The rationale for dyeing food blue to identify aspiration appears to be based on tradition, intuition, acceptance, and assumptions rather than on objective analyses of available evidence [1]. Since blue is not a color found in oropharyngeal secretions, blue dye has been used to detect aspiration without feeding trials [2], as well as added to enteral feedings of critically ill patients [3,4] and to oral feedings of patients with suspected dysphagia [5,6], for over 25 years. New concerns, however, have developed over both the safety [1,7–9] and sensitivity [10–12] of blue dye use.

Methylene blue, although used to dye food, was not approved by the Food and Drug Administration (FDA) for this purpose [13] and side effects and expense resulted in its discontinuance [4,14,15]. Currently, FD&C Blue No. 1 is the color used most often in both enteral feedings and dysphagia evaluations because it is readily available, inexpensive, and simple to use [10], practically nonabsorbable from the healthy gut [16,17], and purportedly sensitive to detection of aspiration [18]. Multiuse nonsterile bottles were first used to store and dispense the dye, but concerns regarding bacterial colonization [19] led to the manufacture of individual use sterile vials (SteriBlu 5-ml vials of 2% Blue No. 1; Nestle Clinical Nutrition, Deerfield, IL). It appeared there were few downsides to blue dye use [18].

Objective analyses, however, found that the blue dye method had poor sensitivity, i.e., a high false positive rate, for detecting aspiration in both enterally fed patients [3,20,21] and during dysphagia evaluations [10–12]. Although the FDA recommended 12 mg/kg/day as the limit for oral FD&C Blue No. 1 intake [22], no consensus guidelines exist regarding volume of dye to add to feedings resulting in large differences across clinicians [4], e.g., the amount of dye added is typically “eyeballed” to the desired color [18]. In addition, amounts of FD&C Blue No.1 far below daily FDA recommendations for healthy subjects have caused discoloration in patients with sepsis [18].

Recently, a number of deaths have been associated with blue dye use [7–9]. Patients were seriously ill but their conditions were improving before they received dyed enteral feedings and turned color. It is known that in the presence of sepsis gastrointestinal permeability increases because of enterocyte death and loss of barrier function at intercellular gaps, with the result that substances not otherwise absorbable, e.g., FD&C Blue No. 1, may be absorbed [23]. Since use of blue food coloring is poorly standardized [4,18], has low sensitivity for detecting dysphagia [3,10–12,20,21], and reports of death have raised safety concerns [7], it was recommended that blue food coloring be abandoned as a marker of aspiration [18,24].

In response to health and safety concerns FD&C Blue No. 1 is no longer being manufactured in individual-use sterile vials, thereby eliminating its use in the hospital setting. Fiberoptic endoscopic evaluation of swallowing (FEES) [5,6] is a widely used and reliable technique to diagnose pharyngeal dysphagia and aspiration and to make appropriate recommendations for therapeutic interventions if dysphagia is present [25–27]. The FEES technique has always added a few drops of colored dye to food as a visual aid in detection of pharyngeal dysphagia and aspiration. Although such a small amount of dye poses very little health and safety risk, since FD&C Blue No. 1 is no longer available not even this small amount of dye can be used. The purpose of the present study is to determine if FEES maintains high intra- and interrater reliability in detecting pharyngeal dysphagia and aspiration without FD&C Blue No. 1 added to food.

Methods

Subjects

Table 1 shows descriptive statistics for 20 consecutive adult inpatients referred for swallow evaluations from the acute care setting of a large, urban, tertiary care, teaching hospital. There were 13 males (mean age = 68 years, range = 38–92 years) and 7 females (mean age = 64 years, range = 47–80 years), with no significant age differences between the genders. There were a variety of diagnostic categories consistent with the inpatient population of a large, urban, tertiary care, teaching hospital, i.e., medicine (N = 6), surgery (N = 5), trauma (N = 5), and neurological (N = 4).

