiration during intubation; rescue MV; re-intubation; lung, head, or neck cancer; or contraindications to suctioning. The parent study enrolled 513 patients with anticipated MV > 36 h (Sole et al., 2019). This secondary analysis includes data from 297 patients who had pepsin A measures.

2.2. Parent study procedures

The intervention group received deep oropharyngeal suctioning using a long, flexible suction catheter every 4 h (Sage, Cary, IL). The control group received a sham suctioning intervention every 4 h. Registered Nurse research assistants (RAs) completed interventions and data collection; both groups received antiseptic oral care with a suction swab every 4 h and tooth brushing every 12 h. Oral suctioning as needed was also a standard of care for all patients. HOB elevation (30°) was assessed every 4 h and ETT cuff pressure (20–30 cm H₂O) every 12 h using a calibrated manometer. Most patients (85 %) were intubated with a polyurethane subglottic suction ETT with a tapered cuff; the others had a traditional PVC ETT with a cylindrical cuff. Patient care decisions related to enteral feeding and distal feeding tube placement were made by the medical team as per usual care.

Data collection included demographic, physiological, and ventilation data. Paired oral and tracheal samples for pepsin A were obtained by RAs every 12 h. Pepsin A enzymatic assays developed by Krishnan et al. (2002) were modified using human data as positive controls and described in detail elsewhere (Talbert et al., 2021). At least one pepsin A value ≥ 6.25 ng/mL was considered positive for microaspiration; this was the lowest sensitivity level to differentiate positive from negative pepsin A (Talbert et al., 2021). Patients were classified as frequent aspirators (abundant microaspiration) when >30 % of individual tracheal samples were pepsin A positive (Nseir et al., 2019; Talbert et al., 2021).

2.3. Statistical analyses

For the secondary analysis, demographic data and select variables were reported by pepsin A status and descriptive statistics. Data were analyzed at the patient-level using Chi-Square (χ2) or Fisher’s Exact test and select variables were analyzed at the event-level (pepsin A) using generalized linear model (GLM), including logistic regression, in R (Version 4.0.2). In GLM, likelihood ratio tests (LRT) were performed to assess for differences in pepsin A status and other categorical variables. Several variables fluctuated between time points, such as enteral feeding and feeding tube location. Patient-level data were determined based on the status of individual variables at the majority of time points per patient unless specified otherwise. For example, patients who received EF for more than half of the time points were classified as EF = Yes. Wilcoxon rank sum test was used to compare infrequent and abundant aspirator groups as well as EF and non-EF groups due to continuous, non-normally distributed pepsin A values. A logistic regression model was fitted to examine the relationship of independent variables with abundant aspiration. Boxplots are reported as graphic summarization of pepsin A levels across groups, where median and interquartile range (IQR) are reported along with the boxplots. Significance level alpha was 0.05.

3. Results

Data were analyzed for 283 of the eligible 297 patients (14 removed for missing pepsin A values) with mean enrollment of 5 days (SD = 2.94). A total of 2408 tracheal samples were analyzed for pepsin A [M = 9 samples per patient (1–29)]. The majority were males (58 %) with an average age of 59.6 ± 18.9 years. Admission diagnoses included medical (35.7 %), surgical (6.4 %), trauma (25.8 %), and neurological (32.2 %). The sample was racially diverse: White (75 %), Black (18 %), Asian (6 %), and other (1 %). No baseline differences were noted between patients who were pepsin A positive vs. negative, including acuity measured by APACHE II.

Most patients received EF 222/283 (78 %) through a small-bore feeding tube or gastrostomy tube (PEG). None of the patients had a large bore gastric tube. Tracheal samples in more than one third of MV patients 111/283 (39 %) tested positive for pepsin A at least one time (Fig. 1). Enteral feeding was started at 2.6 days (SD = 0.94) and time to first positive pepsin A was 3 days (M = 3.09, SD = 1.65) following study enrollment (within 24 h of intubation/MV).

Fig. 1.

Pepsin A measures by patient.

Note: Abundant pepsin was calculated as individual patients with positive pepsin A in >30 % of their tracheal samples.

