Retrospective Analysis of the Clinical Utility of a Protocol for Baseline Clinical Swallow Examination

Introduction

Dysphagia is defined as an alteration in the bolus transit from the mouth to the stomach [1]. Swallowing process may be impaired by neurological disorders and damages, oncological diseases, and the aging process [2]. Prevalence of dysphagia in the general population is about 3-5% [3], while it reaches 16% in elderly [4], 40-60% in patients with stroke [5], and 50-90% in patients with Parkinson’s disease [6]. Complications of dysphagia are aspiration pneumonia, malnutrition, dehydration, reduced QOL and limitations to social participation [7, 8, 9]. Therefore, the identification and the accurate assessment of patients at risk for dysphagia are of critical importance [10, 11].

The typical clinical pathway for swallowing evaluation consists, in order, of a screening, a clinical swallow examination, a mealtime observation and an instrumental assessment. Screening is usually performed by non-specialized professionals and is applied to the whole population of patients at risks of dysphagia [10]. The goal of a swallowing screening is to identify patients requiring an extended assessment of swallowing and promptly decide whether or not to orally feed the patient. Clinical Swallow Examination (CSE) is performed by professionals specialized on swallowing, typically speech and language therapists (SLTs), on patients who are suspected to have dysphagia. The CSE aims to recognize patients with dysphagia, to identify patients requiring an instrumental assessment, to decide the possibility and the type of oral nutrition, and to guide the possible swallowing treatment. Usually, the CSE is appropriately followed by a mealtime assessment to verify an adequate oral intake, the patient’s independence and the dietary prescription. Finally, instrumental assessment provides the final diagnosis of dysphagia, allows to identify pathophysiological mechanisms of swallowing disorders, and adds information to the CSE for decision on oral nutrition and swallowing treatment [12, 13, 14].

In our Intensive Rehabilitation Facility (IRF), considering the high prevalence of dysphagia in neurorehabilitation settings [15], all patients are directly assessed at admission by a SLT using the Bilancio Logopedico Breve (BLB)- swallowing subtest, an Italian baseline CSE protocol [16]. The advantages on the use of the BLB over other available CSE protocols are its shortness and rapidity, making it suitable for the use on a large number of patients. However, although its use has become widespread in different Italian hospitals and rehabilitation units, and its items provide a useful checklist for guiding a meticulous CSE, the BLB is a non-validated protocol. Beside this limitation, the clinical experience on the application of the CSE allowed to qualitatively identify other drawbacks of the protocol, e.g. the absence on an item on indirect signs of aspiration with solids or the different score range for the items.

The aim of the study is to retrospectively analyze the clinical utility of the BLB-swallowing subtest, in particular to identify which items are more relevant for identifying a swallowing disorder and for guiding an adequate diet prescription. This finding might lead to a future revision of the protocol, preliminary to the validation process. The clinical utility is investigated by studying: (i) the association between the scores of the BLB items and the clinician’s decision on the diet; (ii) the association between the scores of the BLB items and the signs of penetration and aspiration on instrumental assessment; (iii) the responsiveness of the BLB items to the diet evolution of the patients within the hospital stay.

Material and Methods

This retrospective study was carried out in Casa di Cura del Policlinico (Milan), a post-acute IRF. The study was conducted in accordance with the Declaration of Helsinki and was approved by the Local Ethics Committee (Milano Area B: Resolution 20_2017). At the time of admission, all patients signed an informed consent form on the use of demographic and clinical data for both clinical and research purposes. Clinical data of the patients and information on medical and rehabilitation procedures carried out during the hospitalization period were recorded on an electronic database management system, routinely used in the clinical practice.

Discussion

The study provides a first critical analysis of the BLB swallowing subtest, an Italian baseline CSE protocol, based on its use in the clinical practice over a 5-years period in a large sample of patients with neurological diseases. It is essential to specify that the study did not aim to validate the BLB swallowing subtest, as the clinical experience accrued since 2010 allowed to qualitatively identify both advantages and limitation in the use of this CSE protocol. Conversely, the results enabled to pinpoint a different clinical utility of the BLB items and may be useful for a revision of the protocol in the future.

Among the 12 items, 4 items showed poor findings according to the analyses (Table 5): pathological reflexes, lip sealing during swallowing, laryngeal excursion, oral residue. The item pathological reflexes showed the greatest ceiling effect (>90%). This result may be a consequence of two aspects. Firstly, the item was dichotomous (0-1) as pathological oral reflexes were only judged as present or absent. Secondly, this symptom reflects a diffuse hemispheric dysfunction and, therefore, a severe neurological impairment, which is restricted to a limited number of patients in the IRF. Concerning laryngeal excursion, the poor association values may be related to the difficulties in assessing this aspect during CSE. Indeed, laryngeal elevation is a critical mechanism for swallowing safety (i.e. avoiding penetration and aspiration) [17]. However, a poor inter-rater agreement (ICC 0.446) was previously reported for its assessment during CSE [18]. Oral residue and lip sealing are relevant aspects to assess as they may reflect difficulties in oral preparation and oral transit, especially with solids. However, they were found to be significant only in respectively two and one analyses performed in the present study. The items showed a clear ceiling effect (>50%), thus reducing the variability in the scores. It may be argued that, although recognizing their important in oral processing of food, the possibility to assess them during CSE, with a small number of swallowing trial, in a standard situation, and with small volumes, may be limited. Conversely, their assessment during a meal observation may provide more valuable information to guide the refinement of diet recommendations.

