Study on Drug-Food Thickening Agents’ Interactions between Warfarin and Prebiotics Used for Viscosity Adjustment

Oropharyngeal dysphagia (OD), a disorder of difficulty in swallowing, affects approximately 8% of the global population [1]. OD is even more prevalent among older adults, affecting approximately 37% of those living independently [2, 3] and 50–68% of those in care facilities [4‒7]. Thickeners are widely used to assist with swallowing during meals and medication intake for patients with OD, with thickeners used in 83.4% of long-term care facilities and in 22.5% of residents [8]. However, there have been cases of thickener use where tablets were excreted in feces without dissolving [9]. Moreover, decreased disintegration and dissolution of medications with thickener usage have been reported [10, 11], suggesting that medications may adhere to dietary fibers, leading to reduced medication absorption [12, 13]. Currently, the primary thickening agent used in the medical field is xanthan gum. In Japan, xanthan gum is considered a soluble dietary fiber, and due to its characteristics of being difficult to digest, it is expected to retain water in the intestines and thicken to improve bowel movements, among other effects. Of note, definitions differ by country; in the USA and Europe, xanthan gum is classified as a food additive, not a dietary fiber. Compared to starch- and guar gum-based thickeners, xanthan gum-based thickeners offer several advantages, including easy viscosity adjustment, high transparency when dissolved (preserving the food’s appearance), and no impact on food flavor. Therefore, it is widely used not only for thickening but also for enhancing food texture. Guar gum and pectin, both soluble dietary fibers, are also commonly used as thickeners. Some soluble fibers, such as oligosaccharides and inulin, are recognized as prebiotics that selectively promote the growth and activation of the host’s beneficial gut flora, thereby improving health [14]. Prebiotics promote gut health, regulate blood sugar levels, and help manage obesity and hypertension [15, 16], making them popular in various dietary supplements. Functional foods, which are specifically designed to enhance nutrient content for specific effects, have become part of everyday life. For instance, foods fortified with dietary fiber that promote bowel movements or support gut flora are commercially available. Consequently, there is a possibility that these foods may be unknowingly consumed alongside medications, leading to potential interactions.

A systematic review investigating the effects of thickeners on oral medications noted that there is an overlap between patients with dysphagia and prescribed multiple medications and those receiving a concentrated liquid diet, emphasizing the need to understand the interaction between concentrated liquid diet and medications. The study also pointed out the need to consider the relationships among the thickener’s main components, concentration, and contact time, along with the influence of pH and digestive enzymes. The review concluded that extra caution is required when administering medications that are difficult to control, such as warfarin and digoxin [17]. We previously encountered a case wherein controlling the prothrombin time-international normalized ratio (PT-INR) of a patient using a thickener while taking warfarin was challenging [18]. Although the thickener was likely a contributing factor, to the best of our knowledge, no previous reports had examined the impact of combining warfarin with thickeners on drug blood levels or clinical laboratory values. Given that thickeners are mixed with medications during administration, understanding their interaction and potential effects on drug efficacy is clinically important. Thus, this study aimed to evaluate the drug-food interactions between warfarin and prebiotics through an exploratory investigation with clinical and basic research studies.

This study focused on patients who were hospitalized at Showa University Hospital between January 2018 and December 2022, took warfarin, used a thickener, and had available data on blood coagulation before and after the combined use of warfarin and a thickener. Of 1,350 patients using warfarin, the PT-INR was measured in 967 patients, and a thickener was used in 33 patients. After excluding those with renal impairment, serum albumin levels <2.5 g/dL or unmeasured, and warfarin usage of ≤5 days, the final study population comprised 14 patients (online suppl. Fig. S1; for all online suppl. material, see https://doi.org/10.1159/000543532).

We retrospectively collected data on age, sex, height, weight, purpose of warfarin use, dosage, concomitant medications, interval (in days) from the start of warfarin to the combination with the thickener, and laboratory findings from the medical records. Data were collected before and after thickener initiation. The precombination value was defined as the measurement taken on the day the thickener was started (or the value immediately prior if not measured on that day), and the postcombination value was defined as the first measurement taken at least 48 h after starting the thickener. Additionally, the Warfarin Sensitivity Index (WSI; PT-INR/warfarin dose [mg/day/m² of body surface area]) [19] was calculated as an indicator of warfarin dose responsiveness.

The pre- and postcombination laboratory data and WSI were compared using the Wilcoxon signed-rank test. Statistical analysis was performed using JMP^® Pro 17.0.0 (SAS Institute, Cary, NC, USA), with significance set at p < 0.05.

