Modified medication use in dysphagia: the effect of thickener on drug bioavailability—a systematic review

Key summary points Aim To understand further the effect of thickener on medications and their pharmacokinetic and therapeutic profiles, using a literature search. Findings Despite dysphagia and polypharmacy being common in older adults, little is known about the effects of altering liquid viscosity on the pharmacokinetics and therapeutic effect of most medications. Message Further work must be undertaken to support clinicians, pharmacists and patients in understanding the effects of thickener. Supplementary Information The online version contains supplementary material available at 10.1007/s41999-023-00896-6.


Introduction
As adults get older there is an increase in the prevalence of polypharmacy [1].In England, more than 1 in 10 people over 65 years take at least eight different prescribed medications each week.This increases as people get older-to nearly 1 in 4 in the over 85's [2].
Dysphagia is a geriatric syndrome associated with many common long-term clinical conditions including stroke, dementia, Parkinson's disease and frailty.Clave et al. report that oropharyngeal dysphagia may affect 30-40% of the population aged 65 years or more.They also report that dysphagia affects more than 50% of people living in nursing homes [3].In hospitalised older adults, dysphagia rates can vary between 7.9 and 86% [4].Thickening fluids is an increasingly used strategy to reduce the risk of aspiration.The increased viscosity of fluids compensates for a swallowing deficit by slowing down the flow of the fluid from the mouth to the oropharynx-allowing time for glottis closure [5].Steele et al. performed a systematic review on the effect of food texture and liquid consistency modification on swallowing physiology and function, they concluded that there is a clear reduction in the risk of aspiration as liquids progress from thin to very thick (studies involving barium swallow under video fluoroscopy in adults with dysphagia) [6].Evidence has been found to show people with dysphagia are three times more likely to experience medication administration errors than those without dysphagia located on the same ward [7]; that is if they are diagnosed at all as often patients with dysphagia are undertreated and underdiagnosed in many medical centres [8].
Given the overlapping profiles of those patients prescribed multiple medications and those receiving thickened fluids, understanding the interactions between thickened fluids and medications and how this may affect medication bioavailability and therapeutic effect is essential.
Thickening agents can be broadly classified into 3 groups: starch based, guar gum based and xanthan gum based, often chosen based on their cost, availability, consistency and flavour.There is variability in how thick fluids are that patients receive.In 2012 the international dysphagia diet standardisation initiative was founded, and the characteristics of fluids, thickened fluids and foods were described.The level of viscosity of fluids ranges from 0 (thin) to 4 (extremely thick) [9].
There have been a few studies reporting the effect of thickeners on drugs, but no systematic review has been performed.We have conducted a systematic review to answer the question of whether thickened fluids affect the bioavailability of oral medications.Furthermore, the level of thickness of fluids, the type of thickener used and the duration of emersion in thickened fluid affects the bioavailability of medications.Additionally, we have reviewed evidence as to whether the bioavailability and therapeutic effect of various medications are differentially affected by thickener.

Methodology
PRISMA methodology was followed to structure the review [10] (see Fig. 1).An experienced librarian performed a literature search of MEDLINE and EMBASE.Search terms included: dysphagia AND bioavailability OR absorption of medicines OR pharmacokinetics AND (with either) Parkinson's/antibiotic /epilepsy/antiviral/thickened diet (EMBASE only).Two of the authors screened search results (titles and abstracts, duplicates were removed at this stage).No software was used in this process.The terms 'Parkinson's', 'antibiotic' 'epilepsy' and 'antiviral' were included so that we could focus our search on time-critical medications and medications that have a critical impact on clinical outcomes.However, when evaluating search results, we have included all medications due to the small number of studies in this field.We excluded unrelated articles, review articles and conference abstracts.We also excluded animal studies and articles relating to nasogastric/percutaneous endoscopic gastrostomy feeds.All articles included were peerreviewed published articles.All ages over 18 years were included in both community and hospital settings.No articles were excluded based on the type of medication.Articles describing in vitro experiments were also included.Articles selected for a full review were selected independently and then agreed upon by both authors.Reference lists from these articles were reviewed and further relevant articles were identified.There was no discrepancy in the selected articles.All relevant articles were then summarised according to study design, number of participants, method and main conclusions.The search was undertaken on 26/02/22 with an additional search on 27/09/23.The two investigators scored the papers using the JADAD scoring system.The JADAD scoring system was used as it allowed easy identification of potential bias in the papers we reviewed [11].

