Received: 16 May 2019 | Revised: 16 July 2019 | Accepted: 23 July 2019 DOI: 10.1111/nmo.13695 ORIGINAL ARTICLE Effect of a gum‐based thickener on the safety of swallowing in patients with poststroke oropharyngeal dysphagia Mireia Bolivar‐Prados1 | Laia Rofes2 | Viridiana Arreola1 | Sonia Guida3 | Weslania V. Nascimento1 | Alberto Martin1 | Natàlia Vilardell1 | 1 3 3 Omar Ortega Fernández | Dina Ripken | Mirian Lansink | Pere Clavé1,2 1 Gastrointestinal Physiology Laboratory, Hospital de Mataró, Universitat Abstract Autònoma de Barcelona, Mataró, Spain Background: Increasing viscosity with thickening agents is a valid therapeutic strat‐ 2 Centro de Investigación Biomédica en Red egy for oropharyngeal dysphagia (OD). To assess the therapeutic effect of a xanthan de Enfermedades Hepáticas y Digestivas (Ciberehd), Barcelona, Spain gum‐based thickener (Nutilis Clear®) at six viscosities compared with thin liquid in 3 Danone Nutricia Research, Nutricia poststroke OD (PSOD) patients. Advanced Medical Nutrition, Utrecht, The Netherlands Methods: A total of 120 patients with PSOD were studied in this controlled, multiple‐ dose, fixed‐order, and single‐blind study using videofluoroscopy (VFSS). A series of Correspondence Pere Clavé, Department of Surgery, Hospital boluses of 10 mL thin liquid and 2000, 1400, 800, 450, 250, and 150 mPa s viscosi‐ de Mataró, Universitat Autònoma de ties were given in duplicate, interrupted in case of aspiration. We assessed the safety Barcelona, Carretera de Cirera s/n, 08304 Mataró, Spain. and efficacy of swallow and the kinematics of the swallow response. Email: pere.clave@ciberehd.org Key Results: A total of 41.2% patients had safe swallow at thin liquid which signifi‐ cantly increased for all viscosities from 71.9% at 150 mPa s to 95.6% at 1400 mPa s Funding information (P < .001). PAS score (3.7 ± 2.3) at thin liquid was also reduced by increasing bolus vis‐ This work has been supported by Danone Nutricia Research, Fundació cosity (P < .001). The prevalence of patients with aspiration at thin liquid was 17.5% Salut del Consorci Sanitari del Maresme, and decreased at all viscosities (P < .01), except at 150 mPa s. Increasing viscosity Furega (Fundació per a la Recerca en Gastroenterologia), Centro de Investigación shortened time to laryngeal vestibule closure (LVC) at all viscosities (P < .01) and re‐ en Red en Enfermedades Hepáticas y duced bolus velocity at ≥450 mPa s (P < .05). The prevalence of patients with pharyn‐ Digestivas (CIBERehd), Instituto de Salud Carlos III (ISCIII), Barcelona, Spain; Fondo de geal residue at each viscosity 37.7%‐44.7% was similar to that at thin liquid (41.2%). Investigaciones Sanitarias, ISCIII (14/00453; Conclusions and Inferences: The prevalence of unsafe swallow with thin liquids is 18/00241)/CIBERehd; and Programa de Estabilización de Investigadores/ very high in PSOD. Increasing shear bolus viscosity with this xanthan gum‐based Programa de Intensificación de la Actividad thickener significantly increased the safety of swallow in patients with PSOD in a vis‐ Investigadora en el Sistema Nacional de Salud (INT15/00026; INT16/00111). cosity‐dependent manner without increasing the prevalence of pharyngeal residue. KEYWORDS aspiration, deglutition disorders, shear viscosity, stroke, swallow response, thickener, xanthan gum Clinical trial registration: Netherlands Trial Register with code NTR5628. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2019 The Authors. Neurogastroenterology & Motility published by John Wiley & Sons, Ltd. Neurogastroenterology & Motility. 2019;31:e13695. wileyonlinelibrary.com/journal/nmo | 1 of 11 https://doi.org/10.1111/nmo.13695 2 of 11 | BOLIVAR‐PRADOS et al. 1 | I NTRO D U C TI O N Key Points Oropharyngeal dysphagia (OD) is a motility disorder characterized by • Oropharyngeal dysphagia (OD) occurs in 45% post‐ difficulty forming or moving the alimentary bolus from the mouth to the stroke patients. Increasing bolus viscosity with thick‐ esophagus and can include aspiration.1 Poststroke OD (PSOD) is clas‐ eners reduces aspirations, but optimal viscosity levels sified in the ICD under the code: 438.82 (ICD‐9) and I69.391 (ICD‐10).2 need to be determined. OD is a prevalent complaint following stroke, with high incidence (45%) • We assessed 7 shear viscosity levels with a xanthan on hospital admission.3 It is associated with poor short‐ and long‐term gum‐based thickener in stroke patients with dysphagia prognosis and several complications, such as malnutrition, dehydra‐ and found a viscosity‐dependent improvement in swal‐ tion,4 and aspiration pneumonia, increasing the risk of mortality 5-7 in lowing safety from 150 mPa s to 800 mPa s through comparison with poststroke patients without OD.5,8-10 It is an inde‐ reduced time to laryngeal vestibule closure and bolus pendent risk factor for prolonged hospital stay and institutionalization velocity. after discharge, and for poorer functional capacity and increased mor‐ • This is the first study to show the full dynamics and tality 3 months after stroke.3 While some patients recover spontane‐ mechanisms of gum‐based thickeners in poststroke OD. ously, 50% assessed 6 months poststroke were found to have chronic OD.11 The pathophysiology of PSOD is characterized by several motor impairments in the kinematics of the swallow response including de‐ pharyngeal residue, and the effects on the biomechanics of the layed laryngeal vestibule closure (LVC) and decreased bolus propulsion swallow response.13,15 12 forces ; also, patients affected by unilateral stroke showed a disrupted pattern of sensory cortical activation after pharyngeal stimulation as a 2 | PATI E NT S A N D M E TH O DS distinctive marker of abnormal sensory integration of swallowing path‐ ways in PSOD.13 2.1 | Study population Thickening agents increase the viscosity of fluids and thin liquids, enhancing the safety of swallow by avoiding aspirations and their This study included 120 PSOD outpatients who were consecutively associated complications,14,15 as stated in a review by the European recruited from March 2016 to December 2017 at the GI Physiology Society for Swallowing Disorders (ESSD).16 Viscosity is a rheological Lab of the Hospital de Mataró, Barcelona following hospital discharge. property which measures the resistance of a fluid to flow, expressed Assuming discordant proportions of 7.5% (safe swallow on thin liquid in SI units as mPa s 