Protocols and assessment procedures in fiberoptic endoscopic evaluation of swallowing: an updated systematic review

Highlights • Assessments of neurological populations do not demonstrate standardization in swallowing videoendoscopy.• Need for standardization of VED protocols for patient diagnosis and management.• None of the studies used the same protocol.• The quality of the studies varied widely mainly in their methodologies and protocols.


Introduction
The fiberoptic endoscopic evaluation of swallowing (FEES) is one of the tests used to assess the function of swallowing and is currently considered a test already established to identify dysphagia in both children and adults. Described in 1988 by Langmore et al. 1 in a scientific article, in 2001 FEES had its procedures detailed. 2 A historical study on FEES has recently been published, 3 describing the evolution of the procedures used to perform the test and its management in dysphagia, as well as some validations in specific populations.
Some published studies on FEES have focused on the validation of protocols for specific populations, such as extubated patients, 4 head and neck cancer patients, 5 tracheostomized patients, 6 vocal fold paralysis patients, 7 osteopathy, 8 and myasthenia gravis. 9 Other studies validated FEES protocols for specific neurological populations, however, not all studies described the protocol used and validated to be used in clinical practice or even to have its study replicated, in order to confirm the results in similar populations.
The focus on the use of FEES in populations of neurological etiology has some particularities, since the examination in the adult neurogenic population may be difficult due to factors such as: the cognitive aspect, 10,11 trunk and head posture during the exam and the occurrence of fatigue, 12---14 which requires the test interruption and makes the diagnosis and treatment plan difficult.
In general, the FEES assessment protocol is divided into three stages: the first is through careful observation of the anatomy, secretions, and visualization of the movements of the nasal structures when a patient is asked to speak and breathe. Some protocols include assessment of sensitivity in the oropharyngeal region by touching the endoscope in specific regions. 15,16 The second stage consists of the direct evaluation of swallowing, offering food and liquids in different consistencies. The third stage consists of verifying postural maneuvers, variations in consistencies and observing eating behaviors, directly identifying postures and food consistencies that favor oral intake in a safer way.
Instrumental assessments to evaluate swallowing function have a gold standard in the FEES and in the videofluoroscopic swallowing study (VFSS). 17 VFSS is a test in which the swallowing process is observed from the time the bolus is captured, passing through all the swallowing phases (oral preparatory phase, oral phase, pharyngeal phase, and esophageal phase). 18 Therefore, this is an exam that determines the degree (mild, moderate, severe) of swallowing changes. In contrast, the FEES observes the pharyngeal phase of swallowing, however it has the advantage of identifying the exact location of the waste in that phase of swallowing, its quantity, and identifying which best maneuvers performs the partial or total cleaning of this residue. In addition, FEES, due to the fact that it does not use radiation, demonstrates greater ease of reproducibility and replicability, both in inpatients and outpatients. Focusing on the adult neurological population, in its evaluation and therapeutic follow-up, FEES is often indicated. 19 Therefore, the use of a specific protocol by speech therapists and physicians during FEES performance facilitates the examination and the clinical diagnosis, offering evidencebased recommendations and reducing the rate of variation. The aim of this study is to identify and describe, based on a systematic review of the scientific literature, the FEES protocol to be used in the adult neurological population with details and the possibility of worldwide standardization.

