The study was done in a large German university hospital. As a part of routine care delivery for patients hospitalised for acute stroke, FEES was performed in case of a pathologic bedside screening procedure, performed by nurses or SLTs. In our department, we use the Gugging Swallowing Screen (GUSS) [13]. If the patient passed the GUSS, no FEES was performed and full oral diet was chosen. If the GUSS indicated possible dysphagia, the patient underwent a CSE by an SLT and FEES by a team consisting of a SLT and a neurologist. FEES was also performed if a patient showed signs of pharyngeal dysphagia during hospitalisation (e.g. wet voice, coughing when drinking, etc.) and if a patient developed signs of infection (productive cough, elevated inflammatory markers). The signs of dysphagia were reported by nurses, SLTs or the treating physicians. The patients with signs of dysphagia were discussed among the “dysphagia experts” of our department and indication for FEES was confirmed; oral diet prior to FEES was chosen as instructed by the GUSS or by clinical judgement of the treating physician. For quality control reasons, our findings gathered from examinations were documented systematically. All FEES were performed in a standardised manner by experienced physicians. The screening process is depicted in Figure 1.
Patients
All stroke patients treated in our department from January 2014 to September 2016 in whom FEES was performed were documented in a standardised manner. Approximately 800 patients per year are discharged from our hospital with the diagnosis of a stroke. Data documented in the database included age, sex, length of stay in hospital, stroke entity (ischemic stroke vs. primary haemorrhage), stroke aetiology (TOAST-criteria in ischemic stroke), NIHSS and mRS on admission and at discharge, localisation of ischemic lesion (left/right hemisphere, bilateral infarctions, brain stem), ischemic vascular territory, presence of risk factors in ischemic stroke (hypertension, atrial fibrillation, diabetes mellitus, hyperlipidaemia, tobacco smoking, cardiovascular disease, previous stroke), occurrence of pneumonia at any given point during hospitalisation (determined by the treating physician due to clinical signs of pneumonia, elevated inflammatory markers in the blood and chest X-ray), treatment on intensive care unit, necessity of intubation and mechanical ventilation lasting longer than 24 hours (excluded from this item were patients intubated for surgery, such as decompressive craniotomy in cerebellar infarction or those who were preclinically intubated and promptly extubated), mortality, presence of dysphagia (as determined by the FEDSS, see below) and type of oral intake (before and after FEES; as determined by the FOIS, see below). For acquisition and use of data for scientific analyses, ethical approval was obtained from the local ethical committee (protocol number 208/16). All patients documented in the database were selected for the current analysis.
FEES
FEES is a videoendoscopic nasolaryngeal swallowing study. We performed FEES following a standardised FEES® protocol, according to Langmore [14]: after applying decongestants (Xylometazolin) and local anaesthesia of the nasal duct using 2% Lidocaine-gel, a small endoscope (about 4 mm in diameter) was introduced through the inferior nasal meatus and the nasopharynx in the oropharynx. The swallowing of saliva and different consistencies of food and liquids and penetration, aspiration, localisation and extent of residues, as well as patients’ reactions (such as coughing), were visualised and documented. By definition, penetration is entering of any material into the airway above the level of the vocal folds, and aspiration is entering of any material below the level of the vocal folds [15]. In the first step of the procedure, anatomical changes, handling of saliva and the movement of swallowing-related structures were tested, then pudding-thick consistency (thickened water), normal water and solid food were introduced. For every consistency, we first used a teaspoon, then a table spoon and in case of water, the patient was asked to take a normal swallow from a cup. For better visualisation, the consistencies were dyed blue, using food colour. All consistencies were applied three times. If a consistence appeared unsafe to test, we skipped it; the consistence was rated as unsafe, if it entered the airway to the level of the vocal folds without ejection from the airway or any aspiration (score 5-8 on Rosenbek’s Penetration-Aspiration-Scale [16]). In the context of this research manuscript, we defined a “relevant dysphagia” as an oropharyngeal dysphagia with a score of 3-8 on Rosenbek’s Penetration-Aspiration-Scale, as this exposes the patient to the risk of pneumonia. Using the findings in FEES, the appropriate oral diet was chosen for the patient. Based on the pathophysiology found in FEES, compensatory and rehabilitative measures to treat dysphagia were carried out as described by Daniels and co-workers [17]. All FEES procedures were performed or supervised by an experienced investigator and lasted about 10 minutes each.
Outcome measurements
Oral intake and dysphagia severity were measured by use of the functional oral intake scale (FOIS) and the Fiberoptic Endoscopic Dysphagia Severity Score (FEDSS), respectively:
FOIS is a seven-tiered scale ranging from 1 = no oral intake at all (NPO= nil per os) to 7 = full oral intake without restrictions (Appendix 1)[18]. For easier readability, the data of the functional oral intake scale were categorised in either NPO (FOIS=1), tube dependency with at least some oral intake (FOIS 2-3), patients without tube dependency with dietary restrictions (single consistency, special preparations or limited specific food) (FOIS =4-6) and oral diet without restriction (FOIS=7). Restriction of the oral diet was defined as a negative change on the FOIS, whereas lowering of restrictions of the oral diet was defined as a positive change. FOIS was documented prior to and after FEES.
There is no standardised way of defining the overall severity of dysphagia. In our department, we use the FEDSS-scale developed by Dziewas and co-workers [19]. The FEDSS is a six-tiered scale originally designed for use in stroke patients (Appendix 2). All parameters were recorded in a standardised way.
For evaluating the value of performing FEES in neurological patients, the following parameters were correlated with baseline data and dependent factors:
- Dysphagia as defined by a FEDSS score of ≥ 2
- The oral intake status as calculated by the FOIS and its overall change and type of change after FEES
Statistical analyses
Absolute and relative frequencies were calculated based on cross-tables. For comparing relative frequencies, we used a two-tailed Fisher’s exact test. Continuous variables were analysed by calculating the median value and the interquartile range (IQR; 25% percentile and 75% percentile). Nonparametric non-paired data were analysed using the Mann–Whitney U-test and paired data using the Wilcoxon-test. Binary logistic regression analysis was used to identify factors associated with the item “change in oral diet”. All statistical analyses were performed using SPSS, release version 22.0 (SPSS©, Inc., IBM Company, 2015, Chicago-IL).