15,17—rheology is the study of the flow and defor‐ and unsafe swallow on main viscosities) and 30% (unsafe swallow on mation of fluids.18,19 Several factors can affect the viscosity of thick‐ thin liquid and safe swallow on main viscosities), a sample size of 95 ened fluids: salivary α‐amylase breaks down starch molecules during patients would be sufficient to have 90% power to detect statistical sig‐ the oral phase of swallow,15 and shear thinning decreases viscosity nificant differences in safe swallowing between each of the three main 18,20 with increasing bolus velocity and shear rate in the pharyngeal viscosities and thin liquid, using a two‐sided Mcnemar's test with an α phase. of 0.5, assuming 20% of patients do not complete the measurements. The ESSD review also recommended (a) the development of Main inclusion criteria were patients older than 18 years, minimum of new thickening agents with less residue, more palatability and, 28 days since diagnosis of stroke, clinical signs or symptoms of swal‐ thus, better compliance (gum‐based thickeners have proven to be lowing dysfunction in the volume‐viscosity swallow test (V‐VST) 22 or 20 better than starch) ; and (b) clinical trials to establish the optimal referral by physician for VFSS or current use of thickened products, no viscosity level for each phenotype of dysphagic patients.16 Few alteration in consciousness, and written informed consent. Main exclu‐ viscosity levels per product have been studied, and the optimal vis‐ sion criteria were need of oxygen therapy, OD not related to stroke, cosity levels for patients suffering poststroke OD have not been history of other neurological disorders or head and neck cancer, xeros‐ determined yet.16 tomia induced by drugs, severe cognitive disorder, incapability to per‐ The aim of this study was to assess the effect of a gum‐based form VFSS, pregnancy or lactation, participation in another research thickener (Nutilis Clear ®) on the safety and efficacy of swallow‐ study, and allergy to any ingredient tested. In addition, for the descrip‐ ing in patients with poststroke OD by evaluating seven different tion of study population, we collected demographic parameters such shear viscosities (150‐2000 mPa s) during swallowing with vid‐ as age, sex, weight, height, type of stroke, time after stroke, severity eofluoroscopy swallowing study (VFSS). There are no previous of dysphagia, nutritional status, comorbidities, medication, and stroke studies that have evaluated such a wide range of viscosities. The severity according to the National Institute of Health Stroke Scale primary objective was to assess the percentage of patients that (NIHSS).23 could swallow safely at each of the three main viscosities (2000, The Ethics Committee of the Hospital de Mataró (Spain) approved 800 or 250 mPa s) compared with thin liquid. Secondary and ex‐ the study protocol, information given to patients about the study ploratory objectives were to assess the effect of all viscosities on and the informed consent form with code 41/15. The study was con‐ penetrations, aspirations, the Penetration‐Aspiration Scale (PAS) ducted according to the principles of the “World Medical Association developed by Rosenbek, 21 the presence and severity of oral and Declaration of Helsinki” (2013) and the International Conference BOLIVAR‐PRADOS et al. | 3 of 11 FIGURE 1 Study design on Harmonization (ICH) guidelines for Good Clinical Practice (GCP, 2.4 | Methods September 1997) as appropriate for nutritional products legislation of Spain where the study took place. This study has been registered in The 2.4.1 | Videofluoroscopy (VFSS) Netherlands Trial register with code: NTR5628. VFSS is a dynamic radiological exploration that evaluates the swallow‐ ing process of boluses of various volumes and viscosities marked with 2.2 | Experimental design a radiopaque iodine contrast.12 Boluses were tested while the patient This was a reference‐controlled, multiple‐dose, fixed‐order, single‐ was seated in a lateral projection. Boluses were prepared with water, blind, and single‐center study. The study procedure (Figure 1) was X‐ray contrast solution (Omnipaque™, GE Healthcare), and the required performed in one single visit. Firstly, the V‐VST—a clinical assess‐ amount of thickener (g) to achieve each viscosity level (mPa s). In our ment tool for dysphagia—was performed on each patient to assess research group, volumes of 5, 10, and 20mL are routinely used in the clinical signs of OD 23-25 and those positive for OD were referred for clinical practice to test the swallowing ability of the patient in an effort VFSS. One week after the completion of the study, a follow‐up call test.12,22 As not all the patients are capable of swallowing the maximum was performed to assess potential adverse events. volume (20mL), 10mL was chosen as an optimal comfortable bolus for During the VFSS, 10mL boluses were given in duplicate to each the patient to swallow in this study. The oral cavity, pharynx, larynx, patient, following the algorithm shown in Figure 1 (only one bolus and cervical esophagus were recorded on video during swallowing. is shown in the algorithm, but two were given if the patient swal‐ VFSS recordings were obtained using a Super XT‐20 Toshiba Intensifier lowed safely). Briefly, the procedure started with thin liquid (when (Toshiba Medical Systems Europe) and recorded at 25 frames/s using aspirations occurred, the second bolus of thin liquid was not admin‐ a Canon DM‐XM2 E video camera (Canon Inc.). The VFSS record‐ istered to protect patients from a new aspiration) and continued ings were analyzed and the measurements obtained using specialized with boluses from the highest viscosity to the lowest. If the patient software (Swallowing Observer; Image & Physiology SL) by an expert aspirated any of the thickened boluses, the study was terminated to blinded clinician.14 VFSS signs. Safety of swallow was assessed by the 16,26 avoid any further aspiration as a safety measure. identification of the PAS score and the prevalence of safe swallows (PAS 1,2), penetrations (PAS 3,4,5), or aspirations (PAS 6,7,8) of each bolus.21 We considered the following signs as indication of impaired 2.3 | Outcome parameters efficacy: piecemeal deglutition, oral, pharyngeal wall, and vallecular or The main outcome parameter was the percentage of patients with safe pyriform sinus residue. The prevalence of residue was