Methods
A systematic review of the literature was carried out guided by the question: ''Is there a protocol for performing FEES and, if so, is it validated to be used universally in patients with neurogenic diseases?'' This study was registered on PROSPERO (CRD42018069428). The databases searched were PubMed/Medline, Cochrane Library, Web of Science and SciELO. The main descriptors related to the investigated theme, crossed, were: FEES evaluation; adults; neurogenic disease; swallowing assessment, as shown in the strategies presented in Table 1. The outcome of each study was considered, that is: presence of a diagnosis of swallowing disorder; impaired laryngeal sensation (at the level of vocal folds); presence of laryngeal penetration of bolus; and/or occurrence of tracheal aspiration.
The review included cross-sectional studies, randomized clinical trials, and longitudinal cohort studies, which used FEES as a standard assessment instrument with neurogenic disease patients. Other types of studies or formats were excluded, as well as cross-sectional studies that included children and/or adolescents. The selection of articles covered the period between 2009 and 2019. The survey of bibliographic data took place between March 2018 and March 2020, based on the aforementioned inclusion criteria ( Table 1).
The first phase of article selection was the exclusion of duplicate studies, followed by the reading and analysis of titles and abstracts of all identified works. Afterwards there was a complete reading of the selected studies, which led to the exclusion of works that did not meet the review criteria. The selected articles were submitted to methodological evaluation, according to the checklist provided by the report Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) 20 for cross-sectional studies, which received a score of 1 when the item was considered, of   33 The arithmetic mean among the assessed works was 17.03 points, with a minimum score of 15.25 and a maximum of 20.25. The minimum score for evaluating the studies was 69% and the average applicability of the studies was 78%. 0 when not contemplated, and of 0.5 when partially contemplated. Afterwards, the averages between the scores assigned by the two evaluators were established. All phases were carried out by two of the authors/researchers, independently. Faced with doubts about whether or not to include the study, the third author/evaluator was called. This study included only articles with at least 70% of the score determined by the STROBE checklist. The arithmetic mean of study scores was 17.86, making up a proportion of 81% of the STROBE score. The included articles were analyzed regarding the possibility of bias, study limitations, number of participants, gender, age, and statistical method ( Table 2). All review procedures presented here were conducted in accordance with Checklist Preferred Reporting Items for Systematic Reviews and Meta-Analyzes (Prisma).