Feeding and related variables are reported according to pepsin A status by individual patient (Table 1). Enteral feeding was associated with positive tracheal pepsin A (χ² = 6.22, p = 0.01). This association held true when analyzed at the patient-level. Almost half of the patients who received enteral feeding most of the time aspirated (87/195). Of the patients who aspirated, the majority received some EF 96/111 (86 %), compared to only 15/111 (14 %) who received no EF (Fig. 2).

Table 1.

Demographics and sample characteristics by patient and pepsin A status (≥6.25 ng/mL).

Variable	Levels	Any positive tracheal pepsin A	No tracheal pepsin A	Total	Chi-squared test statistics	P value
Total	N (%)	111 (39 %)	172 (61 %)	283 (100 %)
Age	<50 years	34 (40 %)	52 (60 %)	86 (30 %)	<0.001	>0.999
	≥50 years	77 (39 %)	120 (61 %)	197 (70 %)
Sex	Female	43 (36 %)	76 (64 %)	119 (42 %)	0.61	0.434
	Male	68 (41 %)	96 (59 %)	164 (58 %)
Enteral feeding	Any enteral feeding	96 (43 %)	126 (57 %)	222 (78 %)	6.22	0.013
	No enteral feeding at any time point	15 (25 %)	46 (75 %)	61 (22 %)
Gastric residual volume	<250 mL	79 (43 %)	103 (57 %)	182 (64 %)	0.79	0.375
	≥250 mL	2 (100 %)	0 (0 %)	2 (0.7 %)
	Unknown	30	69	99
RASS	Deep sedation −3, −4, −5	55 (47 %)	62 (53 %)	117 (41 %)	5.28	0.071
	Light sedation +1, 0, −1, −2	53 (34 %)	102 (66 %)	155 (55 %)
	Anxious/restless/combative +4, +3, +2	3 (27 %)	8 (73 %)	11 (4 %)
Opioids	Yes	92 (43 %)	123 (57 %)	215 (76 %)	4.18	0.041
	No	19 (28 %)	49 (23 %)	68 (24 %)
Sedatives	Yes	50 (44 %)	63 (56 %)	113 (40 %)	1.66	0.198
	No	61 (36 %)	109 (39 %)	170 (60 %)
Paralytics	Yes	4 (4 %)	6 (6 %)	10 (4 %)	<0.001	>0.999
	No	107 (39 %)	166 (61 %)	273 (96 %)

Variables were assigned based on data occurring during the majority of time points unless specified otherwise. Richmond Agitation Sedation Scale (RASS).

Fig. 2.

Pepsin A status by any enteral feeding.

Note: Abundant pepsin was calculated as individual patients with positive pepsin A in >30 % of their tracheal samples.

Abundant microaspiration (positive pepsin A in >30 % of individual patient samples) was present in 48/283 (17 %) of the patients and in 48/111 (43 %) of the aspirators (any positive pepsin A). Abundant microaspiration was also associated with EF (χ² = 10.98, p < 0.001). Median pepsin A levels of frequent (abundant) aspirators were higher than infrequent aspirators, (12.8 ng/mL [IQR 9–23.7] vs. 10.6 ng/mL [IQR 8.25–16.9]; p < 0.001) (Fig. 3). Furthermore, 13/48 (27 %) of frequent aspirators did not receive any feeding at the time of positive pepsin A samples. When analyzed at the event level, positive pepsin A levels at the time of aspiration (≥6.25 ng/mL) were significantly higher at the time of no enteral feeding vs. at the time of enteral feeding, (Median 13.7 ng/mL [IQR 9.07–23.4] vs. 11.3 ng/mL [IQR 8.58–19.2], Wilcoxon rank sum test p = 0.04).

Fig. 3.

Pepsin A levels of abundant (frequent) vs. infrequent aspirators.

Median, upper and lower quartiles, of individual positive tracheal pepsin A samples (6.25 ng/mL) are displayed by patient aspiration status. Infrequent aspirators are patients who had positive pepsin A in fewer than 30 % of their individual tracheal samples and frequent aspirators (abundant) had positive pepsin A in >30 % of their individual tracheal samples. All data points were included in the analysis; however, several data points (200–400 ng/mL) in the frequent aspirator group were removed to create this figure. The total number of positive pepsin A events included in the analysis were infrequent aspirators (n = 120) and frequent aspirators (n = 187).