described as the 21 swallow (PAS score 1 and 2) for the main viscosities (250, 800, and presence or absence of residue in the oral cavity or the pharynx includ‐ 2000 mPa s). Secondary outcome parameters were as follows: (a) safety ing the pharyngeal wall, the vallecula, and pyriform sinus.27 Timing of of swallowing expressed by the mean PAS score,21 and the percentage of oropharyngeal swallow response (OSR) and bolus kinematics. Timing of patients with penetration (PAS score of 3,4,5), or aspiration (PAS score of swallow response was assessed for each bolus given to the patient dur‐ 6,7,8); and (b) the efficacy of swallowing expressed by the presence and ing VFSS.4,27 We measured the time to LVC (time from the glossopalatal severity of oral and pharyngeal residue. Exploratory parameters included junction (GPJ) opening to the LVC) and the total duration of the swallow physiology of swallowing (time to LVC, total duration of swallowing re‐ response. In addition, mean bolus velocity of the bolus between the sponse –LVO–, mean bolus velocity, and bolus propulsion force), distribu‐ GPJ and the upper esophageal sphincter (UES), propulsion forces, and tion of PAS scores, subjective swallowing experience at all viscosities (150, kinetic energy were calculated as published before by our group.4 250, 450, 800, 1400, and 2000 mPa s) and safety and efficacy of swal‐ lowing at the 3 exploratory viscosities (150, 450, and 1400 mPa s). Due 2.4.2 | Bolus rheology to the relevance of the information for patient safety, comparisons on the prevalence of patients with safe swallow and mean PAS scores were also The viscosity levels used in the study were selected according to the performed between all the different viscosities assayed in this study. descriptors of the National Dysphagia Diet Task Force: 1‐50 mPa s 4 of 11 | BOLIVAR‐PRADOS et al. for thin liquid, 51‐350 mPa s for nectar, 351‐1750 mPa s for honey, observed at any of the three pharyngeal locations (pharyngeal wall, −1 28 and >1750 mPa s for pudding viscosities at 25°C and 50 seconds . vallecular, and pyriform sinus), the residue was present (yes); if no For VFSS, we prepared a total of seven different viscosities in 10 mL residue was observed at any of the locations, the residue was absent bolus, consisting of liquid X‐ray contrast as control vs six thickened (no); and if at least one was missing (not performed due to the safety X‐ray contrasts thickened with Nutilis Clear®—consisting of malto‐ rule) and the others were absent, the residue was handled as missing. dextrin, xanthan gum, and guar gum (Nutricia N.V., Zoetermeer, The Data of the duplicates for residue were handled according to the Netherlands)—at each viscosity level. To achieve those viscosities, vary‐ algorithm shown in Figure S1. For efficacy of swallow, an additional ing amounts of thickener were added to 50 mL solution composed of procedure for handling duplicates was used to explore the “worst 1:1 mineral water and the iodine X‐ray contrast: 150 and 250 mPa s vis‐ case” scenario. This selection was independent of PAS score, and the cosities were obtained by adding 0.56 g and 0.75 g, respectively; 450, replicate was selected based on the worst value for the presence of 800, and 1400 mPa s viscosities were obtained by adding 1.27 g, 2.08 g, pharyngeal or oral residue. and 3.81 g, respectively; and 2000 mPa s was obtained with 5.01 g. Statistical tests were conducted two‐sided with a significance level of 5%. All confidence intervals are presented two‐sided with a confidence level of 95%. A resultant probability value of 2.4.3 | Comfortability P < .05 was judged as statistically significant. For the primary out‐ During the VFSS, patients were asked whether they felt comfortable come parameter, percentage of patients that swallow safely, the during the swallowing experience (“I felt comfortable during swallow‐ null‐hypothesis of no effect on safe swallowing of 2000, 800, ing this product”) using a 9‐point Likert scale, at each viscosity level. and 250 mPa s compared with liquid will be rejected if all three Results are presented by using three categories: (a) strongly agree, (two‐sided) P values are <.05 with correct directional decisions. An agree, and moderately agree; (b) mildly agree, undecided, and mildly additional explorative analysis was performed on safety of swal‐ disagree; and (c) moderately disagree, disagree, and strongly disagree. lowing and the mean PAS scores to evaluate the therapeutic ef‐ fects between viscosities. As a post hoc test, the bolus propulsion force was analyzed, and dose‐response curves for the viscosity‐ 2.4.4 | Safety of the product dependent effect of the thickening agent on safety and efficacy All adverse events (AEs) occurring during the study and one week were obtained by representing the prevalence of patients with safe after the procedure (follow‐up telephone call) were recorded and swallowing and those with residue respectively at each level of vis‐ assessed for relationship with the study product according to the cosity using Graphpad Prism 6. guideline of categories described by the World Health Organization and the Uppsala Monitoring Centre (WHO–UMC). 29 3 | R E S U LT S 2.4.5 | Data analysis and statistical methods 3.1 | Sample demographics Binary data were described as relative and absolute frequencies, and Of the 120 patients enrolled, 4 were excluded from the all sub‐ the viscosity levels were compared with thin liquid by applying the jects treated (AST) population because they did not receive any McNemar's test. For ordinal data, the comparisons were done by ap‐ of the thickened viscosities. Additionally, two patients were ex‐ plying the Bhapkar's test; in case of zero counts, the McNemar's test cluded from the per protocol population (PP) because they discon‐ on aggregated categories was used. Continuous data are presented tinued due to reasons other than aspiration which was regarded as mean ± standard deviation (SD), and comparisons were done by as a protocol deviation. The originally planned analysis was on applying a repeated measure mixed model including all six viscosities the intention‐to‐treat population (ITT). However, because there or a paired‐sample Wilcoxon signed rank test in case assumptions were 4 patients in this population who did not receive any of the were not met. The McNemar's test, Bhapkar's test, paired‐sample thickened product, it was decided to present the results for the Wilcoxon signed rank test, and repeated mixed model all take into PP population (n = 114) (Figure S2). The results of the ITT and PP account the within‐subject design and paired data. The statisti‐ populations were comparable. The majority of our population, 76% cal analysis was performed with SAS® software for Windows, SAS (N = 87) were in the subacute phase (28‐180 days after stroke) and Institute Inc. (SAS version 9.4_M1). 