In this prospective study, the protocol established two different age groups, group 1 (50–70 years old) and group 2 (over 70 years old). The study protocol was approved by our hospital’s ethics committee, consent was obtained from all subjects before participation in the study. Twenty-five subjects were included in each group, between January 2021 and December 2022. Subjects in group 1 were healthy volunteers, without any pathology that could impact swallowing. Subjects in group 2 were patients hospitalized in the geriatric department. In the absence of a consensus on the age of presbyphagia, the 70-year age limit was chosen on the basis of data in the literature, and the minimum age required for geriatric hospitalization in our center, which is 70. Exclusion criteria for both groups were the presence of swallowing disorders and pathologies that can influence swallowing (neurological disorders, respiratory condition), history of cervical surgery, cognitive disorders, and allergy to the products used for the tests.

Recordings took place in a quiet room. We used a laryngophone Nauzer® (ref. PLX 300 K) for the acoustic recordings of the pharyngeal phase of swallowing. Two microphones were placed on each side of the neck in front of the trachea, just under the cricoid cartilage (Fig. 1). This is an optimal area for the detection of swallowing sounds [23]. Subjects needed to have a clear neck, no shirt collar, and no jewelry, with hair tied. The laryngophone was connected via an adapter to a tablet Samsung® Galaxy Table 3, used to collect acoustic data. The last food intake had to be at least 3 h before the experimental meal for the study.

Laryngophone in place on a subject’s neck

To avoid a bias related to the dental status of the subjects, the selected meal components did not need to be chewed. The experimental meal was composed of 3 elements with different viscosity: plain water at room temperature, unsweetened yogurt (Danone®), and mashed potatoes (Mouseline®, 125 g in 250 ml of milk and 500 ml of water at 30 °C). The dynamic viscosity measured by the Brookfield method was 1 mPa.s, 300 mPa.s, and 50,000 mPa.s, respectively. The measurement of 100 mL of mashed potatoes, water, and yogurt was done using a measuring glass. The different substances were provided in 3 identical opaque plastic glasses, and two metal spoons were available to eat the mashed potatoes and the yogurt (Fig. 2). This food association (puree-water-yogurt) has already been used in previous works studying the swallowing function [24, 25].

Experimental meal with its 3 components: mashed potatoes, water, and yogurt

The subjects were in a sitting position facing a table, and were asked to eat the experimental meal with the laryngophone in place on the neck: 100 mL of mashed potatoes, 100 mL of water, and then 100 mL of unsweetened yogurt. They were requested to eat at their usual pace. Due to the specificity of the software, which needs to be restarted for each new recording, we have left an identical delay of one minute between each meal component for all subjects.

The laryngophone was connected by an adapter to the tablet. We recorded the subject’s information: name, surname, sex, and date of birth. Then, we manually started and ended the recording for each of the mashed potatoes, water, and yogurt components. We obtained 3 audio files per subject, one per texture tested (mashed potatoes, water, yogurt) in a “.wav” audio file.

All audio recordings were transferred to a computer and analyzed on the Audacity® software, which is a digital audio editing software (Fig. 3). Analyses on this software were performed by two different examiners who had been trained to use the software to analyze acoustic signal of swallowing.

Part of an audio recording on the Audacity® software, showing 2 swallows. S swallow, IS inter-swallow, BS cursor of beginning of swallowing, ES cursor of ending of swallowing

To recognize swallows on audio recordings, the examiners relied on the work of Morinière et al. [14, 26] on the sound components of swallowing. Each recording was treated in the following way: a cursor was placed at the beginning and end of each swallow and the duration of each swallow and inter-swallow was then measured. The swallows identified on the recordings corresponded to the pharyngeal phase of swallowing.

For each subject, different swallowing parameters were studied. The number of swallows (NS) was counted in each audio recording and then summed to obtain the NS of the experimental meal. The duration of each swallow was measured, taking the duration between the cursors at the beginning and at the end of the swallow. For all subjects, the average duration of swallowing (ADS, corresponding the pharyngeal phase of swallowing) for each audio recording, and for the experimental meal, was then calculated. The duration of each inter-swallow was also measured; it corresponded to the duration between the cursor at the end of a swallow and the cursor at the beginning of the next swallow. For all subjects, the average duration of inter-swallowing (ADIS) was then calculated for each audio recording and for the experimental meal. The duration of the meal itself (meal duration: MD) was measured for each meal component, by taking the duration between the cursor at the beginning of the first swallow and the cursor at the end of the last swallow. By adding the sum of the MD of the mashed potatoes, yogurt, and water, we found the MD of the experimental meal (Fig. 4).

Finally, the average swallowing frequency (ASF), corresponding to the average number of swallows per minute, was calculated from the NS and the MD for the experimental meal.

We have thus obtained for all subjects the NS, ADS, ADIS, and MD, for each meal components and for the experimental meal. We also obtained the ASF for the experimental meal. By averaging these parameters for the subjects in each group, the swallowing parameters of group 1 and 2 were calculated and compared.

All time values for this study were measured in milliseconds (ms), and statistical analyses were performed using the software GraphPad Prism version 5. We verified that the collected values were distributed according to a normal distribution with a Shapiro–Wilk test. In case of normal distribution, a Student’s t test was used to compare the data of the two groups. In case of non-normal distribution, a Mann–Whitney test was performed to compare the data of the two groups. The differences were considered significant when p value was less than 0.05, in a two-tailed statistical analysis.