Logistic regression (n = 111 positive pepsin A) was used to examine the association between enteral feeding, opioids, and diabetes with abundant microaspiration. The odds of abundant aspiration were 80 % less in the patients who received enteral feeding (OR 0.20, 95 % CI, 0.06–0.55; p = 0.001) and patients receiving opioids were nearly four times as likely to be abundant aspirators (OR 4.79, 95 % CI, 1.36, 22.87; p = 0.012).

We analyzed feeding tube location in a subset of patients (n = 222) who received enteral feeding at any time (Table 2). Distal feeding tube location was associated with aspiration at the patient-level, (χ² = 5.88, p = 0.03). The majority of patients had post pyloric feeding tubes 140/222 (63 %) while others had small-bore gastric 76/222 (34 %), or gastrostomy (PEG) tubes 6/222 (3 %). When distal feeding tube location was analyzed at the event-level, it was associated with aspiration events (LRT = 7.642, df = 3, p = 0.05) and remained significant when PEG tubes were removed from the analysis. Positive pepsin A was measured in greater than half of the events with post-pyloric tubes (55.6 %) vs. gastric feeding tubes (48.6 %). None of the frequent aspirators had PEG tubes.

Table 2.

Distal feeding tube position by patient and pepsin A status (≥6.25 ng/mL).

Variable	Any positive tracheal pepsin A		No tracheal pepsin A		Total		Fisher’s exact test statistic	P value
	n	%	n	%	n	%
Total	96	43	126	57	222	100
Gastric	30	40	46	61	76	34	5.879	0.03
Post-pyloric	66	47	74	53	140	63
Gastrostomy tube (PEG)	0	0	6	100	6	3

Sub-set analysis of 222 patients who received any enteral feeding.

Opioid administration was also associated with aspiration. Of the patients who received opioids most of the time, fewer than half aspirated 92/215 (43 %). Of the patients who aspirated, a larger proportion of patients received opioids 92/111 (97 %) compared to those who did not receive opioids 19/111 (17 %). No statistical difference was observed with sedative administration. When analyzed at the event-level, a greater proportion of aspiration events occurred when no opioids or sedatives were administered (both 53 %) vs. at the time that opioids or sedative were administered (both 47 %). No association was observed in Richmond Agitation Sedation Scale (RASS) at the patient-level or event-level. Of the patients who aspirated, the majority were either deeply sedated 55/111 (50 %) or lightly sedated 53/111 (48 %). A greater proportion of positive pepsin A measures (55 %) occurred during light sedation at the event-level. Very few patients were restless or combative.

GRV could only be analyzed for a subset of patients (N = 184) due to missing data. Of these, 99 % of patients had GRVs < 250 mL and there was no association with pepsin A status. High GRVs > 250 mL were only experienced by two patients, both of whom aspirated. No association was found between abundant aspiration and GRV.

4. Discussion

Our study adds to the body of knowledge on microaspiration in MV patients based on the following findings: the relationship of tracheal pepsin A and enteral feeding, feeding tube location, abundant aspiration, timing of microaspiration, pepsin A levels, and opioid/sedation administration. This secondary analysis was underpowered to assess for effects of pepsin A on enteral feeding variables because the parent study was powered for pulmonary outcomes and not focused on enteral nutrition. The estimated effect size for positive pepsin A = 0.06; therefore, a total sample of 400 would allow us to detect differences in mean pepsin A levels (alpha level 0.05, 80 % power). We reported most of our findings at the patient-level, although a few key outcomes were also reported at the event level to ensure adequate sample size. Event-level results should be interpreted cautiously because the effect size may be artificially inflated due to increased power, n = 2408 vs. patient level n = 283.

