A total of 272 patients underwent FEES for the evaluation of swallowing functions at our hospital between April 1 and December 20, 2021. Of these, 70 patients who provided consent in writing and had data available for evaluation regarding the study diets (rice-flour jelly, rice-porridge, mashed rice-porridge, and fruit jelly) were included in the study.

FEES was performed using an Olympus ENF-V3 and Olympus OTV-S12 (Tokyo, Japan) according to the manual of the Japanese Society for Dysphagia Rehabilitation [7]. The condition of the pharyngeal and laryngeal regions was observed after insertion of the endoscope, followed by examinations of the swallowing ability of the test foods (colored thickened water, mashed rice-porridge, fruit jelly, rice-porridge, rice-flour jelly, chopped diet, rice, and colored water, in that order). The selection of test foods and the order of examination was altered according to the state of swallowing. A volume of 1–5 cm3 of colored thickened water and colored water using a syringe and one teaspoon (about 3 g) of the other test foods was examined. When pharyngeal residue was observed, multiple swallows, alternate swallows, or head rotation was performed to remove the residue before evaluation of the next test food. When aspiration or laryngeal penetration was observed, throat clearing was encouraged by coughing. When aspiration, laryngeal penetration, or pharyngeal residue was not self-cleared, suction was performed, and the examination was finished. The entire process was recorded on a hard disk drive using HDMI/analog capture (I･O DATA, Kanazawa, Japan).

Rice-flour jelly, rice-porridge, mashed rice-porridge, and fruit jelly were used as the study diets. The rice-flour jelly, rice-porridge, and mashed rice-porridge were prepared at the nutrition department of Fukui University Hospital. The rice-porridge and mashed rice-porridge were the same as those provided to inpatients requiring dysphagia diets at Fukui University Hospital. Rice-flour jelly was prepared by mixing 30 g of high-amylose rice-flour with 40 mL water, adding another 275 mL of boiling water, stirring for approximately 30 s, heating in a microwave oven (Toshiba, Tokyo, Japan) at 500 W for 2 min, and then cooling at 4℃ for 2 h [5]. Rice-porridge was prepared by boiling 1 kg of rice and 9 L of water for 70 min using an electric three-dimensional rice cooker CRAE-150 (Comet Kato, Inazawa, Japan) and removing the supernatant liquid after boiling. Mashed rice-porridge was prepared by mashing rice-porridge for 2 min with an MX-153G professional mixer (Panasonic, Tokyo, Japan). For the fruit jelly, we used Procca Zn Grape (Nutri, Yokkaichi, Japan), which has been approved by the Japan Consumer Affairs Agency as a food for people with dysphagia.

The functional oral intake scale (FOIS) was used to rate the subjects’ diet levels [8] (Appendix 1). The pre-test FOIS was the diet level before the FEES examination, which was extracted from the medical records. The post-test FOIS was determined based on the results of the FEES examination by the physician in charge, rehabilitation physician, and speech-language pathologist [8,9,10].

The food texture (firmness, cohesiveness, and adhesiveness) of the rice-flour jelly, rice-porridge, and mashed rice-porridge was measured by NARO using a RHEONERII RE2-33005C creep meter (Yamaden, Tokyo, Japan) in accordance with the test method for diets for people with dysphagia by the Consumer Affairs Agency [11]. The sample was filled into a stainless steel Petri dish 40 mm in diameter and 15 mm high, fixed to a sample stand, and compressed twice at a compression speed of 10 mm/s and a clearance of 5 mm using a plastic plunger 20 mm in diameter and 8 mm high at 24 °C. Because the mashed rice-porridge was in a liquid state with high fluidity, it was not sufficient to evaluate the texture only by measuring firmness, adhesion, and cohesion. Therefore, for the mashed rice-porridge, the viscosity was measured using a DVNext cone-plate rotational viscometer (AMETEK Brookfield Inc., Middleboro, USA). The viscometer commenced rotating at a speed of 50 s−1 at 20 °C, and the viscosity was measured 2 min later [11]. Each food texture metric was measured three times, and the mean and standard deviation values were presented. The texture of the fruit jelly was taken from the literature [12].

First, we extracted and edited the images of swallowing study diets from the recorded FESS images of participants. Using edited images, three raters (one physician with > 15 years of experience and two speech-language pathologists with > 10 years of experience) evaluated the residue volume of the vallecula and pyriform sinuses based on the Yale Pharyngeal Residue Severity Rating Scale (YPRSRS) on five levels: I (none), II (trace), III (mild), IV (moderate), and V (severe) [13, 14] (Appendix 2). When the three raters agreed upon the level, that level was selected; when the three evaluated levels were divided two to one, the level chosen by the majority was selected; and when the three evaluated levels were different, the middle level was selected.

The vallecula and pyriform sinuses residue of the study diets were compared using the χ2 test and residual analysis. The inter-rater reliability was assessed using the kappa statistics. IBM SPSS Statistics version 22 (IBM, Chicago, USA) was used for statistical analysis with a significance level of p < 0.05.

The University of Fukui Ethics Review Committee approved this study (Approval No. 20200179) and participants provided written consent after the study protocol was explained both in writing and orally.