24% (N = 27) were chronic (>180 days after stroke). Mean age of Safety of swallow of each patient at a particular viscosity level the participants was 76.7 ± 8.9 years, and 54.4% were men. The was expressed as the worst PAS score of the duplicates, and all the MNA‐SF total score indicated that 54.4% of patients were mal‐ parameters of that replicate were analyzed according to the scheme nourished or at risk of malnutrition when enrolled in the study. in Figure S1. Data on safety of swallowing were handled as binary Stroke type was predominantly ischemic 78.1% (n = 89), and the by dividing the patients in two categories: patients who can swal‐ prevalent severity of the stroke valued with the NIHSS was scored low safely (PAS 1‐2) vs patients who cannot swallow safely (PAS 3‐8) (mean ± SD) 7.5 ± 6.8 on admission and 5.3 ± 5.9 on discharge. over the “per protocol” population. The efficacy of swallowing was More details of the epidemiological and clinical characteristics of also handled as binary data (presence or absence): if residue was the population are provided in Table S1. BOLIVAR‐PRADOS et al. | 5 of 11 F I G U R E 2 Percentage of PSOD patients with safe/unsafe swallow at each level of viscosity. “N” represents the number of patients who performed the bolus out of the PP population (114). The percentage of patients with unsafe swallow includes those with aspirations at the former viscosity who discontinued due to the safety rule. Percentage of patients who discontinued at each viscosity: thin liquid (0.0%), 150 mPa s (12.3%), 250 mPa s (8.8%), 450 mPa s (4.4%), 800 mPa s (1.8%), 1400 (1.8%), 2000 mPa s (0.9%). *P < .05; **P < .01; ***P < .001 vs thin liquid F I G U R E 3 Percentage of PSOD patients with safe/unsafe swallow compared between levels of viscosity. Data of patients who discontinued due to the safety rule were imputed with the last observation carried forward. Values are presented for the PP population (114).*P < .05; **P < .01; ***P < .001 and 1.4 ± 1.2 by increasing viscosity from 150 to 2000 mPa s 3.2 | Effect of range of viscosities on prevalence of (all P < .001 vs thin liquid). The distribution of safe swallowing, VFSS signs of OD penetration, and aspiration was significantly different at all vis‐ cosities compared with thin liquid (all P < .001 vs thin liquid). The 3.2.1 | Safety of swallow percentage of patients with penetration and aspiration decreased Primary parameter when viscosity increased (Figure S3). The prevalence of patients Safe swallowing was observed in only 41.2% (n = 47) of the patients with penetrations at thin liquid was 41.2% and ranged between at thin liquid but the percentage significantly increased with the 2.6% and 13.2% for the thickened viscosities. The prevalence of main viscosities (all P < .001 vs thin liquid) (Figure 2). Similarly, safety patients with aspirations showed significant differences (P < .01) of swallowing significantly increased with the explorative viscosi‐ with thin liquid (17.5%) vs all viscosities (0.0%‐4.4%) except for ties compared with thin liquid (all P < .001 vs thin liquid) (Figure 2). 150 mPa s (2.5%, P = .180). Mean PAS score at thin liquid was 3.7 ± 2.3, and it significantly Figure 3 shows the explorative analysis of the between viscos‐ decreased to 1.9 ± 1.4, 1.8 ± 1.6, 1.7 ± 1.6, 1.4 ± 1.2, 1.2 ± 0.6, ity comparisons. Among the different viscosity levels, there were 6 of 11 | BOLIVAR‐PRADOS et al. F I G U R E 4 Percentage of patients with PSOD of the PP population (114) with oral and pharyngeal residue at each viscosity level. “N” represents the population who performed the bolus. *P < .05; **P < .01; ***P < .001 vs thin liquid significant differences between the therapeutic effect of 250 mPa s (78.9%) vs 800 (92.1%), 1400 (95.6%), and 2000 mPa s (91.2%) (all P < .01 vs 250 mPa s), but not between 800 and 2000 mPa s or between 800 and 1400 (P > .05). The maximal therapeutic effect (ceiling effect) was observed at 800 mPa s (92.1% of patients with safe swallowing). 3.2.2 | Efficacy of swallow At thin liquid, pharyngeal residue was present in 41.2% (n = 47) of patients and it did not increase at any of the tested viscosities (37.7%‐44.7%, all P > .05 vs thin liquid) (Figure 4). Oral residue was present in 38.6% (n = 44) at thin liquid and significantly increased at all thickened viscosities (all P < .01 vs thin liquid) (Figure 4). Selecting the duplicate with the “worst case” scenario resulted in comparable results (not shown). 3.2.3 | Dose‐response effect of range of viscosities on safety of swallowing and pharyngeal and oral residue Figure 5 shows the viscosity‐dependent therapeutic effect on safety of swallowing for the tested viscosities. 150, 250, and 450 mPa s offered a protection on safety of swallowing between 71.9% and 82.5% and 800, 1400, and 2000 mPa s a protection between 91.2% and 95.6%. Safety increased in a viscosity‐dependent manner. Pharyngeal residue was not statistically different compared with thin liquid at any of the tested viscosities. Oral residue slightly, but significantly, increased at all viscosities. FIGURE 5 Dose‐response curves for the therapeutic effect of the gum‐based thickener on safety and efficacy of swallowing in patients with PSOD. The upper panel shows the curve of the 3.3 | Effect of range of viscosities on oropharyngeal viscosity‐dependent response represented by the percentage of swallow response (OSR) patients with safe swallows vs the log of the viscosity. The lower panel shows the curve representing the effects on the prevalence 3.3.1 | Timing of OSR of oral and pharyngeal residue vs the log of the viscosity. The shadowed area represents the therapeutic range (150‐800 mPa s) Time to laryngeal vestibule closure (LVC) of the product Time to LVC at liquid viscosity was severely delayed (382.5 ± 139.1 ms) in patients with PSOD. Increasing bolus viscosity (250 mPa s), 304.8 ± 109.6 (450 mPa s), 303.3 ± 94.7 (800 mPa s), ≥150 mPa s shortened time to LVC for all viscosities (Figure 6): mean 300.5 ± 110 (1400 mPa s), and 300.4 ± 107.8 (2000 mPa s) ms LVC for each viscosity was 327.3 ± 108.2 (150 mPa s), 330.1 ± 143.4 (P < .01 vs liquid). Time to LVC was shorter in patients with safe BOLIVAR‐PRADOS et al. | 7 of 11 ≥450 mPa s, caused a significant reduction in bolus velocity for 450 mPa s (0.2835 ± 0.0948; P < .05), 800 mPa s (0.2613 ± 0.0784; P < .001), 1400 mPa s (0.2564 ± 0.0803; P < .001), and 2000 mPa s (0.2729 ± 0.1010; P < .01) vs thin liquid (Figure 7). Bolus propulsion forces Mean bolus propulsion force was 0.041 ± 0.035 mN at thin liq‐ uid. A significant decrease was found at the thickened viscosities (all P < .001 vs thin liquid): 150 mPa s (0.033 ± 0.025), 250 mPa s (0.035 ± 0.032), 450 mPa s (0.030 ± 0.019), 800 mPa s (0.026 ± 0.014), 1400 mPa s (0.025 ± 0.015), and 2000 mPa s (0.028 ± 0.022). 