Verification of Inefficacy Of the Glucose Method In Detecting Aspiration Associated with Tube Feedings

Abstract

While some authors believe that testing for glucose in suctioned tracheal secretions can be used to detect aspiration of glucose-containing formula, others disagree. Previous evaluative studies of the glucose method’s efficacy have lacked adequate statistical power and a gold standard for aspiration. In this animal study, a gold standard for aspiration was used and possessed sufficient statistical power to address the glucose method’s sensitivity and specificity. As such, the results from the study provide the clinician with useful data to decide if the glucose method is appropriate for use in clinical settings.

If repeated aspirations of tube feedings are allowed to occur, the risk for aspiration pneumonia is increased significantly. Hence, clinicians strive to detect aspiration in its early stages so that interventions can be implemented to prevent further aspirations. In the past, blue food dye commonly was added to enteral feedings so that clinicians could observe for blue-tinged tracheal secretions during suctioning. However, this method is unreliable and possibly harmful. Thus, clinicians sometimes test for glucose in suctioned tracheal secretions to try to detect the aspiration of glucose-containing formula. However, there is little research-based information that supports the use of this method; previous evaluative studies have lacked a gold standard for aspiration and adequate statistical power. In this animal study, a gold standard for aspiration was used and possessed sufficient statistical power to address the glucose method’s sensitivity and specificity.

Literature Review

Studies have reported differing views regarding the efficacy of the glucose method in detecting aspiration. One concluded that it is appropriate to use glucose oxidase reagent strips to detect the presence of glucose in tracheal secretions (Metheny, St. John, & Clouse, 1998), while others concluded that the presence of glucose in tracheal secretions does not signal pulmonary aspiration of enteral feedings (Kinsey, Murray, Swensen, & Miles, 1994; Phillips, Meguer, Redman, & Baker, 2003). Some clinicians believe that finding glucose in suctioned tracheal secretions signals the aspiration of glucose-rich enteral feedings (Davis, Arrington, Fields-Ryan, & Pruitt, 1995; Dever & Johanson, 2004; Hussain, Roy, & Young, 2003; Potts, Zaroukian, Guerrero, & Baker, 1993;). However, the literature offers ample diasagreement from others (Elpern, Jacobs, Tangney, & Bone, 1986; Kinsey et al., 1994).

The actual effectiveness of testing for glucose in tracheal secretions to detect aspiration has been difficult to determine because most evaluative studies have lacked adequate statistical power and because a gold standard comparison for aspiration has not been identified (Kinsey et al., 1994; Potts et al., 1993; Winterbauer, Durning, Barron, & McFadden, 1981). Several studies evaluated glucose oxidase reagent dipsticks (Potts et al., 1993; Winterbauer et al., 1981), while another evaluated a laboratory glucose assay (Kinsey et al., 1994).

Establishing the efficacy of the glucose method is considered a worthwhile effort because a reliable bedside test remains an important clinical need. For this reason, an experimental animal model study was undertaken to meet the following objectives:

1. Determine:

   1. the sensitivity of glucose in detecting three forced small-volume aspirations of enteral formula mixed with human gastric juice.

   2. the specificity of glucose in detecting three forced small-volume aspirations of 0.9% sodium chloride solution.

   3. the relationship of the sensitivity of glucose to:

      1. glucose concentrations of the enteral formula (low, moderate, and high).

      2. the cumulative effects of three successive 30-minute forced aspirations, each separated by 90-minutes.

2. Determine the extent to which 2-hour, 4-hour, and 6-hour glucose readings of tracheal secretions were affected by five independent predictor variables.

   1. Glucose concentration of the enteral formulas.

   2. Glucose concentration of the human gastric juice.

   3. Arterial blood glucose level measured at baseline and hourly.

   4. Visible blood in suctioned tracheal secretions.

   5. Volume of tracheal secretions.

Methods

Subjects

The primary subjects were 182 New Zealand white rabbits that weighed approximately 3 kg each (161 experimental and 21 control animals). New Zealand rabbits were chosen because their lungs have no pathogens; thus, these animals are the customary model for pulmonary studies. A sample size of 182 was chosen to achieve sufficient statistical power to address the study’s objectives. In addition, gastric juice was collected from 161 acutely ill humans to mix with the eight enteral formulas. Hyperglycemia was present in approximately 15% of the acutely ill individuals who furnished the gastric juice used in the study. This study was conducted in an animal research laboratory at Saint Louis University School of Medicine. Approval for the inclusion of human subjects was obtained from the Saint Louis University Human Subjects Committee; the animal experiments were approved by the Saint Louis University Animal Care Committee.