The items swallow trigger, oral motor skills, voluntary cough, gag reflex, dentition, and signs of aspiration during liquid showed the best findings in the analysis of the present study (Table 5) and, therefore, seem to be useful for guiding the clinician in the decision of nutrition method and diet type, for identifying the patients with signs of penetration and aspiration on FEES, and for detecting changes of swallowing function over time. The item signs of aspiration during puree was significantly associated with the decision of oral nutrition, the presence of penetration or aspiration on FEES and was sensitive to diet evolution. As expected, no statistically significant association was found with the decision of the diet type (pureed vs. solid diet).

The association between the BLB and the diet recommendation was investigated in two different analyses: (i) comparing patients with NPO vs. oral diet recommendation and (ii) comparing patients with pureed vs. solid oral diet. The difference in the results of the two analyses arises interesting consideration on the items to include in a bedside swallowing protocol based on the targeted aim of the protocol. For instance, assessing oral preparation is relevant for the decision of the type of oral diet, but may be omitted if the tool only aims at supporting the decision of an oral or non-oral nutrition.

Prevention of pulmonary complication (i.e., aspiration pneumonia) is one of the primary goals of swallowing assessment [10]. Instrumental assessment, either with FEES or videofluoroscopy, represents the gold standard for swallowing evaluation [19]. Previous studies documented that CSE is still not strongly predictive of aspiration on instrumental assessment [20]. We assessed the association between BLB items and FEES findings of penetration or aspiration and found a significant association with the items gag reflex, swallow trigger, signs of aspiration with liquid, and with puree trials. Although all the items missed to identify a certain percentage of patients with penetration or aspiration, the value of the CSE in the identification of patients at risk of penetration and aspiration and its advantages over FEES or video fluoroscopy, being low-tech and promptly available in all settings, must be acknowledged. The association between the alteration of the gag reflex and deglutition difficulties is known and it has been previously documented [21]. The item swallow trigger is itself more problematic since a delayed initiation of the pharyngeal phase can be identified only though an instrumental assessment of swallowing [22, 23]. However, the clinical detection of any disturbance or latency in the timing of the swallowing act could reasonably and efficiently highlight a deficit in the transition between the voluntary and the reflex phases of deglutition. Such explanation could support the present finding.

The study has some limitations, mainly related to the retrospective design. Firstly, the data extraction showed some missing values. Patients with more than 2 missing values in the swallowing assessments were not included and the rate of missing value in the BLB items were <5%. The only exception is the item on oral preparatory phase that reached 10% of missing values. This item may be not assessable in some patients, but the retrospective design did not allow to discriminate between missing values and not assessable items. Secondly, the FEES protocol was not standardized, thus variability in the number of trials and volumes may have occurred among patients. Moreover, due to the retrospective nature of the study it was not possible to retrieve more reliable data from FEES reports other than the presence of penetration or aspiration. CSE and FEES interpretation may be prone to clinicians’ subjectivity, which could not be controlled through reliability analysis. Multivariate analysis was not performed due to the low frequency of some cells. However, independent association among BLB items and other factors may not be of primary importance when aiming to identify markers of swallowing impairment and not predictive factors. Lastly, BLB present several limitations: different scoring for each item, high risk of subjectivity in the administration and the interpretation, no standardized training before its administration, no validation study, and no trial with the solid consistency.

In conclusion, the BLB swallowing subtest was critically analyzed based on 5-years clinical experience in an inpatient rehabilitation department. Eight out of 12 items showed a satisfactory performance being associated to diet recommendations and signs of dysphagia on FEES and recording an evolution of swallowing function over time. The present findings support the use of BLB in the clinical practice as a guideline for CSE and highlight the need of a revision of the protocol and a subsequent validation procedure.

Conflict of Interest Statement

• Financial Disclosure: no funding, or financial relationship to disclose. Non-financial disclosure and conflict of interest: Marco Gilardone is one of the authors of the Bilancio Logopedico Breve; the other authors have nothing to disclose.
• Authors’ Contributions N.P., M.G., and M.C. designed the study. M.L., M.V. and G.G. performed patients’ enrolment and data collection. F.M. and N.V. provided the data analysis. N.P., I.D., M.C. and M.G. interpreted the results. N.P. and M.V. took the lead in writing the manuscript. All authors provided critical feedback and helped shape the paper. All authors read and approved the final version of the paper.