Warfarin potassium was obtained from the Pharmaceutical and Medical Device Regulatory Science Society of Japan. Xanthan gum and carboxymethyl cellulose sodium salt (high viscosity) were purchased from MP Biomedicals. Carboxymethyl cellulose sodium salt, low viscosity, and inulin (by enzyme synthesis) were purchased from Tokyo Chemical Industry Co., Ltd. Indigestible dextrin was purchased from Crane Foods Co., Ltd. Guar gum, pectin (from apple), chondroitin sulfate C sodium, sodium alginate (low polymerization and high polymerization), agar, galactooligosaccharides, fructooligosaccharides, λ-carrageenan, and ultrapure water were purchased from Fujifilm Wako Pure Chemical Corporation. All other reagents were of analytical grade. For artificial gastrointestinal fluids, 3F powder to make fasted state simulated gastric fluid (FaSSGF), fasted state simulated intestinal fluid (FaSSIF), and fed state simulated intestinal fluid were purchased from Biorelevant.

Warfarin potassium was prepared at a concentration of 1 mg/mL, and the prebiotics were prepared at concentrations ranging from 0.5% to 2%. A 1-mL aliquot of the warfarin solution was mixed with 9 mL of the prebiotics solution at 25°C. Immediately (0 min) and 10 min after mixing (10 min), 400 µL of the solution was subjected to ultrafiltration using a Centricut Ultra-Mini W-MO (0.22 µm) (Kurabo Industries, Ltd.) at 5,000 rcf and 4°C for 10 min to obtain the sample. To measure the residual rate of warfarin after mixing with the artificial gastrointestinal fluids, the experiment was designed to mimic clinical medication practices. A 1-mL aliquot of 1 mg/mL warfarin solution was mixed with 9 mL of 0.5% prebiotics solution and allowed to stand at 25°C for 10 min. This 10 mL mixture was then combined with 90 mL each of FaSSGF (pH 1.6), FaSSIF (pH 6.5), or fed state simulated intestinal fluid (pH 5.0) and incubated at 37°C for 2 h. The solution was subsequently subjected to ultrafiltration to obtain the sample.

The samples were analyzed using a Prominence high-performance liquid chromatography (HPLC) system (Shimadzu) equipped with a linear ion trap quadrupole mass spectrometer (QTRAP5500, SCIEX). The HPLC conditions included a Cadenza CD-C18 column (4.6 mm × 75 mm, 3 µm particle size) (inTakt Corporation) as the stationary phase. The mobile phase consisted of the following: A, 100 mm ammonium formate; B, acetonitrile: 100 mm ammonium formate (9:1). The flow rate was set to a gradient of 20–100% B (0–15 min). A gradient elution was performed at 0.3 mL/min. Consequently, 5 µL of the sample was injected into the HPLC system. The warfarin concentration was determined from the peak area obtained using the liquid chromatography-electrospray ionization-tandem mass spectrometry system, and the warfarin residual rate was calculated based on the ratio of the warfarin concentration in the sample solution to that in the warfarin control solution. The warfarin residual rates were compared using the unpaired t test. Statistical analysis was performed using JMP^® Pro 17.0.0 (SAS Institute, Cary, NC, USA), with significance set at p < 0.05.

The final analysis included 14 patients (Table 1). In all 14 cases, a product containing 0.642 g of xanthan gum-based thickener per 2 g packet was used. The medication was typically administered with water prepared by dissolving one packet of the thickener product in a cup of water. However, the medical records did not provide information on the exact amount ingested. The median (range) age was 77.5 years (39–85 years). The most common indication for warfarin use was after cardiac valve or vascular graft surgery in 10 patients (71.4%), followed by atrial fibrillation in 3 patients (21.4%). The number of days from warfarin initiation to combined thickener usage was ≥2 weeks for 8 patients, ≥1 week for 5 patients, and 6 days for 1 patient. The median (range) duration for measuring postcombination test values after starting the thickener was 3 days (2–6 days). No pre- to postcombination changes were observed in the aspartate transferase, alanine transaminase, or estimated glomerular filtration rate. However, total protein and albumin levels were significantly increased, while C-reactive protein levels were significantly decreased postcombination. Although no significant difference was found in the PT-INR or WSI, the WSI was decreased postcombination in approximately half of the patients (Fig. 1). Regarding the dosage form, among the eight cases with decreased WSI, four involved tablets and four involved fine granules. Conversely, among the six cases where WSI did not decrease, only one involved fine granules. Overall, 13 of the 14 patients were using medications that may have intensified the effects of warfarin, but none of the patients started additional medications that could weaken the effects of warfarin during the study period. No concomitant medications were changed or discontinued.

