Dysphagia Screening in Parkinson's Disease. A diagnostic accuracy cross‐sectional study investigating the applicability of the Gugging Swallowing Screen (GUSS)

Simple water‐swallowing screening tools are not predictive of aspiration and dysphagia in patients with Parkinson's Disease (PD). We investigated the diagnostic accuracy of a multi‐texture screening tool, the Gugging Swallowing Screen (GUSS) to identify aspiration and dysphagia/penetration in PD patients compared to flexible endoscopic evaluation of swallowing (FEES).


| INTRODUC TI ON
Neurogenic dysphagia is highly prevalent in patients with Parkinson's disease (PD) with significant consequences for nutritional status, quality of life, and outcome. 1 Dysphagic symptoms may occur in the early stages of the disease, 2 however, self-perception of dysphagia might be limited in PD. 3 Typical dysphagia-related findings include oral stage impairments such as dyscoordinated tongue movements, impaired masticatory function, prolonged oral transit time (OTT), and pharyngeal stage impairments including residues, penetration, aspiration, and pharyngoesophageal dysmotility. 4,5 The pathology behind aspiration seems to be a combination of pharyngeal sensory and motor impairments. Additionally, many PD patients have problems with generating sufficient expiratory flow and subglottic pressure to cough effectively. 6 A gold standard method to identify aspiration in PD patients is flexible endoscopic evaluation of swallowing (FEES) 7,8 as it detects dysphagia in more than 50% of subjectively asymptomatic patients. 3 However, instrumental swallowing assessment is associated with costs and burden for the patient and availability depends on clinical settings. This emphasizes the need for a dysphagia screening tool with acceptable sensitivity and specificity to identify aspiration to determine the necessity for continued further instrumental assessment throughout the progression of the disease.
To date, no PD-specific dysphagia screening tool exists. Simple water-swallowing tests, measurements of volume and speed while drinking fluids and swallowing questionnaires are not predictive of aspiration in PD patients. 9,10 Considering the complex pathophysiology behind dysphagia in PD and the increase of impairments with swallowing solid food, 2 dysphagia and aspiration could be noticeable earlier in this patient group when multi-texture screening tools are applied. The Gugging Swallowing Screen (GUSS), 11 a widely used multi-texture tool developed and validated for acute stroke patients, might be applicable. The GUSS includes successive evaluation of saliva swallows and three different bolus textures (semi-solid, liquid, solid) based on the occurrence of established signs of dysphagia and aspiration. A sum score is calculated and interpreted in terms of aspiration risk and dysphagia risk, and diet recommendations are provided. Sensitivity (>96%) and specificity (>50%) of the GUSS to detect aspiration risk in acute stroke patients and good inter-rater reliability (κ = 0.84) have been established. 11,12 Diet recommendations derived from the GUSS sum score were found to be more conservative than those determined by a FEES-based assessment rated with the Fiberoptic Endoscopic Dysphagia Severity Scale (FEDSS), 13 with the GUSS overestimating the need for a non-oral diet (Nil Per Os -NPO) and tube feeding in stroke patients. 12 The objective of this study was to investigate the applicability of the GUSS in patients with PD by: • establishing inter-rater reliability when the GUSS is applied in this patient group • determining the diagnostic accuracy of the GUSS to identify aspiration and dysphagia/penetration compared to a gold standard (FEES) • investigating the agreement between GUSS-based and FEES-based diet recommendations 2 | MATERIAL AND ME THODS

| Ethical approval
Data collection in this prospective cross-sectional study was conducted at the University Medical Center Hamburg-Eppendorf between January 16, and January 31, 2019. It was approved by the local ethics committee of the Medical Council Hamburg. Written informed consent was obtained from all participants.  15 Hoehn and Yahr stage (H&Y) 16 was ascertained; for data analysis, H&Y 2.5 was included in stage 2 and H&Y 3.5 was included in stage 3.

| Procedures
All participants were examined during a single visit to the Department of Voice, Speech and Hearing Disorders in the clinical "on-medication"-stage, which describes the patients' state one hour after medication intake. Each participant underwent two swallowing examinations: (i) the GUSS and (ii) a FEES examination.
The order of the procedures was randomized, and all examiners were blinded to the results of the respective other assessment. Patients were asked to drink the water quickly but not as fast as

| Data analysis
The GUSS examinations were rated by 2 independent and blinded SLPs according to the annotated version of the GUSS and the GUSS manual (https://www.dysph agie-trapl.at/guss-formu lar-deuts ch/). 11 19 In setting the cutoff scores for aspiration and dysphagia/ penetration, we replicated procedures of the previous GUSS validation studies. 11,12 "Aspiration" was defined as a GUSS cutoff score of ≤14 and a PAS score of ≥6. "Dysphagia/penetration" was defined as a GUSS cutoff score of ≤19 and a PAS score of ≥3. Instead of the terms "aspiration risk" and "dysphagia risk" used in the GUSS, we used "aspiration" and "dysphagia/penetration" in the following analyses to standardize reporting of the results. Diet recommendations were derived from the FEES examination using the FEDSS 13 score. The FEDSS scaling was slightly adjusted as shown in Table 2 to allow comparison with the GUSS' recommendations.

| Statistical procedures
Quantitative data describing participants' characteristics were Inter-rater reliability of the two independent GUSS ratings was established with the Spearman rank correlation coefficient (r s ). Interrater agreement of the FEES ratings was determined by Bravais-Pearson's correlation coefficient.
Accuracy of the GUSS (compared to FEES) to detect aspiration and dysphagia/penetration was analyzed with a 2 × 2 contingency All statistical tests were two-tailed, and a priori significance levels were set at p < 0.05. Data were analyzed with the statistical software package SPSS, version 24.0 (IBM USA).

