Distributed under Creative Commons Cc-by 4.0 Research Article Nutritional Supplements : Taste Preferences in Patients with Malignant Haematological Disease during Active Treatment

Background: Several studies have indicated that cancer patients have significantly altered taste sensitivity without specifying the preferences. One of the related problems is low compliance to nutritional therapy with oral nutritional supplements (ONS) in patients suffering severe weight loss. Objective: We wanted to investigate taste preferences and sensoric characteristics among three usually used ONS in patients with malignant haematological disease during cytotoxic treatment. Design: Tested drinks were: Protin® (protein-enriched-milk, ARLA), Nutridrink® (NUTRICIA) and hospital-produced drink of buttermilk and egg (RH-drink), all with vanilla taste. Protein-contents were the same (5-6 g/100 ml). Forty-one consecutive patients tasted the three ONS in a randomized, blinded setup

with high-dose chemotherapy and often stem-cell transplantation, changes in taste and smell appear with particular difficulty in differentiating sour and bitter (Epstein et al, 2002). In these hyper metabolic cancer patients, protein-intake is important due to gluconeogenesis, with a loss of lean body mass (LBM). This loss of LBM can potentially be reduced by protein supplements as an integrated part of dietary counselling, but it is only shown in patients with colorectal cancer undergoing radiotherapy . However, poor compliance often results in patients not reaching their nutritional goals (Ravasco, 2005), and part of this are tasterelated problems. Nutritional supplements are developed in an industrial setting and analysis of taste is very often done in a laboratory by a censoring panel. This setting and these censors are quite different from the patients. Patients with gastro-intestinal cancers seem to prefer fresh-milk-based supplements to UHT milk based (Ultra High Temperature) and fruit juice based, also during chemotherapy. However, vitamins and minerals were added to both the UTH-milk based and the fruit based products, and all were hyperosmolar (Rahemtulla et al, 2005). No similar studies have been published for patients with malignant, haematological diseases.
This study aimed to determine the patients' preferences for three different ONS including non-vitamin-enriched and freshly made products. We wanted to examine taste perception, and to assess the reproducibility of the taste assessments by means of a Visual Analogue Scale (VAS) in patients with malignant haematological disease.

VAS (Visual Analogue Scale)
The reproducibility of VAS is relatively well documented in other scientific areas and is described as a valid measure of subjective phenomenon, but carries a phenomenological problem with an underlying acceptance of linearity, which might not be true (Rødbotten, 1997). In this study a unipolar continuous scale was used to evaluate the sensory variables: sweet, sour, bitter, salt, thickness, gritty and metal and the palatability (the ability to drink a glass of the product (150 ml)). All the VAS-scores were performed immediately after tasting, and at the end of the experiment patients rated the three (four) products 1-3(4), as one was repeated.

Selection of Products
Three milk-based products were selected, as these were the three mostly used products in the department: Nutridrik® from Nutricia, Koldskål (RH) from the kitchen in Rigshospitalet and Protin® from Arla. The products had the same flavour, Vanilla and had similar protein content (Nutridrik® 6g/100ml, Protin® 5.7g/100ml and RH 5 g/100ml). Nutridrik® is a UHT milk based, hyperosmolar product from Nutricia with added vitamins and minerals with an osmolarity of 450 mOsm/l (in 100 ml: 630 kJ, 18.4 g carbohydrate, 5.8 g fat). RH is a fresh cold buttermilk based product with eggs made in the Kitchen on Rigshospitalet from day to day. The recipe for one litre is: Junket 3 dl, buttermilk 6 dl, egg-yolk 80 g, sugar 45 g, and the osmolarity is 653 mOsm/l. Junket, egg and buttermilk is whippet, sugar is added and the taste is adjusted with vanilla sugar and lemon (in 100 ml: 356 kJ, 9.3 g carbohydrate, 3.1 g fat). Protin® is a fresh milk based product with added milk-proteins from the diary, Arla with a shelf life of 10 days (open 3 days), osmolarity 718 mOsm/l (in 100 ml: 540 kJ, 12 g carbohydtate, 6.3 g fat)..

Taste samples
The products were served in a plastic cup (4 ml) at room temperature in a random rotating order on the same day. The patients were blinded to the products. The samples were served according to a preformed protocol with random allocation of the drinks, and the repeated drink in a random position in the sequence. One of the samples was repeated in every experiment to quantitate reproducibility.