Our findings were consistent with what is known in the literature. We confirmed that a high percentage of MV patients experienced microaspiration of gastric secretions (39 %), which was higher in patients who received any EF (43 %) vs. patients who received no EF (25 %). Previous studies of EF reported similar pepsin findings in critically ill patients, although the specific gastric protease was not identified and assumedly included both pepsin A and C (Dewavrin et al., 2014;Metheny et al., 2006; Schallom et al., 2015). Our study did not measure pepsin C, which is also located in oral mucosal cells (Hallal et al., 2015). Pepsin A, a proteolytic enzyme specific to the stomach, was used as a proxy measure of microaspiration (Hallal et al., 2015; Krishnan et al., 2002). Although an association between EF and microaspiration was reported previously (Metheny et al., 2008; Schallom et al., 2015), the measurements of pepsin A as a biomarker confirm the source of aspiration as gastric in origin. Due to variable assay methods, no reference standard exists for Pepsin A measures (Nseir et al., 2011).

Although enteral nutrition has been associated with microaspiration in other studies, the clinical importance of microaspiration of gastric pepsin A into the trachea remains unknown. Recent studies have ruled out acid alone as an independent source of airway inflammation, yet concluded that significantly more cellular damage occurs when pepsin is present (Bulmer et al., 2010; Hurley et al., 2019). Evidence suggests that pepsin is a mediator for inflammatory and neoplastic disease when exposed over time (Klimara et al., 2020). An analysis of pepsin A in a larger subset of our study population showed a relationship between aspiration and VAC, although VAP was non-significant (Talbert et al., 2021). Using calculated measures for ARDS (PaO₂/FiO₂ ratio, Berlin criteria), positive pepsin A was associated with moderate and severe ARDS suggesting that the inflammatory action of pepsin A may have been a contributing factor (Talbert et al., 2021). Nseir et al. found less abundant gastric microaspiration in patients who were enterally fed vs. parenterally fed, although both groups experienced gastric microaspiration; no differences in patient characteristics were reported (Nseir et al., 2019).

Metheny suggested that abundant aspiration may be a more valid measure of microaspiration compared to positive pepsin alone, to avoid over-reporting small volume microaspiration events (Metheny et al., 2008). Definitions for abundant microaspiration vary widely. Abundant microaspiration has been reported as positive pepsin in >25–74 % of specimens obtained from individual patients (Dewavrin et al., 2014; Metheny et al., 2006; Metheny et al., 2008; Nseir et al., 2011; Nseir et al., 2019; Talbert et al., 2021). We defined abundant or frequent aspirators as patients who tested positive for pepsin A in >30 % of their individual tracheal samples. Our findings indicated that patients who were abundant aspirators had significantly higher pepsin A levels compared to those who were infrequent aspirators. We are unaware of other studies reporting this finding and it is unknown whether pepsin A levels have clinical significance. We found a high percentage of abundant aspirators 13/15 (87 %) in MV patients who did not receive EF, yet aspirated. Confirmed by logistic regression, the patients who were not fed were more likely to have abundant aspiration. Additionally, the median pepsin A levels at the time of aspiration were higher at the time of no enteral feeding vs. enteral feeding. Our results suggest that abundant aspirators may be a specific subset of patients, such as those with gastrointestinal (GI) dysfunction, including EF intolerance (EFI) and/or pre-existing gastroesophageal reflux disease (GERD), which were not measured in our study. Diabetes was associated with higher pepsin levels in older adults receiving enteral feeding but was not significant in our sample (Ding et al., 2021). GERD was associated with positive pepsin in several outpatient studies (Klimara et al., 2020; Upendran et al., 2020). Gastroesophageal reflux occurs in healthy individuals but the frequency and/or duration are increased in the critically ill (Schörghuber & Fruhwald, 2018).

We may be the first to report on the timing of positive pepsin A events in relation to initiation of MV and onset of EF. The average time to first positive pepsin A event was 3 days following study enrollment, which was within 24 h of MV. We likely missed the majority, if not all, of intubation-related aspiration events due to the short half-life of pepsin A in the trachea. Because our EF data were collected every 12 h, we were able to detect aspiration within 24 h of EF in some patients, suggesting that the relationship between microaspiration and EFI should be further explored. EFI occurs in more than one third of MV patients (Reintam Blaser et al., 2014; Reintam Blaser et al., 2021) within 3 days (median) of EF (Gungabissoon et al., 2015). EFI’s delayed gastric emptying is associated with increased length of stay and mortality (Reintam Blaser et al., 2014; Reintam Blaser et al., 2021).