Variables

Independent variables

The primary independent variable was the glucose content of eight enteral formulas, in addition to a control solution of 0.9% sodium chloride. The 182 animals were assigned randomly to one of the eight enteral formulas or saline control solution (n = 19 to 21 per group). Researchers selected eight enteral formulas to represent low glucose content (three formulas), moderate glucose content (two formulas), and high glucose content (three formulas). Formulas with different glucose concentrations were chosen because proponents of the glucose method to detect aspiration have recommended using an enteral formula with a relatively high glucose concentration of at least 200 to 300 mg/dL (Potts & Zaroukian, 1995). Glucose concentrations in the formulas used in the current study ranged from < 20 mg/dL to 466 mg/dL (see Table 1).

Table 1.

Estimated Glucose Concentration in Enteral Formulas Used in Study and Number of Animals (N) Receiving Each

Formula	Estimated Glucose Concentration (mg/dL)	Number
Low Glucose Content
Glucerna® (Ross Laboratories, Columbus, OH)	< 20	20
Isocal® (Mead Johnson Nutritionals, Evansville, IN)	60	21
Jevity® (Ross Laboratories, Columbus, OH)	95	20
Moderate Glucose Content
Pulmocare® (Ross Laboratories, Columbus, OH)	134	20
Osmolite® (Ross Laboratories, Columbus, OH)	261	21
High Glucose Content
Pediasure® (Ross Laboratories, Columbus, OH)	354	20
Two Cal HN® (Ross Laboratories, Columbus, OH)	401	20
Traumacal® (Mead Johnson Nutritionals, Evansville, IN)	466	19

The second independent variable was the glucose content of the human gastric juice that was mixed with the enteral formula. On the day preceding each experiment, human gastric juice was collected from the nasogastric tubes of 161 acutely ill hospitalized adults; all of the patients had been fasting for at least 4 hours and no medications had been received enterally within the hour preceding specimen collection. Split samples were used to measure the glucose concentrations in 157 of the 161 specimens by an automated glucose analyzer and glucose oxidase reagent strips. Because the readings were highly correlated (r = 0.96, p < 0.001), only readings from the glucose oxidase reagent strips are reported. While 50 (28%) of the gastric juice specimens contained no glucose, the mean glucose concentration in the remaining 107 specimens was 99.1±16.6 mg/dL (range 5 to 800 mg/dL). The gastric juice was refrigerated until the following morning, when it was mixed in equal amounts with one of eight enteral formulas. The eight enteral formulas did not differ according to the amount of glucose in the gastric juice with which they were mixed, F (7,149) = 1.69, p = 0.12.

Procedure

The animals were anesthetized, intubated with a 3.5 mm uncuffed endotracheal tube, and mechanically ventilated with a pediatric ventilator (VIP BIRD Infant Pediatric Ventilator; Minster, OH). An attached warm air respiratory humidifier (Fisher & Paykel Healthcare; Auckland, New Zealand) was used to facilitate the collection of secretions during suctioning. Continuous monitoring for hemodynamic and acid-base status was performed. A tidal volume of 50 mL with a positive end-expiratory pressure of 2 cm H₂O was used in all animals, with the frequency altered to control acid-base balance. Arterial glucose measurements were made at baseline and hourly thereafter by glucose oxidase reagent strips. A volumetric infusion pump was used to instill three separate boluses of fluid (either human gastric juice mixed with enteral formula, or 0.9% sodium chloride solution) intratracheally via a 1.22 mm catheter introduced through the endotracheal tube into the mainstem bronchus.

At the beginning of each experiment, 0.4 ml/kg of the fluid was infused over a 30-minute period. The infusion then was stopped, and 90 minutes were allowed to elapse before endotracheal suctioning was performed with a 6.5 Fr. catheter attached to a 20-mL pediatric mucus trap. The 90-minute period was selected to allow dilution of the infused fluid by local respiratory secretions. The suctioned tracheal secretions were inspected visually for blood before being taken to a research laboratory where the glucose concentration was measured by an automated glucose analyzer and glucose oxidase reagent strips.