Table 1 Patient characteristics
Full size table

Equipment and Materials

FEES equipment consisted of a 3.6-mm-diameter flexible fiberoptic rhinolaryngoscope (Olympus, ENF-P3); light source (Olympus, CLK-4); camera (ELMO, MN401E); VHS recorder (Panasonic AG1330); color monitor (Magnovox, RJ4049WA01); and digital swallowing workstation (Kay Elemetrics Corp., Model 7200). FD & C Blue No. 1 (Nestle SteriBlu), puree consistency (vanilla pudding [yellow]), and liquid consistency (skim milk [white]) foods were used.

Procedures

The basic FEES protocol [5,6] was done at bedside, with the patient in an upright position or as upright as tolerated, and without administration of topical anesthesia to the nasal mucosa, thereby eliminating any potential adverse anesthetic reactions and ensuring a reliable physiologic evaluation [28].

The 20 subjects were randomized to receive food either with or without FD&C Blue No. 1 (Fig. 1). Nine subjects received blue-dyed food and 11 subjects received nondyed food. In the blue-dyed food group, two drops of blue dye were mixed with 60 cc of pudding and 120 cc of milk, respectively. In the nondyed food group, vanilla pudding (yellow) and skim milk (white) were ingested. The order of food presentation was the same for dyed and nondyed foods, i.e., three puree boluses (3–5 cc each) followed by three liquid boluses (3–5 cc each).

Fig. 1
figure 1

Videophotographs of food trials with (A) blue-dyed bolus, (B) non-blue-dyed bolus of vanilla pudding (yellow), (C) non-blue-dyed bolus of skim milk (white), and (D) monochrome food bolus (black and white).

Full size image

The occurrence of four events was rated for each subject: (1) the stage transition characterized by depth of bolus flow to at least the vallecula prior to the pharyngeal swallow; (2) evidence of bolus retention in the vallecula or pyriform sinuses after the pharyngeal swallow; (3) laryngeal penetration defined as material in the laryngeal vestibule but not passing below the level of the true vocal folds either before or after the pharyngeal swallow; and (4) tracheal aspiration defined as material below the level of the true vocal folds either before or after the pharyngeal swallow.

Data Analysis

Three speech–language pathologists experienced in interpreting FEES results independently and blindly reviewed the digitized videotape three times. Using real-time analysis with repeat viewing as needed, inter- and intrarater reliability ratings for depth of bolus flow before the swallow, evidence of bolus retention after the swallow, laryngeal penetration, and tracheal aspiration were made. The first two viewings maintained appropriate color contrast and were viewed four months apart. The third viewing occurred two months after the second viewing using only a monochrome, i.e., black and white, image thereby eliminating any influence of color on ratings (Fig. 1). Intrarater agreements were expressed in percentages and the kappa statistic was used to measure interrater reliability. The kappa statistic is a proportion of agreement between raters after chance agreement has been excluded, i.e., κ = (observed agreement − expected agreement)/(number of paired judgments − expected agreement). The upper limit of the kappa representing total agreement between the raters is 1.00, while a kappa of 0.00 represents agreement at chance level. Kappa scores in the range of 0.41–0.60 indicate moderate agreement, 0.61–0.80 indicate substantial agreement, and 0.81–1.00 indicate excellent agreement [29].

Results

Table 2 shows intrarater agreement for the four variables, i.e., bolus flow, bolus retention, laryngeal penetration, and tracheal aspiration, with blue-dyed, non-blue-dyed, and monochrome recordings. Intrarater agreements for blue-dyed versus non-blue-dyed food trials were 100% for all four variables and monochrome trials ranged from 95% to 100%.