Common measures to assess for microaspiration include alpha-amylase and pepsin assays of tracheal secretions (Griton et al., 2021; Metheny et al., 2011; Nseir et al., 2019; Sole et al., 2019); both are quantifiable and easy to use in routine practice (Nseir et al., 2021). A limitation to both biomarkers is the unknown timing for clearance from tracheal secretions (Nseir et al., 2021). Amylase originates in the saliva, specifically measuring microaspiration from an oral source (Sole et al., 2020) versus, pepsin which measures both gastric and oral sources, and pepsin A which is specific to the stomach. Perhaps point of care testing will be available in the future if pepsin A can be established as an important biomarker for microaspiration.

Contrary to other studies, we found a difference in distal feeding tube location when we analyzed at both the patient-level and event-level. We found a significant increase in aspiration with post-pyloric feeding tube placement. Previous studies found an association between post-pyloric feeding tube placement (duodenum or jejunum) and reduced risk for aspiration in the critically ill (Heyland et al., 2001; Metheny et al., 2011), yet our results did not support these findings. Heyland et al. (n = 33) found decreased gastroesophageal regurgitation and a trend towards decreased aspiration in duodenal tubes, measured by oral and tracheal radioisotopes (Heyland et al., 2001). Metheny et al. (n = 428) reported 12–18 % fewer positive tracheal pepsin with post-pyloric vs. gastric feeding tubes (Metheny et al., 2011). Our study showed more positive pepsin A events at the time of post-pyloric feeding tube position. At the event-level, we performed statistical analysis using GLM which is a robust test for repeated measures designs. Distal feeding tube position was not consistent within patients across time. Feeding tubes frequently migrate from the initial gastric placement to a post-pyloric placement within 24 h following insertion (Bourgault et al., 2020). At the patient-level we classified feeding tube location by the distal tube position held for the majority of time. Positive pepsin A was measured in a greater number of post-pyloric tubes vs. gastric tubes at both the patient and the event-level. We considered that pepsin A vs. pepsin measures used in other studies may have contributed to the difference in findings, yet our overall microaspiration rates using pepsin A as a biomarker were consistent with microaspiration measured by pepsin in the literature. Post-pyloric feeding tubes are typically recommended for patients at high risk for/or diagnosed with EFI to reduce risk for gastric reflux and microaspiration (Alkhawaja et al., 2018; Reintam Blaser et al., 2017; Singer et al., 2019). High acuity has been associated with increased risk for EFI (Elke et al., 2015), which may increase risk for gastric microaspiration, yet we found no difference in acuity levels between the patients who aspirated vs. those who did not. No aspiration was experienced by patients with PEG tubes, although the volume was low.

No prior relationship between GRV and EFI has been established, likely because GRV cutoffs (150–500 mL) and measurements vary widely (Reintam Blaser et al., 2014). At the time of the parent study, GRV was being performed as part of usual practice and measures of ≥250 mL were considered high. Although no consistent relationship was found between pepsin and GRV, Metheny et al. reported increased aspiration frequency as GRVs increased; aspiration risk was higher when patients had two or more GRVs ≥ 200 mL or one or more GRVs ≥ 250 mL (Metheny et al., 2008). Our study found no relationship between pepsin A and GRV, possibly because the majority of our patients had low GRV measures (<250 mL) and only 2 patients experienced high GRVs. Approximately one third of our GRV data were missing, assumedly due to difficulty obtaining aspirate from post-pyloric tubes. High GRV, redefined by practice guidelines as GRV volume ≥ 500 mL over a six-hour period (McClave et al., 2016; Metheny, 2021; Singer et al., 2019), is often used as a marker for EFI in combination with other vague clinical signs, such as vomiting, diarrhea, abdominal distension, and unexplained abdominal pain (Gungabissoon et al., 2015; Reintam Blaser et al., 2021). Routine GRV monitoring is no longer recommended because it was a poor indicator of EFI and/or aspiration risk and its use frequently resulted in unnecessary nutritional interruptions which has been associated with increased mortality (Martindale et al., 2020; Reignier et al., 2013). In our study, only two patients had high GRVs (≥250 mL) and they both aspirated. While this is clinically interesting, our data were insufficient to be meaningful.