At hour 2, an additional 0.4 ml/kg of the appropriate fluid was infused over a 30-minute period. Again, the infusion was stopped and 90 minutes were allowed to elapse before endotracheal suctioning was performed. At hour 4, this process was repeated. Thus, by the end of the 6-hour experiment, each animal had received intratracheally a total volume of fluid (either enteral formula mixed with gastric juice, or 0.9% sodium chloride solution) equivalent to 1.2 mL/kg body weight. Normal saline was not instilled intratracheally during any of the suctioning procedures to avoid diluting the secretions.

Measurements

Tracheal secretions’ glucose (TSG)

All of the suctioned tracheal secretions were analyzed for glucose by two methods (an automated glucose analyzer and glucose oxidase reagent strips). Because split samples from the 541 specimens had highly similar glucose values (r = 0.98, p < 0.001), only readings from the glucose oxidase reagent strips are reported. The TSG served as the dependent variable.

The automated glucose analyzer used in the project had a linear range of readings from 0 to 500 mg/dL (Yellow Springs Instruments, Model 2300; Yellow Springs, OH). When a sample with a glucose concentration of 500 mg/dL was encountered, it was diluted and analyzed again. The concentration was then calculated according to dilution factor.

Chemstrip bG (Boehringer Mannheim; Indianapolis, IN) glucose oxidase reagent strips were used to make the visual glucose readings. First, a drop of the specimen was placed on the two test pads of the glucose oxidase reagent strip. After 2 minutes had elapsed, the colors of the test pads were compared visually to the chart provided by the manufacturer; color gradations on the chart corresponded to glucose concentrations of 20, 40, 80, 120, 180, 240, 400, and 800 mg/dL. If the colors on the test pad were darker than those for 240 mg/dL, an additional 60 seconds were allowed to elapse before the final reading was made. All of the visual glucose interpretations were made by the same registered nurse research assistant.

Blood glucose (BG)

Only glucose oxidase reagent strips (Chemstrip bG) were used to measure hourly arterial blood glucose concentrations. Per manufacturer directions for measuring glucose in blood, the closest match to the lower pad on the test strips was made, and then the closest match to the upper pad was made; the two numbers were averaged for an estimated blood glucose reading.

Visible blood in tracheal secretions

The mucus trap in which each tracheal specimen was collected was held against a white background and visually inspected for blood. Blood was said to be either present or absent; no attempt was made to quantify the amount of blood present. All of the observations for blood in the tracheal secretions reported in this study were made by the same registered nurse research assistant prior to performing the visual glucose test.

Data Analysis

Various threshold values for TSG as a marker for aspiration of glucose-containing enteral formula have been described (Potts et al., 1993; Winterbauer et al., 1981). Researchers estimated sensitivity and specificity of the glucose oxidase reagent strips by using four glucose threshold concentrations: 20 mg/dL, 40 mg/dL, 60 mg/dL, and 80 mg/dL. Sensitivity was estimated using TSG data from the 161 experimental animals that received intratracheal instillations of enteral formula plus gastric juice. Specificity was estimated for each of the four threshold values using TSG data from the 21 control animals.

The percentage of missing TSG data for the 161 experimental animals was low, ranging from a high of 6.2% at 2 hours to a low of 1.9% at 4 and 6 hours. For the 21 control animals, the percentage of missing TSG data was 28.6% at 2 hours, 33.3% at 4 hours, and 9.5% at 6 hours. Nearly all of the missing data occurred because of inadequate volume of the tracheal secretions to perform the glucose analyses.

A 3 × 3 repeated measures analysis of variance was used to determine the effects of formula glucose concentration (low, moderate, and high) and time of measurement (2, 4, and 6 hours) on sensitivity. To examine what predictors explained variation in TSG measurements in the 161 experimental animals at 2, 4, and 6 hours, investigators used hierarchical multiple regression analysis for each time period. For each regression analysis, only animals with non-missing data for the TSG dependent variable and at least 75% of the predictors were used. Missing predictors were estimated with the overall sample mean. To examine predictors of 2-hour TSG, researchers entered formula glucose concentration at Step 1 and gastric glucose concentration at Step 2. At Steps 3, 4, and 5, they entered baseline, 1-hour, and 2-hour blood glucose concentration, respectively. At Steps 6 and 7, they entered presence of visible blood in tracheal secretions and volume of tracheal secretions. To examine predictors of 4-hour TSG and 6-hour TSG, researchers entered the same predictors at Steps 1, 2, 3, 4, and 5 as for 2-hour TSG. For 4-hour and 6-hour TSG, 3-hour and 4-hour blood glucose values were entered at Steps 6 and 7. For 6-hour TSG, 5-hour and 6-hour blood glucose values were entered at Steps 8 and 9. Visible blood and volume of tracheal secretions at the hour of the TSG measurement were entered at the last two steps.