3.4 | Comfortability Comfortability while swallowing scored highest at thin liquid (66.3%), and it decreased significantly to 46.3% and 31.3% during swallowing the main viscosities 800 and 2000 mPa s, respectively (Figure 8). Categories of comfortability were differently distributed at all viscosities compared with thin liquid (all P < .001 vs thin liquid), except for 150 and 250 mPa s (Figure 8). 3.5 | Safety of the product. Adverse events (AEs) A total of 16 adverse events occurred in 11 patients out of the 116 in the AST population and were considered unrelated or unlikely to be related to the study product. The most frequent AEs were mild gastrointestinal disorders (14 AEs in 10 patients): diarrhea, nausea, abdominal distension and pain, dyspepsia, and stomatitis. No serious AEs were reported during or following the study. 4 | D I S CU S S I O N FIGURE 6 Time to LVC at each viscosity level. The upper panel The main result of this study is that increasing bolus viscosity with shows mean time to LVC at each viscosity. The lower panel shows the xanthan gum‐based thickener Nutilis Clear ® significantly in‐ time to LVC plotted against safe/unsafe swallow at each viscosity creased the safety of swallow in patients with PSOD in a viscos‐ level. Time to LVC was delayed in patients with unsafe swallowing at all viscosity levels except for 2000 mPa s. Time to LVC <160 ms ity‐dependent manner. The study also shows that these patients (green line): safe swallowing as established in a study with healthy presented a pattern of OD with highly prevalent and severe signs volunteers.4Time to LVC ≥340 ms (red line): cutoff time to detect of impaired safety and efficacy of swallow, aspirations and oro‐ the presence of unsafe swallowing in poststroke patients according pharyngeal residue, a severely impaired swallow response, and a to previous studies.12 *P < .05; **P < .01; ***P < .001 high prevalence of malnutrition or risk of malnutrition. Together (PAS 1‐2) vs unsafe swallow (PAS 3‐8): significant differences were these characteristics place these patients at high risk for severe detected in all viscosities except for 2000 mPa s (Figure 6). nutritional and respiratory complications. Impaired safety of swal‐ At thin liquid, the total duration of the swallow response was low in these patients with PSOD is associated with a severe delay 1020.9 ± 220.8 ms and significantly decreased to 947.1 ± 228.7, in time to LVC. An unexpected, but very relevant, result of this 998.8 ± 472.1, 944.1 ± 180.2, 943.1 ± 221.4, 953.5 ± 225.3, and study was that increasing viscosity with this gum‐based thicken‐ 943.2 ± 234.8 ms at 150, 250, 450, 800, 1400, and 2000 mPa s, re‐ ing agent significantly improved airway protection mechanisms spectively (all P < .01 vs thin liquid). by reducing time to LVC. Increasing bolus viscosity also caused a slight, but significant, increase in oral residue and decreased tongue propulsion forces, and decreased bolus velocity at high 3.3.2 | Bolus kinematics viscosity levels without any significant effect on pharyngeal resi‐ Mean bolus velocity due. Finally, the study shows that the gum‐based thickener is safe Poststroke patients included in the study presented a mean bolus and well tolerated in patients with PSOD as reflected by the low velocity at liquid of 0.3138 ± 0.1265 (m/s). Increasing bolus viscosity, number of AEs. 8 of 11 | BOLIVAR‐PRADOS et al. The chronic PSOD population is a phenotype of patients with have a significant effect on the safety of swallowing and 800, 1400, OD that is growing in Europe, due to the increasing incidence of and 2000 mPa s showed a similar level of protection. Aspiration is stroke events (from 1.1 million per year in 2000 to an estimated 1.5 the most severe impairment in swallowing safety. For this param‐ million per year in 2025,30 the progressive increase in the prevalence eter, the minimal viscosity with a significant effect was 250 mPa s, of stroke survivors, and the high prevalence of OD among these pa‐ which suggests a therapeutic range starting at 250 mPa s. However, 5 3 tients (50%‐81%), even among those with mild strokes (45%). We current results on aspiration can be considered inconclusive to es‐ and others have found that mild stroke survivors are at high risk of tablish the lower level of the therapeutic range because the study 3,31 malnutrition and that aspiration pneumonia is the main cause was not powered for this parameter. Low sample size, which was of 1‐year mortality among them.6 The main result of our study is the partly driven by the safety rule, might have prevented us from find‐ viscosity‐dependent effect on safety of swallow with this xanthan ing significant effects on aspiration at the lowest tested viscosity, gum‐based thickening agent in these patients with PSOD allowing that is, 150 mPa s., which was proven effective with regard to safe safe deglutition in almost all these poststroke survivors with OD. swallowing. For the main viscosities tested, significant differences The therapeutic range of this thickening agent in this phenotype of in the therapeutic effect on safety of swallow vs liquid were found, patients is 150‐800 mPa s, as 150 mPa s was the lowest viscosity to and increasing bolus viscosity above 800 mPa s did not cause any further significant increase in the safety of swallow in this pheno‐ type of patients. As far as we know, this is the first study to assess the effect of seven different viscosities in patients with PSOD. Our results suggest that, using this specific thickening agent, healthcare providers can cover the therapeutic needs of this