Results
Five hundred seventy results were found in the initial search.A further 23 articles were identified by the authors following reviews of reference lists.Following a review of the abstracts (performed independently by 2 of the authors) 543 publications were rejected.50 received a full text review and a further 32 were rejected, with the remaining 18 articles included in the review.All studies reviewed were small single-centre studies.The rationale for rejecting studies included study design (e.g., animal studies), study type (e.g., review articles and conference abstracts).
Sixteen papers published between 1984 and 2021 were reviewed (outlined in Table 1).Five studies were in vivo and the remaining eleven studies were in vitro.
The in vitro studies mostly evaluated the disintegration and dissolution times of various medicines, comparing control medicines to medicines that had been immersed in thickened fluid for various periods of time.The number of medicines studies ranged from 1 to 45. Endpoints in the in vivo experiments included serum/urine concentrations of medication and therapeutic effect (e.g., serum blood glucose levels when evaluating the effect of thickener on diabetic medications).

The differential effects of type of thickener in vitro
In 2014 Manrique et al. [12] compared 5 commercial thickening agents at 3 thicknesses (mild level, moderate level and extremely thick level) by assessing their effect on the dissolution of atenolol.At 'mild' thickness level no significant delay in the dissolution of atenolol was observed compared to the control.At 'moderate thickness' there was variability between brands on dissolution rate and products based on xanthan gum were noted to delay dissolution.Extremely thickened fluids retarded dissolution rates [12].In 2020 Takahashi compared fluoroquinolone film-coated tablets mixed with starch-based food thickener and xanthan gum-based food thickener-finding that starch-based thickener resulted in a more 'evenly dispersed' product [13].Matsuo's study in 2020 also bore out  the differential between thickeners by comparing guar gum to xanthan gum finding that emersion in guar gum increased disintegration time of magnesium oxide tablets although the study was powered to compare thickeners to the control making the significance unclear.Furthermore, this difference was only noted in some, but not all brands, of magnesium oxide tablets [14].

The differential effects of thickened fluid viscosity
In 2018 Tomita found that magnesium oxide tablets that had been immersed in higher concentrations (0.6 or 0.8% w/v) of xanthan gum thickener took longer to disintegrate in water compared to tablets immersed in 0.2 and 0.4% w/v [15].The effect of thickener concentration was also demonstrated in Manrique's study in 2014 (as noted above) [12].

The effect of the duration of time emersed in thickened fluid
Tomita's study in 2017 demonstrated that Glufast tablets immersed in xanthan-gum thickener for 1 min had a shorter disintegration time compared to tablets immersed in thickener for 10 min however the significance of this was unclear as the study was powered to compare both groups to the control (i.e., not immersed in thickener) [16].In 2018 the same group once again demonstrated this principle, this time with magnesium oxide tablets.They found that tablets immersed in 0.6% thickener for 10 min had a significantly longer dissolution time than those immersed for 5 min.Similar results were born out for tablets immersed in 0.8% thickener [15].In 2020 Matsuo's group noted that immersion of magnesium oxide tablets for 10 and 30 min in xanthan and guargum-based food thickening agents caused disintegration delay and non-disintegration, however, tablets immersed for 1 min only quickly disintegrated [14].In 2021 when the group were comparing the effect of different solvents, they found that magnesium tablets immersed in a thickener for 30 min had delayed disintegration times compared to 1 min [1].

The differential effect of thickening different types of liquids
Ruis-Picazo's study in 2020 compared the dissolution rates of medications including atenolol, candesartan and valsartan.They investigated whether the pH of the thickened fluid affected the dissolution rate.They found that thickener did not impact the dissolution rate of atenolol at pH1.2 but had a marked effect on dissolution rate observed at pH 4.5 and 6.8 compared to control.In the case of candesartan, thickener slowed dissolution rates at all pH levels tested.Similar results were observed for valsartan with the exception of pH 1.2 [18].Matsuo's study in 2020 showed that sports drinks (pH 3.8) and apple flavoured thickened drinks (pH 3.5) had less of an effect on the disintegration time of tablets than roasted green tea (pH 6.5) and green tea flavoured thickened drinks (pH 6.6) [19].Conversely in 2021 Matsuo compared the disintegration times of 4 medications (magnesium, senna, furosemide and aspirin) in 12 different solvents (including water, tea, juice and coffee) and concluded that different solvents did not differentially affect disintegration rates [17].