Results

Sensitivity of the glucose method varied according to the threshold TSG concentration (20, 40, 60, or 80 mg/dL) and the time of data collection (2 hours, 4 hours, and 6 hours). As shown in Table 2, sensitivity worsened (and specificity improved) as the TSG threshold value increased. Significant declines in sensitivity were accompanied by significant declines in mean glucose concentrations in the tracheal secretions across each 2-hour period (p < 0.01); the mean glucose concentration in the experimental animals’ tracheal secretions was 118.1 ± 54.7 mg/dL at hour 2, 104.4 ± 58.8 mg/dL at hour 4, and 79.3 ± 52.6 mg/dL at hour 6. Low, moderate, and high glucose concentrations in the enteral formulas had no differential effect on sensitivity for any of the four threshold values: F (2, 143) = 1.36, p = 0.26 (20 mg/dL); 1.29, p = 0.28 (40 mg/dL); 0.46, p = 0.64 (60 mg/dL); and 2.69, p = 0.08 (80 mg/dL). Descriptive statistics for the predictor variables are presented in Table 3, along with correlations between these variables and TSG concentrations in the 161 rabbits that received intratracheal instillations of enteral formula plus gastric juice.

Table 2.

Sensitivity of Glucose Method for High, Moderate, and Low Glucose Formulas and Specificity (Using Control Saline Group) by Glucose Threshold Value (20, 40, 60, 80) and Time of Suctioning (2, 4, 6 Hours)

Time of Suctioning	Tracheal Secretion Glucose Concentration Threshold Values
	≥ 20 mg/dL	≥ 40 mg/dL	≥ 60 mg/dL	≥ 80 mg/dL
Sensitivity - High Glucose Formula
2 hours	100% (54/54)	98% (53/54)	96% (52/54)	85% (46/54)
4 hours	98% (57/58)	93% (54/58)	90% (52/58)	74% (43/58)
6 hours	90% (53/59)	78% (46/59)	71% (42/59)	53% (31/59)
Sensitivity - Moderate Glucose Formula
2 hours	100% (40/40)	98% (39/40)	98% (39/40)	80% (32/40)
4 hours	93% (38/41)	88% (36/41)	83% (34/41)	68% (28/41)
6 hours	85% (34/40)	70% (28/40)	60% (24/40)	45% (18/40)
Sensitivity - Low Glucose Formula
2 hours	98% (56/57)	95% (54/57)	84% (48/57)	68% (39/57)
4 hours	98% (58/59)	97% (57/59)	90% (53/59)	58% (34/59)
6 hours	97% (57/59)	90% (53/59)	68% (40/59)	44% (26/59)
Specificity - Normal Saline Control
2 hours	40% (6/15)	60% (9/15)	73% (11/15)	100% (15/15)
4 hours	57% (8/14)	79% (11/14)	79% (11/14)	100% (14/14)
6 hours	58% (11/19)	68% (13/19)	90% (17/19)	95% (18/19)

Table 3.

Descriptive Statistics and Correlations Between Predictor Variables and Tracheal Secretion Glucose In Rabbits with Tracheally Instilled Enteral Formula/Gastric Juice (N = 161)

		Correlations with Tracheal Secretion Glucose (TSG)
Predictor Variables: Mean ± S.D (mg/dL)		2 Hours	4 Hours	6 Hours
Formula glucose	224.8 ± 158.1	0.23 *	0.15	0.05
Gastric glucose	64.2 ± 132.5	0.39 *	0.26 *	0.10
Baseline blood glucose (BG)	111.6 ± 26.2	0.19	0.15	0.09
1-hour BG	112.6 ± 28.7	0.38 *	0.36 *	0.24 *
2-hour BG	118.8 ± 39.6	0.43 *	0.47 *	0.32 *
3-hour BG	110.3 ± 39.7	—	0.40 *	0.38 *
4-hour BG	111.2 ±43.2	—	0.47 *	0.48 *
5-hour BG	114.4± 45.8	—	—	0.54 *
6-hour BG	117.9 ± 48.8	—	—	0.46 *
Visible blood in secretions (1=present, 0=absent)
2-hour	0.28 ± 0.45	0.01	—	—
4-hour	0.35 ± 0.48	—	0.09	—
6-hour	0.50 ± 0.50	—	—	0.17
Volume of tracheal secretions (ml)
2-hour	0.24 ± 0.18	0.03	—	—
4-hour	0.35 ± 0.26	—	0.13	—
6-hour	0.46 ± 0.32	—	—	0.01

The sample sizes used above were n=143 (2-hour), n=152 (4-hour), and n=153 (6-hour).