Results
The selection of studies carried out by 2 independent evaluators in three databases, as described in the methodology, found a total of 3724 published studies based on the crossed descriptors ''fiberoptic endoscopic evaluation of swallowing AND protocols of assessments'' and ''endoscopic swallowing assessment AND assessment procedures'' . Based on the personalized search and the inclusion of the descriptor ''neurology'', aiming at studies performed with adult neurological population, 77 articles were selected on the PubMed platform, 14 in the Cochrane Library, 10 in Scielo, making a total of 101 articles. In search of article eligibility and from reading abstracts, articles that did not meet the inclusion criteria of the study were excluded, 28 articles remaining for full reading, which were analyzed by two blinded evaluators.
Finally, 21 studies published between 2009 and 2020, considered as updated publications, were selected because they were the only ones with complete FEES protocols described in the publications. The description of the search for the articles is shown in Fig. 1 based on PRISMA guidelines (Fig. 1).
Of these, 18 studies are cross-sectional, 02 crosssectional longitudinal and 01 application of a protocol developed through a cohort. The neurological diseases assessed in the selected studies were: progressive supranuclear palsy (PSP), amyotrophic lateral sclerosis (ALS), elderly patients hospitalized with some health condition, Parkinson's disease (PD), stroke, myotonic dystrophy type 1 (DM1), myasthenia gravis (MG), elderly >60 years, traumatic brain injury (TBI), and vascular dementia. The description of the articles and the methodological characteristics of the articles are shown in Table 3.
The methodological evaluation of the studies using the STROBE report, by individual evaluation of two blinded and independent evaluators, and the hypothesis for this review found 21 studies which were selected for this systematic review, 5 of them with high score, 21---25 and one highlighted due to its results and statistical analysis performed. 25 When verifying the risk of bias within studies, some studies have exposed their limitations and were found in Warnecke et al. 26 that to avoid the expectation bias, the FEES performed in the study were evaluated randomly by two independent judges; that is, out of the order in which the exams were performed. Alvarenga et al. 23 reported as one of the limitations of the study a probable sample bias, since the patients who accepted it probably did so due to presenting swallowing symptoms. Braun et al. 24 reported a possible selection bias when researching patients in the intensive care unit, demonstrating that the study sample was more severely affected. Suntrup-Krueger et al. 25 pointed out limitations on possible biased results of the study due to the fact that in the intensive care unit there is a high knowledge of post-extubation dysphagia. Nienstedt et al. 27 and Pflug et al. 28 stated that they minimized selection bias and Warnecke et al. 29 commented on a possible selection bias based on the inclusion and exclusion criteria of the study. Shapira-Galitz et al. 30 suggested a possibility of assessment bias at the time when the researchers needed to read the questionnaire for patients over the phone. D'Ottaviano et al., 31 Leder et al., 22 Mandysova et al., 21 Marian et al., 32 Gozzer et al., 33 Imaizumi et al., 34 Schröder et al., 35 Pilz et al., 36 Somasundaram et al., 37 Souza et al., 38 Souza et al., 39 Umay et al., 40 and Farneti et al. 41 did not describe the bias assessment.
The number of patients evaluated in the selected studies ranged from 1 to 961, all of them being adults and or elderly, both men and women, and 6 of the studies carried out an evaluation in a control group compared to age matching. 25,27,28,30,36,40 Among the protocols presented in the 21 studies, 7 of the articles used protocols from the institutions where the research was carried out or even protocols only detailed in the articles; 2 used the protocol by Dzeiwas et al. 42 ; 2 studies used the Langmore protocol 1 ; 2 used the Langmore protocol 43 ; 4 used the Langmore protocol 2 ; 1 article performed FEES evaluations with protocol by Warnecke et al. 9 ; 1 study evaluated using the brief bedside dysphagia screening test 21 ; 1 used the FEES levedopa-test 26 ; and 1 performed evaluations without the use of food using the protocol by Hyodo et al. 44 However, even though studies have mentioned the use of the same protocol to perform the FEES, in the description of the protocols, we observed differences in the chosen consistencies, foods and volumes, not characterizing the same protocol. The way in which FEES images were captured was not discussed in the articles, which is why it was not discussed in this systematic review.
Detailing the protocols, only 3 of the studies cited the international dysphagia diet standardization initiative (IDDSI) 45 as a basis for standardizing the food consistencies offered during the FEES of the studies, with thick and liquid consistencies being offered in these studies. One of the studies included the evaluation of liquid in addition to the other consistencies mentioned. Three of the articles presented better and more accurate details of the protocols used to perform FEES, Warnecke et al., 26 Shapira-Galitz et al. 30 and Souza et al. 38 , specifying the tested consistencies, declaring the foods that were used for each consistency, the quantity offered at each moment, and how many times each consistency was offered. Of the 21 studies, 11 mentioned having used food coloring to contrast the color of the food in relation to the structures, blue or green and possible secretions present during the exam.
Among the outcomes assessed by de Lima Alvarenga et al., 23 Gozzer et al., 33 Souza et al. 38 , and Souza et al., 39 there were: posterior oral leakage, pharyngeal residues, laryngeal penetration, laryngotracheal aspiration,                     and laryngeal sensitivity. The study by Souza et al. 39 evaluated mainly residues, all of them according to each tested consistency. The studies by Warnecke et al., 9 Warnecke et al., 26 Marian et al., 32 and Braun et al. 24 used the severity scale for dysphagia in their outcomes, the fiberoptic endoscopic dysphagia severity score (FEDSS). 9 As for the autonomy and capacity for oral intake based on the Functional Oral Intake Scale (FOIS), they were assessed as outcomes by Leder et al., 22 Farneti et al., 41 Imaizumi et al., 34 and Shapira-Galitz et al. 30 In the study by Imaizumi et al., 34 the International Classification of Functionality was also used for the assessment. The detailed outcomes in each study separately are shown in Table 4.