phenotype of dys‐ phagic patients by using a viscosity between 150 and 800 mPa s. A major question that arises from these results is how to pre‐ scribe the optimal viscosity level of this thickening agent to PSOD. Firstly, these products should be labeled appropriately to promote their safe use.32 Secondly, accurate clinical methods should be used to diagnose OD and to prescribe which viscosity is the most appro‐ priate for each patient with PSOD, as not all these patients can be assessed by instrumental exploration.34 Multiple consistency meth‐ ods for clinical diagnosis of poststroke OD—such as GUSS and the V‐VST—have been recently recommended in PSOD in a guideline developed by the ESO and the ESSD and can be adapted to these viscosities.33 For the V‐VST—that uses only two levels of thickened F I G U R E 7 Mean bolus velocity from GPJO to UESO at each viscosity—250 and 800 mPa s can be considered as the most appro‐ viscosity level. Bolus velocity was reduced above 450 mPa s. priate; 800 mPa s as the viscosity providing the maximal significant *P < .05; **P < .01; ***P < .001 vs thin liquid therapeutic effect for this thickener; and 250 mPa s for patients with F I G U R E 8 Comfortability while swallowing the product. The comfortability while swallowing the product at each viscosity level was evaluated by using a 9‐point Likert scale to the following sentence: “I felt comfortable while swallowing this product.” Likert scale score is divided into three categories for each viscosity. For the statistical analysis, these three catergories and the category of missing values were used. “N” represents the population who answered the question, the category of missing values is not shown in the figure. *P < .05; **P < .01; ***P < .001 vs thin liquid BOLIVAR‐PRADOS et al. | 9 of 11 less severe safety impairment as a safe and comfortable intermedi‐ poststroke patients developed new signs/symptoms of ineffective ate value providing a significant therapeutic effect vs thin liquid and swallow related to poor functional, nutritional, and health status and vs 800 mPa s. 22,24 institutionalization.35 Another study on stroke patients concluded Thickeners are widely used in poststroke OD as a compensa‐ that tongue weakness was also caused by reduce muscle mass of tory therapeutic strategy to avoid aspiration. In a previous study swallow muscles and poststroke sarcopenia.36 Our present results of on similar patients with PSOD, it was found that thickening liquids a reduced bolus propulsion force with the higher viscosities further with either modified starch (MS) or xanthan gum‐based (XG) thick‐ suggest that stroke patients also need specific nutritional and reha‐ eners had a strong therapeutic effect on safety of swallow. 20 The bilitation procedures to increase bolus propulsion forces and tongue prevalence of safe swallow using MS and XG thickeners increased strength by fighting poststroke sarcopenia. Interestingly, pharyngeal with bolus viscosity reaching up to 89%‐92% of patients with PSOD residue, more related to pharyngeal clearance caused by pharyngeal at higher viscosity levels (4000 mPa s for MS and 1700 mPa s for constrictors, was unaffected by increasing viscosity.35 XG), above those used in the present study. In this previous study, In the present study, increasing shear viscosity was obtained by the MS thickener strongly increased pharyngeal residues, whereas adding increasing amounts (grams) of the gum‐based thickener to a the XG increased oral residue at 1700 mPa s but did not increase mixture of water and contrast agent. The obtained shear viscosity is pharyngeal residue at any viscosity. Timing of airway protection the independent variable for this study. Besides shear viscosity, other mechanisms (LVC) and bolus velocity were not affected by either rheological proprieties such as elasticity, adhesiveness, and cohesive‐ of the thickener agents. 20 This was one of the first studies showing ness and different extensional viscoelastic behaviors also may play a an advantage for XG thickeners over MS in PSOD, due to its strong role in swallow physiology.18 The assessment of the effect of exten‐ therapeutic effect on safety, low pharyngeal residue, and amylase sional flows on viscosity of thickening agents is now under develop‐ resistance. The present study is a step forward as the therapeutic ment, and the potential influence of these rheological properties on effect on safety of swallow is also very high (92.1% for 800 mPa s) swallow safety and efficacy in patients with OD is still unknown. and is achieved at lower viscosity levels, the absence of pharyngeal Our study has some limitations. The first one arises from its ex‐ residue is similar, Nutilis Clear® is unaffected by amylase, and—a perimental design as we included a pass/fail safety rule to protect new finding—increasing viscosity with this thickener causes a sig‐ patients from dangerous and unnecessary repeated aspirations. Due nificant reduction of time to LVC over thin liquid. Videofluoroscopic to our design, not all patients received all the viscosities, especially studies have shown that the time to LVC is a critical event in the the lowest levels. This is a quite common situation in pharmacologic/ occurrence of penetrations and aspirations, causing unsafe deglu‐ physiologic studies, to minimize the possibility of serious AEs to pa‐ tition, and time to LVC ≥340 ms predicts unsafe swallow in chronic tients, for example, during progressive effort tests.27 A similar “safety PSOD patients. 