^*p < 0.01

Because of limited effect of formula glucose concentration on sensitivity, and because there was considerable variation among experimental animals on TSG, researchers examined predictors of 2, 4, and 6 hour TSG using multiple regression analysis. As shown in Table 4, concentration of glucose in the formulas explained 5% of the variation in TSG at 2 hours, but it was not a significant predictor of tracheal glucose (the proxy for aspiration) at 4 and 6 hours. Gastric glucose was a stronger predictor than formula glucose, explaining three times as much variation at all three times. However, gastric glucose ceased to be a significant predictor of tracheal secretion glucose at 6 hours. Blood glucose was the most consistent predictor of TSG, with its overall contribution increasing from 17.1% at 2 hours, to 25.2% at 4 hours, and to 32.2% at 6 hours. Although the presence of visible blood in the secretions and volume of secretions increased from 2 to 6 hours, neither visible blood nor volume made any consistent contribution in explaining TSG. At all three time periods, 60% or more of the variation in TSG remained unexplained (see Table 4). Figure 1 depicts the percentage of variability in the glucose concentration in the tracheal secretions explained by the glucose concentrations in the infused formula and gastric juice and by the concentration of glucose in the animals’ arterial blood.

Table 4.

Multiple Regression Results for Rabbits with Tracheally Instilled Enteral Formula/Gastric Juice (N=161)

	% of Variance in Tracheal Secretion Glucose Explained by Predictors
Predictor Variables	2 Hours	4 Hours	6 Hours
Formula glucose	5.2 *	2.3	0.3
Gastric glucose	15.3 *	6.4 *	1.0
Baseline blood glucose (BG)	4.0 *	2.4	0.9
1-hour BG	9.2 *	10.0 *	4.8 *
2-hour BG	3.9 *	8.7 *	5.1 *
3-hour BG	—	0.1	4.5*
4-hour BG	—	4.0 *	8.7 *
5-hour BG	—	—	7.9 *
6-hour BG	—	—	0.3
Visible blood in secretions (1=present, =absent)
2-hour	0.0	—	—
4-hour	—	0.0	—
6-hour	—	—	0.1
Volume of secretions (ml)
2-hour	1.7	—	—
4-hour	—	0.6	—
6-hour	—	—	3.1 *
Total % variance explained	39.3	34.5	36.6
Adjusted % variance explained	36.1	30.3	31.6

The sample sizes used above were n=143 (2 hours), n=152 (4 hours), and n=153 (6 hours).

^*p < 0.01

Figure 1.

Percent of Variability in Tracheal Glucose Concentration Explained by Glucose in Formula, Gastric Juice, and Arterial Blood

Discussion

As shown in Table 2, simultaneously occurring sensitivity and specificity values ≥ 90% were not observed at any of the four TSG threshold concentrations, regardless of the type of formula used or the hour of data collection. The most acceptable sensitivity/specificity combinations were observed after the first (2-hour) aspiration event at TSG threshold concentrations of 60 and 80 mg/dL. Thus, although a TSG concentration ≥ 20 mg/dL is viewed by some authors as indicative of aspiration (Davis et al., 1995; Potts et al., 1993), current findings do not support this assumption because almost half of the control animals’ secretions had glucose values this high.

Several investigators have advocated the addition of 10 grams of glucose to 500 ml of enteral formula to increase the formula’s glucose concentration to approximately 120 mmol/l (Hussain et al., 2003; Young, 2001). However, current researchers found that high glucose formulas did not exhibit significantly higher sensitivity than the moderate and low glucose formulas (see Table 2). Noteworthy is that the sensitivity of all three enteral formula groups (high, moderate, and low glucose) decreased to about 50% at the time of the third aspiration event (6 hours). Although it is unclear why this decreased sensitivity occurred, it is possible that shifts of glucose into the cells due to increased lung injury were responsible.