Viscosity
The viscosity was determined by the Laboratory of Rheology and Texture, The department of Food Technology, The Technical University of Denmark by the shear stress/shear rate (Pa s). Shear stress is the force that moves the liquid, and shear rate (1/s) is the velocity gradient.

Subjects
Forty-one patients from the haematological clinic in Rigshospitalet participated in the study (table 1). Patients were asked for participating consecutively on rounds if they fulfilled the inclusion criteria. Inclusion criteria: patients in treatment in one of the four haematological units, no distinction for cancer type, gender or type of treatment. The time interval from the last chemotherapy or radiation therapy varied from two weeks to two months. All the patients had severe eating problems at the time of investigation, and all had mucositis visually diagnosed. All patients were of legal age and mentally able to understand written and verbal information. Exclusion criteria: use of parenteral nutrition or tube feeding, lack of ability to communicate and/or cooperate. All patients gave written consent.
The protocol was approved by the regional ethical committee.

VAS
The patients' median and average VAS scores for the 3 drinks were similar (table 2)   The taste characteristics are shown in figure 1. The differences in sweet, sour and sensory variables are illustrated. The VAS evaluation of the three products as a hole was not markedly different; their spider web had similar looks. We found no systematic relation between the VAS evaluations and the time passed from the preceding therapeutic session.

The sensory differences of the three products
The results were detailed further (table 4). The patients judged that Nutridrik® had a significant sweeter taste than Protin® (p=0.041) and RH (p=0.021); the patients liked Protin® (p=0.002) and RH (p=0.005) significantly better than Nutridrik®. The patients judged that RH (p=0.008) had a significantly more sour taste than Nutridrik®. No significant differences were found between Protin® and RH with regard to any of the sensory variables.

Reproducibility
Identical (+/-10% in cm) VAS-score was seen in 39 of the 41 patients (95%), when blindly testing the same sample twice in random order mixed-in with the other samples.

Viscosity
The difference between the fresh milk products (Protin® and RH) and the commercial (Nutridrik®) was marked ( Table  5 and fig  2).

Statistics
As results were distributed very skew, nonparametric methods were applied, Wilcoxon-, Mann-Whitney rank-sum tests, and Kruskal-Wallis test.

Discussion
Dysgeusia associated with weight loss due to lack of appetite and altered eating pattern influence the patient's Quality of life. Dysgeusia is caused by many factors, . In our setting, we found no systematic influence of the time-interval from the last chemotherapy or radiation therapy, but all the patients had mucositis. Even though individual differences in taste perception are many and complicated (Stevens, 1996), and despite the complex effects of disease, dissemination of disease, and treatment (Hutton et al, 2007), results seem to have some uniform directions. Fresh, milk based products seem to have a higher acceptability, but agreement is not complete. Several explanations might contribute. We know that hyperosmolarity is a problem for patients with mucositis, but all three drinks in our experiment were considerably hyperosmolar, making this factor an unlikely explanation. Texture was also markedly different ( fig 2). Whether this is a determining factor or not cannot be answered with our present knowledge, but the enriched milk and the buttermilk-drink were very similar, and they were both evaluated as positive by the patients. Nutridrink showed different results, and was evaluated very negative. This may indicate that texture could be part of an explanation. We tried to interview the patients for details related to their own scoring, but gave it up, as most patients felt too tired after the testing sessions. Smell and viscosity as well as other textural variables could be factors involved.
The advantages of the commercial "fullnutritional" drinks are their content of micronutrients, their very long shelf-life, and the control procedures related to the manufacturing. The micronutrients might well be the reason for the reduced palatability, but future studies are needed to elucidate this.
All the published studies about taste disturbances in patients with malignant diseases are single meal tests. In the future, we would need long time studies using the results from the meal-studies to elucidate whether there are positive effects on nutritional status during the course of treatment, as compliance in the long run is determining the clinical value. Weight as well as body composition and quality of life should be used as outcome variables.

Conclusion
This study gives an indication of patients' preference with significant higher preference and palatability for fresh milk products than a commercial, milk based product. However, the basic sensory parameters studied with VAS gave no clear distinction between the three products, and the underlying basis for the preference is not obvious, and merits further investigations. The used test-method was well functioning with a very high reproducibility-rate.