The calibration curve for warfarin created using the peak areas measured by the liquid chromatography-electrospray ionization-tandem mass spectrometry system showed linearity from 1 × 10⁻⁸ mg/mL to 1 × 10⁻⁴ mg/mL, with a detection limit of 1 × 10⁻⁸ mg/mL (blank + 2 standard deviations). The intraday variability of the calibration curve was 3.7–16.0% (coefficient of variation, n = 7). Accordingly, the sample solutions were diluted to achieve a warfarin concentration of 1 × 10⁻⁵ mg/mL for analysis.

Among the 14 prebiotics, we found a significant decrease in the warfarin residual rate for xanthan gum (0 min: 62.9 ± 3.6%, 10 min: 56.5 ± 4.3%), guar gum (0 min: 66.9 ± 4.5%, 10 min: 62.6 ± 1.6%), and pectin (0 min: 50.7 ± 3.9%, 10 min: 54.7 ± 4.1%). The residual rates for xanthan gum and guar gum exhibited time-dependent decreases. Subsequently, under varying concentrations of xanthan gum, guar gum, and pectin, the warfarin residual rates decreased as the prebiotic concentrations increased. After mixing these three prebiotics with warfarin and reacting the solutions with artificial gastrointestinal fluids, xanthan gum and guar gum had the lowest residual rates with FaSSGF, while pectin showed the highest residual rate with FaSSGF (Fig. 2).

The response to warfarin varies between individuals, and the PT-INR can substantially change depending on the patient’s condition and concomitant medications. From our past experiences, we noted the potential for warfarin and prebiotics to affect the PT-INR, prompting us to conduct this study. In the clinical research study of 14 patients, no significant differences were observed pre- to postcombination in the anticoagulation indicators of PT-INR and WSI. This may be attributed to the effective control of the PT-INR through warfarin dosage adjustment. However, the WSI was decreased postcombination in approximately half of the patients, indicating that there were patients with reduced warfarin dose responsiveness. As none of the concomitant medications are known to diminish the effects of warfarin, the thickeners may have contributed to the decrease in the WSI. Interestingly, patients with reduced WSI were more likely to use fine granules compared to those without reduced WSI. When administering medications to patients with dysphagia, healthcare providers often crush tablets and mix them with thickened water or enteral feedings, which may lead to interactions during the preparation stage. Since powdered medications have a larger surface area than tablets, they have greater contact with the thickening agent, making them more prone to interactions and potentially reducing absorption. A limitation of the clinical research study is its small sample size. Given that our hospital is an acute care facility, only few patients had conditions stable enough for monitoring, resulting in a small number of participants for analysis. Additionally, this was a single-center, retrospective study; hence, it did not sufficiently address individual factors affecting the PT-INR, such as changes in patient condition or fluctuations in albumin levels. Moreover, genetic polymorphisms related to warfarin metabolism, common among the Asian population, were not investigated. Future large-scale studies on patients with stable PT-INR control on long-term warfarin therapy are needed to expand upon our findings.

In the basic research study, among the multiple prebiotics tested, the warfarin residual rate decreased with xanthan gum, guar gum, and pectin in a concentration-dependent manner. Many thickeners primarily contain xanthan gum or guar gum. To simulate the in vivo dynamics after oral administration, the warfarin and prebiotics mixture was reacted with artificial gastrointestinal fluids to measure the warfarin residual rate. The prebiotics concentration matched that of the thickened water and was mixed with the artificial gastrointestinal fluid at a ratio deemed physiologically appropriate. Xanthan gum and guar gum exhibited the highest warfarin residual rates in FaSSIF with a neutral to mildly acidic pH and the lowest residual rates in FaSSGF with a strongly acidic pH. In contrast, pectin showed a higher residual rate in FaSSGF. Warfarin potassium is a low-molecular-weight compound (346.42 g/mol) and exists in the form of warfarin ions in aqueous solutions. The ratio of molecular to ionic forms changes with the pH of the solution, increasing the molecular form in acidic conditions and the ionic form in alkaline conditions. This change may lead to chemical or physical interactions with prebiotics.