Previous studies showed an association between tracheal pepsin and heavy sedation and opioid administration (Metheny et al., 2006; Metheny et al., 2008) Only opioids were significant at the patient level in our study. Of the patients who aspirated, 97 % received opioids the majority of time. Without event-level analysis, it is not known if opioid administration occurred at the time of aspiration. When analyzed at the event-level, we found more aspiration events in the absence of opioid and sedative administration at the time of data collection. Sedation levels by RASS were not significant with either analysis, although a greater proportion of aspiration events occurred during periods of light sedation when analyzed at the event-level. Previous studies have shown that body position or routine nursing practices, such as oral care have been associated with ETT movement and leakage of oral secretions past the ETT cuff into the trachea (Chair et al., 2020; Kim et al., 2009). Patient movement may occur more frequently in lightly sedated patients compared to those who are less anxious or sedated, possibly contributing to increased ETT movement and microaspiration (Chair et al., 2020).

Lying flat in bed has also been associated with positive tracheal pepsin (Metheny et al., 2002). We are unable to determine if there was any variation in HOB elevation in the parent study; therefore, we could not assess for an association between HOB and positive pepsin A. The standard of care was to always elevate the HOB to 30°, which was verified every 4 h. Recommended HOB elevation to reduce risk of aspiration for enterally fed, MV patients is 30–45° (Taylor et al., 2016). HOB elevation is often underestimated (Martí-Hereu & Arreciado, 2017) and not consistently maintained; variation may occur due to transport off the unit, or during procedures and basic patient care (Lizy et al., 2014). The multimodal preventative measures controlled during the parent study, HOB elevation to 30°, maintenance of ETT cuff pressures at 20–30 cmH₂O, and regular oral suctioning (Leasure et al., 2012; Rawat et al., 2017) should be part of standard care to minimize risk for gastric microaspiration. Other measures such specialized endotracheal tubes with subglottic suctioning are thought to mitigate the risk for microaspiration of secretions from the oral cavity (Chair et al., 2020; Rawat et al., 2017), yet a small RCT found no difference in tracheal amylase when subglottic suctioning was used (Griton et al., 2021). Although routine oral care with CHG has become a practice standard, recent studies have demonstrated that CHG may cause more harm than good (Blot et al., 2022; Dale et al., 2021; Hellyer et al., 2016).

Fifteen patients who were not enterally fed at the time of their tracheal samples experienced aspiration and most 13/15 (87 %) were frequent aspirators. Because the parent study focused on a pulmonary intervention and outcomes, variables such as gastroesophageal reflux (GERD), EFI, vomiting, and coughing were not collected. In this small volume of aspirators who were not receiving feeding at the time of their aspiration, no associations were found with other EF variables. Data were also unavailable on chronic conditions such as GERD and laryngopharyngeal reflux (LPR), which have both been associated with microaspiration (Klimara et al., 2020). Other possible contributing factors to gastric microaspiration which were not directly assessed during this study include loss of integrity of the lower esophageal sphincter, gastric distension, and/or oropharyngeal dysphagia due to desensitization of the pharyngoglottal adduction reflex (Kostadima et al., 2005). Feeding characteristics of abundant aspirators should be explored in a larger sample.

Most studies suggest that positive pepsin events are under-identified. Microaspiration events may be missed due to pepsin’s short half-life in the lungs and feasibility limitations to obtain timely samples. Pepsin prefers the acidic environment in the stomach (pH < 5.5) (Bardhan et al., 2012; Kahrilas & Kia, 2015). In the more alkaline tracheobronchial environment (pH 7.5), pepsin can maintain its proteolytic function for an undefined time before it becomes inactive, yet it has remained stable to pH 8 (Johnston et al., 2007; Piper & Fenton, 1965). Tracheal pepsin was detected up to 4–6 h in rabbits (Metheny et al., 2004) and oral pepsin was detected up to 90 min in humans (Knight et al., 2005). In the laryngopharynx, pepsin becomes inactivated at a pH of 6.5 but remains stable up to 24 h and can reactivate if the pH is decreased, such as a subsequent reflux event (Johnston et al., 2007; Piper & Fenton, 1965). It is theorized that pepsin may reactivate inside of the epithelial cells due to lower intracellular acidity, thereby causing cellular level inflammatory changes (Johnston et al., 2007). Tracheal samples in the parent study were obtained every 12 h, suggesting that we missed microaspiration events.