| FEES results
All FEES examinations were tolerated without complications.

| Inter-rater reliability of the GUSS in patients with PD
Four GUSS videos had to be excluded from the inter-rater analysis due to insufficient acoustic quality. The two independent GUSS ratings of the remaining 47 participants showed high inter-rater reliability (r s = 0.8; p < 0.001). Disagreements resulted from different ratings on the parameters coughing (voluntary cough or cough after swallow) and voice change. Ratings of rater #1 (that had been obtained online during the examination) were used for further analyses.

| Accuracy of the GUSS vs. FEES to detect aspiration and dysphagia/penetration
The GUSS detected aspiration with a sensitivity of 50% and a specificity of 51.3% (

| Diet recommendations: GUSS vs. FEDSS vs. self-report
No significant agreement was found between the GUSS vs.

| DISCUSS ION
Swallowing impairments are frequent in patients with Parkinson's disease and may occur long before patients notice and report their difficulties. 3 There is a need for disease-specific screening tools that allow for repeated evaluation and early identification of aspiration and dysphagia to ensure timely and efficient referral to instrumental assessment. In our study, the applicability of the Gugging Swallowing Screen (GUSS) 11 in patients with PD was investigated by establishing inter-rater agreement and determining the diagnostic accuracy to detect aspiration and dysphagia when applied in this patient group.
Inter-rater reliability of the GUSS in our cohort of PD patients was good, and by this, previous results in stroke patients were confirmed. 11,12 Aspiration was identified with a sensitivity of 50% and specificity of 51% indicating that the GUSS in its current form is not applicable as a screening instrument to predict aspiration in PD patients with acceptable accuracy. Dysphagia/penetration was identified with higher sensitivity (73%); however, this is considerably lower than in stroke patients (>96%) 11,12 and below our determined accuracy threshold (≥80%). Furthermore, with a specificity of 36% in detecting dysphagia/penetration, the intended benefit of a dysphagia screening tool-that is, to enhance efficiency of referral to further instrumental assessments-cannot be met, and over-referral is very likely. While likelihood ratios indicate that the GUSS is associated with the occurrence of aspiration and dysphagia in this patient group, the diagnostic contribution to establish these diagnoses is rather small. Therefore, the test should not be used as the only screening measure but in combination with expert clinical and instrumental swallowing examination.
Our data replicate previous findings that the GUSS' diet rec- Our results also provide approaches to possible modifications of the GUSS that might enhance its applicability in PD patients. The most crucial GUSS parameters were "coughing," "voice change" and "delayed swallowing," while "reduced vigilance", "drooling" and "absent swallowing response" were not detected in any patient. These findings corroborate previous results. 23 The low specificity of the GUSS may be partly due to its focus on the perceptual evaluation of "effective coughing" and "voice change" after bolus intake. This is probably a crucial discriminating factor of swallowing screening tests  Table 2 PD patients has been widely suggested [25][26][27] and might increase the validity of the GUSS as a dysphagia screening tool in this patient group.
"Wet voice" or "voice change" is a commonly used indicator in swallowing screening tests, 28 however, the validity of this parameter to identify aspiration is not straightforward. Some studies indicate that inclusion of this parameter enhances sensitivity in water-swallowing tests, 29 others found that sensitivity varies with increasing volumes and different bolus consistencies in multi-texture screening tools. 30 Voice quality is particularly difficult to assess in patients with PD as their voice quality deteriorates as the disease progresses and impairments in vocal cord function can interfere with perceptual evaluation of wet voice. 31 In a study with PD patients and experienced raters, wet voice was the least perceived vocal abnormality compared to hoarseness and tremor and its inclusion into a multi-texture swallowing screening contributed to specificity but not to sensitivity in detecting aspiration. 30

| Limitations
The GUSS was developed and validated for acute stroke patients, and the same applies to the FEDSS. Thus, a direct transfer of these procedures to PD patients might be questionable in the first place.
However, we aimed to investigate the applicability of an existing and widely applied, validated multi-texture screening tool in this patient group. Data were collected during a short time of patient recruitment, while a longer data collection period could have enhanced the sample size in our study. However, this limitation is partly balanced by the distribution of H&Y severity stages in our sample corresponding to the normal distribution across the population.

| CON CLUS ION
The multi-texture screening tool GUSS in its current form, although applicable with good inter-rater reliability, is not able to detect aspiration in PD patients with sufficient accuracy. We suggest a modified version of the GUSS that incorporates adaptations of the parameters cough effectiveness, voice change, and delayed swallowing to account for the complex pathophysiology of swallowing disorders in PD patients. Our data confirm that diet recommendations drawn from screening tools should be verified by adequate clinical and instrumental assessment approaches as the complex progressive limitations of swallowing physiology in PD patients need a holistic approach to dietary adaptations.

ACK N OWLED G M ENTS
We would like to thank MD Odette Fründt and Ole Schlickewei for their help with data acquisition and Dr. Carola de Beer for her advisory support.

AUTH O R CO NTR I B UTI O N S
UF, JR, CP, JCN, and AN performed study conception and design, data collection, data analysis and interpretation, drafting/revising manuscript, including medical, and final approval. BS and JCK performed data collection and data interpretation and contributed to drafting and revising the manuscript. CP, CB, MPN, and CG were clinical supervisors, involved in protocol writing, experimental session, and interpretation of data and also contributed to drafting and revising the manuscript.

D I SCLOS U R E
The authors declare that they have no conflict of interest.