Results and Discussion

Model Fluids

The flow curves for the three edible model fluids are shown in Figure 2. All were designed to have the same viscosity at 50 s−1 of 0.15 Pa s. The Boger fluid should ideally have the same constant shear viscosity as the Newtonian, but the limitation of using edible xanthan instead of for example, PAA makes it slightly shear‐thinning. The amount of added xanthan was optimized to be as low as possible, to maintain constant shear viscosity, while still giving a positive normal stress while sheared. The effect on viscosity of having a dilute lemonade instead of water as solvent was negligible, as the viscosity of the lemonade (viscosity = 1.02 ± 0.003 mPa s) was very close to that of water (1.0016 mPa s). A Boger fluid with PAA instead of xanthan was also prepared and measured to give a reference for a corresponding non‐edible Boger fluid. PAA is more elastic and gives a Boger fluid with a more constant shear viscosity as shown in Figure 2.

FIGURE 2: Viscosity as a function of shear rate for the model fluids studied, Newtonian, Bogerxanthan, shear‐thinning. A Boger fluid with PAA instead of xanthan is shown for reference, . The solid line indicates 50 s−1. Error bars for the standard deviation were smaller than the symbols.

Similar model fluids have previously been prepared for other purposes (Koliandris et al. 2011) as well as for swallowing studies (Nyström et al. 2015). The concentrations and viscosity levels previously reported had the same magnitude, but variations between batches and suppliers make exact comparisons impossible.

All model fluids had IDDSI 2 when measured using the IDDSI syringe test (Steele et al. 2024). The test uses a standardized syringe to measure remaining fluid after 10 s dripping. The rheology of this test has been translated to Power law parameters n and K obtained by viscometry (Lecanu et al. 2024). When plotting n and K in a diagram, the IDDSI levels appear as diagonal bands as shown in Figure 3. All three model fluids fall into the IDDSI = 2 band, even though the low viscosity of the shear‐thinning fluid pushes it toward the limit of IDDSI = 1.

FIGURE 3: IDDSI levels adopted from Lecanu (Lecanu et al. 2024) with the three model fluids inserted: Newtonian, Bogerxanthan, shear‐thinning. The dashed lines are extrapolations from the results by Lecanu. K and n are the Power Law parameters describing a shear‐thinning fluid (n < 1). A Boger fluid with PAA instead of xanthan is shown for reference, . Error bars for the standard deviation were smaller than the symbols.

The fluids were tested as is and no artificial saliva or saliva‐like buffer was added. In the INFOGEST protocol all samples are diluted with equal amounts of saliva buffer (Brodkorb et al. 2019). For boluses of solid food we have previously shown that this ratio is similar to what you find in real boluses (Stading 2021; Stading et al. 2023). For beverages it can be argued that there is less dilution by saliva as the ingestion is faster, but there are no quantitative results published from expectorated boluses.

As an estimation, a semi‐dilute xanthan solution has a viscosity, which is proportional to the concentration squared. A 50:50 dilution will thus reduce the viscosity to 0.52 = 25% of the original viscosity. This would typically mean one step in IDDSI level as can be seen from Figure 3, for example, by comparing the IDDSI level for K and 0.25 K at n = 1. This is a qualitative discussion, but the point is that studying IDDSI level 2 fluids would correspond approximately to IDDSI 3 fluids diluted by saliva. Higher IDDSI levels are rarely relevant for beverages, and lower, that is, IDDSI 1, needs to be investigated to test if mainly their elastic properties are strong enough to contribute. Similarly, although the model fluids cover a significant part of the IDDSI 2 band, other thickeners like starch, and commercially available thickeners should be investigated as well.

A positive normal stress during shear was used as an indication of elasticity in optimization of the Boger fluid. For low viscosity fluids, the normal stress is often overshadowed by inertia, and it is not an absolute measure of fluid elasticity. The normal stress of the Boger and shear‐thinning fluids is shown in Figure 4a. The shear‐thinning xanthan solution had lower shear rate dependence than Boger fluids, which can depend on several physical mechanisms. Xanthan forms a weak network, which easily can disentangle and thus build less elastic stress. The extended, non‐aligned chains at low shear rate mean slower relaxation, and the polymer contribution decreases with shear due to the shear‐thinning (Del Giudice et al. 2017; Zirnsak et al. 1999). In the Boger fluids, maltodextrin masks shear‐thinning.

FIGURE 4: (a) Normal stress as a function of shear rate for the shear‐thinning and Boger fluids, and (b) extensional viscosity as a function of extension rate for the same fluids: Bogerxanthan, shear‐thinning. A Boger fluid with PAA instead of xanthan is shown for reference, . The dashed blue line for the Newtonian fluid indicates 3 × shear viscosity.

Extensional viscosity is a more quantitative measure of fluid elasticity than the normal stress. The shear‐thinning and Boger fluids showed substantial extensional viscosity as shown in Figure 4b, especially compared to 3 × shear viscosity plotted for the Newtonian fluid. A Boger fluid containing PAA instead of xanthan is shown for reference, and it had the highest extensional viscosity due to the high elastic contribution of the PAA. A Trouton ratio could be calculated for the lowest extension rate and was in the range 5–90 for the fluids.