Both guar gum and xanthan gum have high-molecular-weight structures. Xanthan gum maintains a stable structure under pH 2–12 conditions [20]. Therefore, the decrease in the warfarin residual rate in FaSSGF is attributable to the molecular form of warfarin entering the complex high-molecular-weight structure. In contrast, the residual rate of warfarin was approximately 20% higher in the more neutral FaSSIF. The increase in the ionic form due to the high pH of FaSSIF may have led to chemical bonding. Although this has not been fully elucidated, using thickeners while fasting rather than after meals can have a smaller impact on drug efficacy when taking warfarin. Regarding pectin, low-methoxy pectin (with a methoxylation rate of 8–12%) was used in this study. Pectin hydrolyzes at a pH of ≤3 [21, 22], which suggests that interactions may have been disrupted in the acidic FaSSGF, leading to a higher warfarin residual rate. Furthermore, because pectin has a linear structure, physical effects are unlikely, with chemical bonding (adsorption) between functional groups potentially being the main mechanism of interaction. For patients using gastric acid suppression agents, a higher gastric pH could prevent the hydrolysis of pectin, potentially affecting drug efficacy.

A limitation of the basic research study is that it focused on warfarin in its powder form and did not consider other formulations. In general, the absorption rate of warfarin is not affected by the timing of meals, and it has almost 100% bioavailability following oral administration. However, there have been reports where tablets taken with thickeners did not disintegrate and were excreted as whole or where dissolution varied depending on the tablet type [13, 23]. Thus, it is possible that the residual rate of warfarin tablets could be further reduced, potentially leading to a decrease in the PT-INR. This finding is particularly interesting because it contradicts the clinical trial results. The influence of dosage form on the strength of the interaction has been noted and needs further clarification for different warfarin formulations. Another limitation is that the reactions were conducted under separate conditions for gastric and intestinal fluids. Given that warfarin is absorbed in the stomach and upper small intestine, examining changes in warfarin residual rates by sequentially reacting with gastric fluid followed by intestinal fluid would be necessary to better reflect real-world conditions.

The three prebiotics that demonstrate a decrease in warfarin residual rates are classified as plant-based water-soluble dietary fibers. Xanthan gum is derived from microorganisms, while guar gum and pectin are derived from cell walls. Other plant-based prebiotics, such as sodium alginate, carrageenan, and agar from seaweed, as well as inulin from roots, did not affect the warfarin residual rates. In addition, animal-derived prebiotics (e.g., chondroitin sulfate) and those classified as oligosaccharides (e.g., indigestible dextrin, galactooligosaccharides, and fructooligosaccharides) and synthetic or modified polysaccharides (e.g., carboxymethyl cellulose) did not influence the warfarin residual rates. Since these findings suggest that the origin and structure of prebiotics may play a role in drug interactions, further investigation is needed to fully understand these mechanisms.

In conclusion, the use of thickeners containing xanthan gum, guar gum, or pectin during warfarin therapy may weaken the effects of warfarin due to drug-food interactions. Prebiotic products are classified as foods, and thus, research on their interactions with pharmaceuticals is limited. Accordingly, additional studies are needed as it is important to provide clinicians and patients with accurate information on the safe use of prebiotic products and pharmaceuticals to ensure patient well-being through appropriate drug delivery.

We acknowledge Y. Hirade, K. Onuki, and K. Kitami for their assistance in the LC-MS analysis. We would also like to thank M. Kaneki, Y. Ozeki, T. Michikawa, T. Ohtsuka, S. Sugisawa, and Prof. H. Arakawa for the great discussion.

Information about the aim of the study was posted on the Showa University website, and an opt-out informed consent protocol was used for the use of participant data for research purposes. The study protocol and the consent procedure were reviewed and approved by the Showa University Ethics Committee (Approval No. 21-181-A, date of decision August 31, 2022).

The authors have no conflicts of interest to declare.

This work was supported by JSPS KAKENHI Grant No. JP22K11862. The funder had no role in the design, data collection, data analysis, and reporting of this study.

H.W. and K.K. collected and analyzed the data and wrote the manuscript. M.O. collected and analyzed the data. K.T. and M.C. contributed to the review and editing of the manuscript and provided resources. K.K. conceived the study and obtained funding. All authors contributed to the conception and design of the work, interpretation of the data, review and editing of the manuscript, and approval of the final version of the manuscript. All authors agreed to be accountable for all aspects of the work. K.K. is the guarantor of this work and, as such, has full access to all the data in the study.

Data are not available due to ethical reasons. Pseudonymized data can be acquired from the corresponding author upon reasonable request.