The differential effect of thickened fluid on various medications
In 2020 Matsuo's group investigated the disintegration time of 40 different tablets comparing non-immersed tablets to those immersed in 1% xanthan gum-based thickener, 3% xanthan gum-based thickener and a commercially available thickened drink for 1 min.They found that the medication being tested did differentially impact the effect of thickener on the disintegration rate [15].It was not clear whether a particular property of some tablets contributed to this.Brave-Jose's observational study in 2022 bore out similar results.They investigated 45 medicinal products and used a starch-based thickener.They found that some drugs were incompatible with thickener and observed the formation of clumps that did not dissolve.Specifically, they noted this for almagate, ibuprofen and macrogol [20].
Sugiura et al. [21] compared the disintegration time of control tablets with tablets that had been immersed in 0.9% w/v xanthan gum thickener for 1 min.Disintegration time did significantly increase for all 7 tablets when immersed in thickener.There was only a weak correlation between the duration of tablets control disintegration and duration after immersed in thickener though (r = 0.431).

The effect of thickener on the therapeutic efficacy of some medications
In the first study of its kind, in 1984, Hupponen et al. conducted a double-blind study whereby volunteers received either digoxin or penicillin medication taken with either guar gum or a placebo preparation (wheat germ) mixed with water [22].Serum and urine concentrations of digoxin were measured by radioimmunoassay.Serum penicillin concentrations were measured microbiologically according to Reeves et al.They found that the mean peak digoxin concentration was lower when the medication was taken with guar gum compared to placebo granules (0.05 < p < 0.1) but similar amounts of digoxin were recovered by 24 h in urine samples.A significantly lower mean peak penicillin concentration (p < 0.01) was obtained on guar gum compared to placebo.There was no difference in time to reach peak penicillin concentration.24 h urine samples were not available for penicillin.It should be noted, however, that in this study guar gum was not being considered for its properties as a thickening agent (moreover as a treatment for diabetes) and thus it is not clear if the viscosity of the guar gum is similar to what is used in clinical practice for thickening.Furthermore, this was a small study (n = 10) only looking at two medications.
In 2015 Tomita et al. conducted an observational study on patients who were admitted to hospital wards during a 14 day period-who were all taking magnesium oxide (n = 147).They found that 21 patients were observed to be using thickening agents.Whether they disintegrated magnesium oxide into a small amount of water prior to adding thickener was observed, and the dose taken was observed.Whether they took additional laxatives was also observed.The mean dose of magnesium was 1705 mg/day in those who took a thickened fluid without disintegrating it first, 1122 mg/day in those who disintegrated into water and then added thickener.The average dose in those who took it with thin fluids was 1380 mg/day.25% of patients who did not use thickener took an additional laxative.33% who took magnesium with thickened fluids without disintegration, and 60% of those who took magnesium with thickened fluids after disintegration, used another laxative [23].This study did not qualify whether thickener significantly affected the amount of magnesium used, or whether there was a statistically significant difference in the requirement for a second laxative.It is also not clear whether other clinical conditions that may have affected laxative use were accounted for.
In 2017 Tomita et al. performed a single-centre prospective cross-over study they compared the effect of immersion in xanthan gum Vs control on the therapeutic effect of mitglinide tablets (N = 5).They found a significant difference in blood glucose levels from administration to 90 min.The greatest difference in blood glucose levels was at 60 min (123 mg/dL for immersed tablets Vs 82 g/dL for non-immersed tablets.p < 0.01) [1].This study corroborated findings from one the group did in 2016 assessing whether immersion in thickener impacted the therapeutic efficacy of another diabetic medication (voglibose) which once again demonstrated a significant difference in blood glucose levels between control and test group (immersed in 3% thickener for 10 min, N = 9) [24].
Interestingly in 2019, Tomita's group demonstrated, like in vitro, the in vivo effects of immersion in thickener may differentially affect the therapeutic efficacy of different medications.In a single centre cross-over design study, they compared the time taken to reach a maximum systemic concentration of levofloxacin after emersion of tablets in thickener, jelly water and control.They found no significant difference in time taken to maximum system concentration and no significant difference in maximum systemic concentration [2].