The method of Hyperbolic Contraction Flow determines the total measured stress on a hyperbolic nozzle through which the fluid is forced to extend. The stress is a sum of extensional and shear stresses, and the calculated shear stress is subtracted before calculating the extensional viscosity. For the Boger fluids, this means that only a small extensional component remains; thus, the higher standard deviation for the Boger fluids.

The flow during swallowing of the model fluids was evaluated using a model of the human throat, the “Gothenburg Throat” model (Stading et al. 2019), see Figure 1. The model can mimic healthy as well as impaired swallowing by setting the timing of opening and closing of the epiglottis, airways and the UES. The velocity measurements used in the model have previously been evaluated against video fluoroscopy (Qazi et al. 2020) and the same type of model fluids by a group of dysphagia patients (Nyström et al. 2015). The bolus volume was set to 30 mL which is relevant for drinking beverages although patients with severe dysphagia dysfunction would use smaller boluses (Colevas et al. 2022). The initial speed of the bolus set to 0.3 m/s is within the normal range of typically 0.1–0.5 m/s (Ekberg 2019; Qazi et al. 2019). For impaired swallowing the closing of the epiglottis was delayed 1 s and the airways and UES were open. Figure 5 shows the flow of the swallowed fluid bolus during passage through the pharynx. Videos in slow motion of the bolus flow is found in the Supporting Information and in the Youtube play list https://youtube.com/playlist?list=PLqLiVcF3GKy2Cz3QHIjJB9NvDY2ZwAfid&si=42BmZfqwc51osNNj.

FIGURE 5: Snapshots of boluses passing the pharynx of the Gothenburg Throat model during bolus flow in side view (left) and front view (right). (a) Healthy swallowing of water, (b) impaired swallowing of water, (c) impaired swallowing of the Newtonian fluid, (d) impaired swallowing of the Boger fluid and (e) impaired swallowing of the shear‐thinning fluid. The arrows point to fluid entering the airways. The white moving part is the epiglottis. Videos of the swallowing events are found in the Supporting Information and in the Youtube play list https://youtube.com/playlist?list=PLqLiVcF3GKy2Cz3QHIjJB9NvDY2ZwAfid&si=42BmZfqwc51osNNj.

The snapshots in Figure 5 show the bolus flow when most of the bolus has passed the pharynx, and the most interesting feature is flow into the airways (aspiration). Water is difficult to swallow for a person with dysphagia dysfunction as the flow is turbulent and the bolus splits up in droplets. When comparing Figure 5a,b, there is water in the airways for the impaired swallowing in Figure 5b, but not in 5a. If the fluid is thickened from 1 mPa s of water to 0.15 Pa s with maltodextrin, there is still flow into the airways as shown in Figure 5c, which means that although higher viscosity, the Newtonian fluid still causes aspiration. When thickening the fluid with maltodextrin and xanthan (Boger fluid) there is no flow into the airways as shown in Figure 5d. The Newtonian and Boger fluids have the same shear viscosity, but the Boger fluid is substantially more elastic due to the addition of xanthan. Similarly, Figure 5d shows that the shear‐thinning fluid also prevents flow into the airways. The conclusion so far is that both elasticity and shear‐thinning are helpful to prevent aspiration.

Figure 6 shows the bolus flow of water and the model fluids in more detail. The colors denote the velocity of the bolus at a specific point along a line from the transducer into the pharynx, that is, depth is plotted on the horizontal x‐axis. The transducer was placed just above the entrance to the trachea. Time is on the vertical y‐axis, which means that the front of the bolus is close to the x‐axis and the delayed flow is higher on the y‐axis. The velocity‐time curves are averages of at least three measurements.

FIGURE 6: Velocities of boluses shown by the color scale, of the model fluids along a beamline through the center of the pharynx (see Figure 1), for increasing time of flow. (a) water, (b) Newtonian fluid, (c) Boger fluid, and (d) Shear thinning fluid.

The fastest flow occurs early in time and close to the wall, which also is obvious from the videos of the bolus flow. Figure 6a clearly shows that the water bolus breaks up in parts indicated by the separated peaks, contrary to the other fluids where the bolus is more cohesive.

When comparing the bolus flow of the Newtonian fluid (Figure 6b) with that of water (Figure 6a), the Newtonian fluid is more cohesive and does not break up in droplets. It is still not sufficiently cohesive to prevent flow into the airways. Flow continues close to the wall over the whole measured period, which is in line with the Newtonian character of the fluid.

The Boger fluid (Figure 6c) is spatially more cohesive than the Newtonian bolus although flow continues close to the wall. The small improvement in cohesiveness is, however, sufficient to avoid flow into the airways. The shear‐thinning bolus is even more cohesive and has less flow over time along the wall, which may be explained by the shear‐thinning behavior, which results in a higher apparent viscosity at low shear rates.