Although more frequent sampling would be beneficial from a measurement perspective, obtaining tracheal specimens every 2 or 4 h is not feasible, since samples must be obtained via endotracheal suctioning or bronchial alveolar lavage (BAL), which is more invasive compared to oral sampling. Also, pepsin A is unstable at room temperature, so continuous sampling methods are ruled out. It is unknown whether inflammatory changes to tracheal endothelium from pepsin A are related to dose (pepsin level), frequency, or cumulative effects of each over time. Our findings suggest that future research should explore abundant microaspiration, and the relationship of the timing and dose of microaspiration with EF, and capture variables related to EFI and GERD. Patients with gastric dysfunction may require additional aspiration prevention interventions. Additional research should focus on patient outcomes related to pulmonary inflammation, such as ARDS, acuity level, and length of stay to better understand the effects of microaspiration in this population.

4.1. Application to nursing practice

This study reinforces the need to maintain mitigation efforts to prevent gastric microaspiration in MV patients, whether they are receiving enteral feeding or not, such as head of bed elevation 30–45°, airway cuff pressures 20–30 cm H₂O, and regular oropharyngeal suctioning. Despite bundled mitigation efforts, 43 % of MV adults receiving enteral feeding experienced gastric microaspiration measured by pepsin A. Strategies such as improved securement of ETTs may help to mitigate movement of the tube and possible aspiration. Our study suggests that distal feeding tube location does not prevent gastric microaspiration in this population. Microaspiration also occurred in MV patients who were not receiving enteral feeding. Our findings also reinforced the practice to avoid routine GRV monitoring. The average onset of microaspiration was within 24 h of starting enteral feeding and frequent aspirators had higher levels of pepsin A. Timing of aspiration onset might suggest association with EFI, but our secondary dataset did not contain variables to assess for this. Nurses should be aware of other signs and symptoms of EFI, such as nausea, vomiting, diarrhea, unexplained abdominal pain, and/or abdominal distention and alert the interdisciplinary team if EFI is suspected (Reintam Blaser et al., 2021).

4.2. Limitations

Because this was a secondary data analysis, we were limited to the methods and data collected in the parent study which focused on pulmonary outcomes; therefore, some variables related to EF were not available. Secondary data analysis can introduce bias because the research questions and analyses are developed after completion of the original study. No mechanism was available in the parent study to monitor HOB elevation and ETT cuff pressure between data collection periods, so it is possible that there may have been some undocumented variation (Martí-Hereu & Arreciado, 2017). As frequently seen in clinical research, some variables changed at the different data collection points, such as feeding tube location, formula rate, and sedation, which complicated analyses at the patient-level. Due to the variation of several independent variables at the patient-level and suboptimal sample size, generalized linear mixed model could not be performed. Generalized linear mixed model is a more robust test for repeated measures design because it evaluates the temporal effects of the variables and can detect variances in variables that have small effect sizes.

5. Conclusion

Our analysis confirmed the relationship of gastric microaspiration and EF in MV ventilated patients using tracheal pepsin A (≥6.25 ng/mL) as a biomarker. More than one third of the sample aspirated (34 %), yet a greater proportion of the patients who aspirated (43 %) were receiving EF. Pepsin A levels at the time of aspiration events were higher without feeding vs. feeding. Only 25 % of patients without any EF aspirated; however, most of these patients had abundant aspiration, suggesting a possible association with gastric dysfunction such as EFI or a pre-existing condition such as GERD. Additionally, pepsin A levels in tracheal samples of abundant aspirators were higher compared to infrequent aspirators. In contrast to other studies, more aspiration events occurred in tubes with post-pyloric placement. At the event level, most aspiration episodes were captured in the absence of opioids or sedatives, although RASS scores were not significant. Research with a larger sample size, including a more expansive set of EF variables and pre-existing conditions influencing gastric microaspiration is recommended.