Discussion
There is an expanding field of in-vitro and in-vivo studies evaluating the potential therapeutic impact of thickeners on the medications that they are taken with.Variables such as the type of thickener (starch based, xanthan gum based or guar gum based), the viscosity of the thickened fluid and the duration tablets are immersed in thickener seem relevant.Several in-vitro studies have demonstrated that these variables significantly affect outcome measures including disintegration and dissolution rates.Whether these variables would significantly impact therapeutic outcomes in-vivo remains to be seen, although two out of three single-centre studies have demonstrated that the general concept of immersion in thickener does affect therapeutic outcomes [16,24].It should be noted however that all participants in these studies were healthy and so what the therapeutic impact on glycaemic control in a diabetic patient taking either mitglinide or voglibose with thickener in the long term remains to be seen.
The type of liquid used with thickener has demonstrated differing significance in various studies in vitro.Given the variation of outcomes in these experiments, and the potential clinical relevance (noting changing pH through the digestive tract) further investigation in this field may prove relevant.It may aid understanding of both the significance of the type of liquid that thickener is added to and the relationship between in-vitro and in-vivo experiments-where the pH of the liquid used is recorded and is comparable to that of the digestive tract.It may also be important to consider whether enzymes within the digestive tract are significant in terms of whether thickener effects therapeutic outcomes.
Both in-vitro and in-vivo studies seem to show that medications (and their therapeutic effect) may be differently affected by the use of thickener.Two in-vitro studies each reviewing 40 + medications have shown that immersion in thickener differentially effects disintegration rates and that several medicines form clumps that do not dissolve when immersed in thickener [19,20].The differential effect of medication choice on whether therapeutic impact is likely to be affected by thickener in vivo is also borne out by comparing the studies of Tomita et al. [16,24,25].They demonstrated in two small single-centre studies that the therapeutic effect of two diabetes medications is impacted by immersion of tablets in thickener, whereas the systemic concentration of the antibiotic levofloxacin (and thus assumedly its therapeutic impact) is not affected.
Analogous to the studies reviewed here Wright et al. investigated the effect of a novel gel-based swallowing aid on systemic salicylate levels and platelet function when delivered with 300 mg aspirin tablets in a cross-over study of twelve healthy volunteers.This gel formation had been developed to ease the administration of tablets to patients with dysphagia.They found that when aspirin tablets were administered with the gel formation systemic salicylate levels were reduced and platelet function was increased.They concluded that bioequivalence could not be assumed [28].
Although not the focus of this review, the shared properties between this gel based swallowing aid and thickener should be considered (e.g.viscosity).On the other hand, although not the focus of this study, it is worth noting that two animal studies have not demonstrated a significant effect of the administration of medications with thickener on their bioavailability.Ilgaz et al. [29] found that in rabbits regardless of thickness level, administration of levetiracetam with thickening agents does not affect the bioavailability of the drug.Gotoh et al. [30] compared the administration of famotidine without food, with yoghurt and with thickener in rats.They found no difference in plasma concentrations in all 3 groups (with levels being taken up to 6 h after administration).
A limitation of this review is that the initial search used only two databases.Search terms included: dysphagia AND bioavailability OR absorption of medicines OR pharmacokinetics AND (with either) Parkinson's/antibiotic/epilepsy/ antiviral/thickened diet (EMBASE only).This was with a view to targeting our review of medicines whereby bioavailability is critical to efficacy.During the initial screening of titles/abstracts it became apparent that research into the effect of thickener on the bioavailability of oral medications is a small field and as such we did not exclude any studies based on oral medications being tested.
To conclude, these initial small single-centre studies suggest that thickener may affect the therapeutic efficacy of at least some medications.The overall clinical relevance of this remains to be seen but may be of particular significance when considering medications such as warfarin, digoxin and some anti-epileptic medications and antibiotics (with narrow therapeutic drug levels).Further larger-scale studies are required to evaluate what the therapeutic impact of thickener is on a much bigger range of medications, factoring in other variables that have been shown to be relevant in vitro including the type of thickener, the viscosity of thickener and duration of immersion.In the meantime, this review highlights the importance of doctors and speech and language therapists to consider carefully the necessity of oral thickener and the route of prescribed medications.

Fig.
Fig. 1 PRISMA flow diagram outlining research methodology