An unexpected finding was that the glucose concentration in human gastric juice was a stronger predictor of TSG than was the glucose concentration in the enteral formulas. Based on an animal study by Davenport (1966), current researchers hypothesized that human gastric juice would contain negligible amounts of glucose. However, as reported earlier, they found that 72% of the gastric juice specimens contained glucose concentrations ranging from 5 mg/dL to 800 mg/dL (mean = 99.1±16.6 mg/dL). Although the source of the gastric juice glucose is unknown, it may have been from glucose-based medications or feedings previously administered through the patients’ nasogastric tubes. Although the gastric juice was collected at least 1 hour after medications given by mouth or tube and at least 4 hours after any previous feeding, slowed gastric emptying may have allowed glucose from these sources to remain in the stomach.

As indicated in Table 4, the strongest predictor of TSG at the second and third aspiration events was the arterial blood glucose concentration. This supports findings reported by other investigators of acutely ill humans (Kinsey et al., 1994; Meert, Kshama, Daphtary, & Metheny, 2002). The relationship between arterial and tracheal glucose concentrations is difficult to explain. Kinsey et al. (1994) postulated that a transudative process might allow blood glucose to enter tracheal secretions in pathologic situations.

It is possible that visible blood in tracheal secretions can produce positive glucose readings because blood itself contains glucose (Meert et al., 2002). However, current researchers found no significant difference in glucose concentrations in bloody and non-bloody secretions (see Table 4).

Conclusion

The variables examined in this study (enteral formula glucose, gastric juice glucose, arterial blood glucose, blood in secretions, and volume of the tracheal secretions) explained about one-third of the variance in TSG following the three aspiration events (see Table 4). Despite the hypothesis that the glucose concentration in the aspirated enteral formula would be the primary predictor of TSG, the strongest predictor by far was the animals’ arterial blood glucose concentrations. An unexpected finding was that gastric juice glucose explained considerably more variability in TSG than did enteral formula glucose. Findings do not support the common assumptions that aspiration of high-glucose formula is easier to detect than is aspiration of low-glucose formula, or that bloody tracheal secretions contain significantly more glucose than non-bloody secretions.

In conclusion, findings indicate that the glucose method lacks sufficient sensitivity and specificity to warrant its use in detecting repeated aspirations. As noted earlier, sensitivity decreased sharply from the first to the third aspiration event. An even greater problem is the method’s poor specificity, because factors other than aspiration of glucose-containing formula strongly influence TSG concentration.

Recommendations For Practice

Because there is no bedside test capable of detecting small-volume aspirations, added emphasis should be placed on precautions to prevent aspiration whenever possible. A semirecumbent position is recommended for all at-risk patients (Collard, Saint, & Matthay, 2003; McClave & Dryden, 2003). In addition, a recent consensus group (McClave et al., 2002) on aspiration in critically ill patients recommended that:

• reverse Trendelenburg position be used when a semirecumbent position is not possible.

• small-bowel feedings (as opposed to gastric feedings) be used to reduce the incidence of gastroesophageal reflux and possibly the incidence of aspiration.

• continuous feedings be used instead of bolus feedings in high-risk patients.

• use of narcotics and sedatives be kept to a minimum.

• feedings be abruptly stopped when overt regurgitation or aspiration occurs.

• high-risk patients be moved to a monitored unit with adequate staffing.

• gastric residual volume assessment be used together with clinical assessment to minimize the risk for aspiration, with gastric residual volumes > 500 ml indicating the need to withhold feedings and to reassess tolerance, and gastric residual volumes in the range of 200 to 500 ml prompting careful bedside evaluation and initiation of methods to reduce risk; even though residual volumes< 200 ml seem to be well tolerated, there should be ongoing evaluation of risk.

Future research to detect aspiration should focus on more reliable indicators of aspiration during tube feedings. For example, pepsin is a far more sensitive and specific indicator of aspiration of gastric contents than either the dye method or the glucose method (Metheny et al., 2004; Metheny et al., 2002a; Metheny et al., 2002b). A limitation of the current study is that it was performed on animals and thus did not represent typical findings in clinical settings where patients are receiving continuous tube feedings. Obviously, however, the animal model study was needed to evaluate the glucose method in the presence of a gold standard for aspiration. Findings from the currently reported animal model study are quite similar to findings from studies performed on humans; that is, the presence of glucose in suctioned tracheal secretions is far more likely to be related to blood glucose concentrations than to the presence of aspirated enteral feedings (Kinsey et al., 1994; Phillips et al., 2003).