Table 2 Intrarater agreements (%) for the three raters with blue-dyed, non-blue dyed, and monochrome food trials.
Full size table

Table 3 shows combined kappa values for the three raters with blue-dyed, non-blue-dyed, and monochrome food trials. Kappa values indicated that agreement was above chance for all four variables under all three viewing conditions. Specifically, moderate to substantial agreement (0.56–0.71) occurred for blue-dyed food trials, substantial to excellent agreement (0.75–1.00) occurred for non-blue-dyed food trials, and moderate to excellent agreement (0.60–0.87) occurred for monochrome food trials. Aspiration demonstrated the highest values in each of the viewing conditions, with perfect agreement (1.00) observed for non-blue-dyed food trials.

Table 3 Combined interrater kappa values for blue-dyed, non-blue-dyed, and monochrome food trials
Full size table

Table 4 shows combined interrater kappa values for the three raters comparing blue-dyed versus monochrome trials and non-blue dyed versus monochrome trials. All variables in both comparisons were in the substantial-to-excellent-agreement range (0.63–1.00). Specifically, kappa values for blue-dyed versus monochrome trials ranged from substantial to excellent agreement (0.75–0.89) and non-blue-dyed versus monochrome trials ranged from moderate to excellent agreement (0.63–1.00).

Table 4 Combined interrater kappa values comparing blue-dyed versus monochrome trials, and non-blue-dyed versus monochrome trials
Full size table

Discussion

The principal finding of the present study was that the FEES technique [5,6] maintains high intra- and interrater agreements for detecting the major symptoms of pharyngeal stage dysphagia and aspiration with the use of regular, non-blue-dyed food, i.e., yellow pudding and white skim milk. Substantial to excellent overall intra- and interrater agreements were observed among the three conditions, i.e., blue-dyed food, non-blue-dyed food, and monochrome food trials, and for all four variables associated with pharyngeal dysphagia, i.e., bolus flow, bolus retention, laryngeal penetration, and tracheal aspiration. It appears that the important variable in detecting both bolus flow to and location in the pharynx and larynx is how well a bolus reflects light, i.e., it must be brighter than the tissue it is resting on. Therefore, the endoscopist can be assured of reliable FEES results using regular, non-dyed food trials.

Some endoscopists may have considered changing to green food dye following reports of systemic problems with blue dye [1,7–9]. There have been no studies in the literature, however, regarding the safety or sensitivity of using green or any other colored food dyes. The present study’s high intra- and interrater agreements for pharyngeal dysphagia and aspiration using regular, non-dyed foods do not support the use of any food dye during FEES.

An attempt was made to define the minimal volume of bolus necessary for a rater to identify bolus flow, bolus retention, laryngeal penetration, and tracheal aspiration. This proved to be difficult for a small percentage of specific trials. However, overall intrarater agreements, at both the initial and repeated videotape viewings, were extremely high, indicating that consistent internal criteria were maintained. The minor discrepancies in the interrater agreements were due to trace, i.e., clinically insignificant, volumes of food being identified by one rater but not by the others. It is not felt that disagreements on such small bolus volumes would change the diagnosis or influence diet recommendations for any patient. Rather, the discrepancies appear to be an artifact of the narrowly defined research criteria.

There are a number of areas for future research. Food colors other than white milk and yellow pudding should be investigated to determine if high intra- and interrater agreements for pharyngeal dysphagia and aspiration are maintained during FEES. Also, it would be of interest to determine if ratings by novice clinicians are different from those of experienced endoscopists in identifying depth of bolus flow, bolus retention, laryngeal penetration, and aspiration for dyed versus non-dyed food presentations during FEES.

Conclusions

The reliability of FEES to detect the critical features of pharyngeal dysphagia and aspiration is consistently high using either blue-dyed or non-blue-dyed foods. As a result, the endoscopist can reliably identify depth of bolus flow, bolus retention, laryngeal penetration, and tracheal aspiration using regular, non-blue-dyed foods, i.e., yellow pudding and white skim milk. Use of blue dye during FEES, therefore, may be abandoned as a marker of pharyngeal dysphagia and aspiration [18,24].