Discussion
The objective of this systematic review of identifying a standardized and validated protocol for endoscopic evaluation of swallowing in patients with underlying neurogenic disease has not been achieved. All studies evaluated and selected for this study used described and detailed protocols, but none were validated. Mandysova et al. 21 developed a dysphagia screening test for bedside application and validated it based on FEES, but it is not a validated FEES protocol. Dziewas et al. 42 have been mentioned in some of the articles as a validated FEES protocol, however it is the development and validation of a new score for the assessment of dysphagia severity, which does not correspond to a validated FEES protocol for the neurological population. The Hyodo score 44 was developed and validated to identify the presence and degree of dysphagia, indirectly, that is, without the use of liquids and food, only based on secretion management and intraoral sensitivity in neurological patients, consequently not characterizing a FEES protocol, but an evaluation score.
As for all studies that cited the Langmore protocol, in different years of publication (1988,1998,2001) and updates, these are considered guidelines, as the author says in a recent 2017 article, 3 requiring validation for different populations. Regarding the protocols presented by the selected studies, 7 out of 21 studies used the consistency of pureed, liquid, and soft & bite sized, 6 of the studies used pureed, liquid and solid. The other studies used different consistencies, as an example of liquid and slightly thick only. Considering the existence of consistency standardization based on IDDSI, an international diet standardization initiative 45 since 2015 and updated in 2019, only Souza et al., 38 Souza et al., 39 and Gozzer et al. 33 used international validation. IDDSI favors the use of the same nomenclature worldwide, facilitating the standardization and validation of assessment protocols such as FEES for neurological populations, for example. However, 13 of the 21 studies were published after 2015, the year in which IDDSI was created and published, and did not use international standardization, making it difficult to validate a protocol.
The volumes presented in the study protocols were widely different, varying from studies testing all consistencies in 3, 5 and 10 mL, others testing only 10 mL three times and others identifying volumes as the size of the spoon offered, with the teaspoon being specified, to studies that did not specify volumes or number of offers or even the utensils used.
Some of the utensils used in some studies were straw, glass, but without specifying the size or even diameter of each utensil, making it difficult to understand the quantity offered to the study participant. Therefore, the use of spoons, straws or cups without specifying size and quantity cannot be characterized as a description of volumes, as there are different diameters and sizes for each of these utensils. Additionally, studies citing the same protocol in the description of the article used different consistencies and volumes, consequently not being the same protocol.
Finally, the presence of a speech therapist to provide food and guide the swallowing of patients during FEES alongside the doctor who performs endoscopy was mentioned in 11 of the 21 selected studies. It is clear that the evaluation of swallowing and its possible changes during FEES is of exclusive medical responsibility in Brazil, while the role of the speech therapist is to monitor the evaluation and verify the patient's responses to body maneuvers for food. Teamwork is the gold standard for an accurate diagnosis and determination of the appropriate therapeutic plan for that patient.
Regarding the quality of the studies, even considering a high cut-off point, most studies were evaluated as good or satisfactory, making it difficult to consider any of these protocols for standardization and/or validation for FEES in the neurological population. However, three of these studies were the ones that best detailed their procedures and protocol for FEES evaluation. 26,30,38 Three studies brought all the items meticulously detailed since the tested consistencies, one of them using the IDDSI, the volumes offered in units of ''mL'' or ''cc'', mentioned the use of food coloring to favor contrast in relation to the structures and secretions of the organism, food and liquids offered according to each consistency, how many times each one was offered, and which utensils were used to offer each consistency. The way in which FEES images were captured was not discussed in the articles, which is why they are not highlighted in this systematic review.
However, none of these protocols have been validated for the neurological population. The lack of a validated protocol makes it difficult to widely standardize the assessment and changes the therapeutic approach, since the entire rehabilitation is based on a detailed, accurate and reliable assessment. Adequate and correct diagnosis is the basis for any rehabilitation and management of dysphagia in adult patients with neurological disease.
The importance of having a standardized and validated protocol for neurological populations, including specific populations, is crucial because each neurological disease has its own particularities and pathophysiology, as well as the presentation of dysphagia. Dysphagia characteristics vary widely according to the neurological diagnosis. The literature details these differences regarding dysphagia in Parkinson's disease, 46 dysphagia in supranuclear progressive paralysis (PSP), 47 dysphagia in amyotrophic lateral sclerosis (ALS), 48 as well as dysphagia in traumatic brain injury (TBI), 49 among others.
Some of the limitations faced in this systematic review were the possible bias in evaluating the quality of the studies, due to the fact that one of the blinded evaluators was the same to account for the averages of the items and the final score and the non-use of the third evaluator in this phase of the study. In addition to the possible bias in the selection of studies, as we have delimited more recent research (from the last eleven years), we may have left out historically important research for this topic, although mentioned in the introduction and discussion.
The lack of a standardized and validated protocol for the adult population with neurogenic diseases significantly limits a detailed, accurate, and focused assessment of the possible swallowing difficulties faced by these patients. The clinical diagnosis of dysphagia may be underestimated or overestimated according to the protocol used and outcomes assessed. This systematic review, the first in the field, highlights the need to validate protocols with a focus on adults with underlying neurogenic diseases considering the characteristics of dysphagia and its pathophysiology. Adequate, reliable and accurate diagnosis is the basis for the management of swallowing in these populations.

Conclusion
The reliable reproducibility of the protocols is only feasible in three of the articles, even with different protocols, but none were standardized or validated for the adult neurological population.

Conflicts of interest
The authors declare no conflicts of interest.