20 Such a delay in time to LVC in PSOD associated rule” was used in all our previous studies with thickening agents as with impaired safety of swallow was also observed in this study, al‐ requested by the Ethical Committee.16,25 Because it is clinically rele‐ 4 most doubling the time to LVC of healthy people, and was slightly vant information, in‐between viscosity comparisons were performed above that previously described in comparable patients.12 Reduced by imputing the data of the missing values from the safety rule by car‐ pharyngeal sensitivity and impaired conduction and cortical integra‐ rying the last observation forward. As a consequence of the design of tion of pharyngeal sensory inputs at the stroke site is a key feature the study, care should be taken interpreting these results. However, of chronic PSOD and has been closely associated with impaired this design and our interpretation is the safest from the patient's per‐ safety of swallow and delayed time to LVC.13 In fact, sensory feed‐ spective. Another limitation is that the transversal design of the study back from the bolus is critical to tailor the motor component of the does not provide information on longer term clinical outcome, for in‐ swallow response. Therefore, the reduction in time to LVC caused stance whether the observed improved safety of swallowing with the by the thickening agent suggests a mode of action beyond a simple thickener agent results in fewer respiratory infections. Future longi‐ “compensatory” effect.12,13 Another relevant result of the study is tudinal randomized clinical trials should be performed to confirm the that increasing viscosity—which is a measure of the fluid resistance translation of the strong therapeutic effect of the gum‐based thick‐ to bolus flow—reduces bolus propulsion force and bolus velocity ener on swallowing safety into clinical outcomes including incidence at viscosities greater than 450 mPa s. This effect might explain the of nutritional and respiratory complications.3 Nutritional support and slight, but significant, increase in oral residue as tongue strength is oral care must also be included in these protocols. reduced in these patients.12 This result agrees with a previous study In summary, this study shows that increasing bolus viscosity with from our group which concluded that impaired safety of swallow in Nutilis Clear® causes a strong viscosity‐dependent effect on safety chronic poststroke patients was caused by specific impairments in of swallow in PSOD without increasing pharyngeal residue. Our study swallow response such as a delay in the airway protection mech‐ suggests that the therapeutic effect of the thickener might be caused anisms and weak tongue propulsion force.12 Those results led to a by specific effects on oropharyngeal physiology (mainly time to LVC claim that treatments for these patients should be targeted to im‐ and bolus velocity). To optimize this strong therapeutic effect, clini‐ prove these critical biomechanical events (delay in LVC and reduce cians must provide early diagnosis of PSOD and the prescription of tongue strength). We recently studied the natural history of swallow the required appropriate viscosity by multiconsistency clinical and/or function during the 3‐month period after stroke and found 26% of instrumental methods. This might be appropriate to reduce nutritional 10 of 11 | BOLIVAR‐PRADOS et al. and respiratory complications and improve the prognosis of patients Alberto Martin https://orcid.org/0000-0002-6805-764X with PSOD. We believe these findings will have implications for cur‐ Natàlia Vilardell https://orcid.org/0000-0002-1357-9889 rent clinical practice. Our study clearly shows that the therapeutic effect of thickening agents depends on shear viscosity levels, with Omar Ortega Fernández https://orcid.org/0000-0001-6178-3748 a therapeutic range of 150‐800 mPa s for this xanthan gum‐based Pere Clavé http://orcid.org/0000-0002-0696-8560 thickener multiple consistency methods for clinical diagnosis, and management of poststroke OD can be adapted to this viscosity range for this specific phenotype of patients with OD. This information will REFERENCES improve clinical practice by providing the specific levels of viscosity to 1. Clavé P, Shaker R. Dysphagia: current reality and scope of the prob‐ cover the therapeutic needs of this phenotype of dysphagic patients. lem. Nat Rev Gastroenterol Hepatol. 2015;12(5):259‐270. Fluid thickening must be integrated into compensatory multimodal 2. World Health Organization. International Statistical Classification treatments, such as the minimal‐massive intervention 37 or neurore‐ of Diseases and Related Health Problems 10th Revision. 2016. Available from: http://apps.who.int/class​ifica​tions/​icd10/​brows​ habilitation approaches, aiming to restore swallow function.38 e/2016/en. Accessed February 6, 2019. 3. Rofes L, Muriana D, Palomeras E, et al. Prevalence, risk factors and complications of oropharyngeal dysphagia in stroke patients: a co‐ AC K N OW L E D G M E N T S hort study. Neurogastroenterol Motil. 2018;30(8):1‐10. 4. Rofes L, Arreola V, Romea M, et al. Pathophysiology of oropharyngeal We would like to thank all the participants of this study for their avail‐ dysphagia in the frail elderly. Neurogastroenterol Motil. 2010;22(8):1‐9. ability and cooperation, all the members of our Dysphagia Team and 5. Martino R, Foley N, Bhogal S, Diamant N, Speechley M, Teasell R. the X‐ray Department of Hospital de Mataró for performing this pro‐ Dysphagia after stroke: incidence, diagnosis, and pulmonary com‐ ject during 2 years, and also Jane Lewis for correcting the English. We plications. Stroke. 2005;36(12):2756‐2763. 6. Hilker R, Poetter C, Findeisen N, et al. Nosocomial pneumonia after would also like to thank Danone Nutricia Research and the statistical acute stroke: implications for neurological intensive care medicine. group of Danone Nutricia Research, especially Egbert Biesheuvel, for Stroke. 2003;34(4):975‐981. their support in the data analysis. Finally, we would like to thank the 7. Ickenstein GW, Stein J, Ambrosi D, Goldstein R, Horn M, Bogdahn Department of Medicine of the Universitat Autònoma de Barcelona, U. Predictors of survival after severe dysphagic stroke. J Neurol. 2005;252(12):1510‐1516. Mateu Serra, and Elisabet Palomeras from the Research Unit of the 8. Guyomard V, Fulcher RA, Redmayne O, Metcalf AK, Potter JF, Hospital de Mataró, and Dr Elena Escubedo and Dr Jordi Camarassa Myint PK. Effect of dysphasia and dysphagia on inpatient mortal‐ from the Faculty of Pharmacy, University Barcelona, for their advice ity and hospital length of stay: a database study. J Am Geriatr Soc. on the statistical and pharmacological content. 2009;57(11):2101‐2106. 9. Smithard DG, O'Neill PA, Park C, et al. Complications and outcome after acute stroke. Stroke. 1996;27:1200‐1204. C O N FL I C T O F I N T E R E S T 10. Barer DH. The natural history and functional consequence of dysphagia after hemisphere stroke. J Neurol Neurosurg Psychiatry. Declaration of personal interests: Clavé has served as consultant and 1989;52:236‐241. received research funding from Danone Nutricia Research. Guida, 11. Mann G, Hankey GJ, Cameron D. Swallowing function after stroke: prognostic factors at 6 months. Stroke. 1999;30(4):744‐748. Ripken, and Lansink are employees of Danone Nutricia Research. 12. Vilardell N, Rofes L, Arreola V, et al. Videofluoroscopic assessment of the pathophysiology of chronic poststroke oropharyngeal dys‐ phagia. Neurogastroenterol Motil. 2017;29(10):1‐8. AU T H O R C O N T R I B U T I O N 13. Cabib C, Ortega O, Vilardell N, Mundet L, Clavé P, Rofes L. Chronic post‐stroke oropharyngeal dysphagia is associated with impaired Clavé is a submission's guarantor. Clavé, Rofes, Vilardell, and Lansink in‐ cortical activation to pharyngeal sensory inputs. Eur J Neurol. volved in study concept and design. Arreola, Rofes, Martín, Nascimento, 2017;24(11):1355‐1362. Ortega, and Bolívar‐Prados involved in selection of patients. Rofes, 14. Rofes L, Arreola V, Mukherjee R, Swanson J, Clavé P. The effects of a Vilardell, and Bolívar‐Prados involved in acquisition of data. Clavé, xanthan gum‐based thickener on the swallowing function of patients Guida, Ripken, and Bolívar‐Prados involved in statistical planning and with dysphagia. Aliment Pharmacol Ther. 2014;39(10):1169‐1179. 15. Gallegos C, Brito‐de la Fuente E, Clavé P, Costa Aet al. Nutritional analysis. Clavé, Guida, Ripken, Lansink, and Bolívar‐Prados involved in aspects of dysphagia management. In: Toldrá F, ed. Advances in Food analysis and interpretation of data. Clavé, Guida, and Bolívar‐Prados in‐ and Nutrition Research. 1st ed., Vol. 81. London, UK: Elsevier Inc.; volved in drafting of the manuscript. Clavé and Bolívar‐Prados involved 2017:271‐318. in revision of bibliography. All authors approved the final version of the 16. Newman R, Vilardell N, Clavé P, Speyer R. Effect of bolus viscosity on the safety and efficacy of swallowing and the kinematics of the manuscript, including the authorship list. swallow response in patients with oropharyngeal dysphagia: white paper by the European Society for Swallowing Disorders (ESSD). Dysphagia. 2016;31:232‐249. ORCID 17. Barnes HA. Shear‐Thinning Liquids. In: University of Wales, ed. A Handbook of Elementary Rheology, Chapter 9. 1st ed., Aberystwyth: Mireia Bolivar‐Prados https://orcid.org/0000-0002-8896-1150 University of Wales; 2000:55‐60. Viridiana Arreola https://orcid.org/0000-0003-3913-317X 18. Brito‐de la Fuente E, Ekberg O, Gallegos C. Diagnosis and Treatment, Rheological Aspects of Swallowing and Dysphagia. In: Eckberg O, Weslania V. Nascimento https://orcid.org/0000-0003-1267-3475 ed. Dysphagia. 1st ed., London, UK: Elsevier Inc.; 2012:493‐506. BOLIVAR‐PRADOS et al. | 11 of 11 19. Partal P, Franco JM. Non‐Newtonian Fluids. In: Windhab E, ed. 33. ESO and ESSD guideline for the diagnosis and treatment of post‐ Rheology, vol. I. Singapore: Encyclopaedia of Life Support Systems stroke dysphagia. In: 4th European Stroke Organisation Conference, (EOLSS); 2010:96‐119. Gothenburg; 2018. 20. Vilardell N, Rofes L, Arreola V, Speyer R, Clavé P. A comparative 34. Cabib C, Ortega O, Kumru H, et al. Neurorehabilitation strat‐ study between modified starch and xanthan gum thickeners in post‐ egies for poststroke oropharyngeal dysphagia: from compen‐ stroke oropharyngeal dysphagia. Dysphagia. 2016;31(2):169‐179. sation to the recovery of swallowing function. Ann N Y Acad Sci. 21. Rosenbek JC, Robbins JA, Roecker EB, Coyle JL, Wood JL. A pene‐ 2016;1380(1):121‐138. tration‐aspiration scale. Dysphagia. 1996;11:93‐98. 35. Arreola V, Vilardell N, Ortega O, et al. Natural history of swallow 22. Clavé P, Arreola V, Romea M, Medina L, Palomera E, Serra‐Prat function during the three‐month period after stroke. Geriatrics. M. Accuracy of the volume‐viscosity swallow test for clinical 2019;4:1‐13. screening of oropharyngeal dysphagia and aspiration. Clin Nutr. 36. Sporns PB, Muhle P, Hanning U, et al. Atrophy of swallowing mus‐ 2008;27(6):806‐815. cles is associated with severity of dysphagia and age in patients 23. Goldstein LB, Bertels C, Davis JN. Interrater reliability of the nih with acute stroke. J Am Med Dir Assoc. 2017;18:635.e1‐635.e7. stroke scale. Arch Neurol. 1989;46(6):660‐662. 37. Martín A, Ortega O, Roca M, Arús M, Clavé Civit P. Effect of a min‐ 24. Rofes L, Arreola V, Mukherjee R, Clavé P. Sensitivity and spec‐ imal‐massive intervention in hospitalized older patients with oro‐ ificity of the eating assessment tool and the volume‐viscosity pharyngeal dysphagia: a proof of concept study. J Nutr Heal Aging. swallow test for clinical evaluation of oropharyngeal dysphagia. 2018;22(6):739‐747. Neurogastroenterol Motil. 2014;26(9):1256‐1265. 38. Perlman A, Booth B, Grayhack J. Videofluoroscopic predictors of 25. Clavé P, De kraa M, Arreola V, et al. The effect of bolus viscosity aspiration in patients with oropharyngeal dysphagia. Dysphagia. on swallowing function in neurogenic dysphagia. Aliment Pharmacol 1994;95:90‐95. Ther. 2006;24(9):1385‐1394. 26. Hind JA, Nicosia MA, Roecker EB, Carnes ML, Robbins JoAnne. Comparison of effortful and noneffortful swallows in healthy middle‐ aged and older adults. Arch Phys Med Rehabil. 2001;82(12):1661‐1665. S U P P O R T I N G I N FO R M AT I O N 27. Ortega O, Rofes L, Martin A, Arreola V, López I, Clavé P. A compar‐ ative study between two sensory stimulation strategies after two Additional supporting information may be found online in the weeks treatment on older patients with oropharyngeal dysphagia. Supporting Information section at the end of the article.     Dysphagia. 2016;31(5):706‐716. 28. National Dysphagia Diet Task Force. National dysphagia diet: stan‐ dardization for Optimal Care. Chicago, II: Am Diet Assoc; 2011. How to cite this article: Bolivar‐Prados M, Rofes L, Arreola V, 29. Oreagba IA, Usman SO, Olayemi SO, et al. The use of the WHO‐ UMC system for standardized case causality assessment. Uppsala et al. Effect of a gum‐based thickener on the safety of Uppsala Monit Cent. 2014;48(3):194‐203. swallowing in patients with poststroke oropharyngeal 30. Truelsen T, Piechowski‐Jozwiak B, Bonita R, Mathers C, dysphagia. Neurogastroenterol Motil. 2019;31:e13695. https​:// Bogousslavsky J, Boysen G. Stroke incidence and prevalence in doi.org/10.1111/nmo.13695​ Europe: a review of available data. Eur J Neurol. 2006;13:581‐598. 31. Dávalos A, Ricart W, Gonzalez‐Huix F, et al. Effect of malnutrition after acute stroke on clinical outcome. Stroke. 1996;27:1028‐1032. 32. Clavé P. ESSD labelling of thickening agents for patient safety. In: 8th ESSD Congress, 15th Session Shaping the Future session: rehabili‐ tation in Dysphagia; 2018.