Risk of malnutrition and health-related quality of life in community-living elderly men and women: The Tromsø study

The Tromsø Study is a health survey of the population of Tromsø, a medium-sized town in Norway. The 6th cross-sectional survey was conducted between October 2007 and December 2008. All independently living inhabitants aged 65 to 87 years (6,098) were invited, of which 4,017 (66%) participated by going to a study center for data collection. After exclusions (21 persons because height or weight had not been measured due to various disabilities, 412 persons because of missing weight loss information and additionally 298 persons due to lack of response to the EQ-5D health state descriptive system), 3,286 subjects (i.e., 54% of the eligible subjects) were included in the analyses of the EQ-5D-index. The second part of the EQ-5D exercise (the EQ VAS scale) had 1,306 respondents.

The regional board of research ethics approved the survey, and each participant gave written informed consent prior to inclusion in the study.

At the study center, the participants, who were instructed to wear no shoes and light clothing, had their weight (kg) and height (cm) measured to the nearest decimal using a Jenix DS-102 stadiometer (Dong Sahn Jenix Co., Ltd., Seoul, Korea). Body mass index (BMI) was calculated as weight divided by the square of height (kg/m²). The participants were asked in a self-administrated questionnaire whether they had any involuntary weight loss during the last 6 months. If they had, they were asked how many kilograms (kg) had been lost. Weight loss was grouped as follows: below 5%, between 5 and 10% or above 10% of body weight prior to weight loss.

Based on the BMI and degree of weight loss, each subject was categorized into low, medium or high risk of malnutrition according to the malnutrition universal screening tool (‘MUST’) (Fig. 1). The ‘MUST’ tool is one of the nutritional screening instruments recommended by the European Society for Parenteral and Enteral nutrition [11], and it was originally developed by the British Society of Parenteral and Enteral nutrition (www.​bapen.​org.​uk) [12]. It includes an acute disease component with no nutritional intake for >5 days, which normally necessitates hospitalization [12]. As participation in this study required the ability to independently visit a research center, the acute diseases component was set to zero. The weight loss question was slightly modified to state a time span of the “last 6 months”, but this encompasses the time span of “the past 3–6 months” in the original ‘MUST’ tool.

Based on BMI alone, we further categorized the participants into six categories (<20.0, 20.0–22.4, 22.5–24.9, 25.0–27.4, 27.5–29.9 and ≥30 kg/m²), thereby including the WHO (World Health Organization) definitions of overweight (25.0–29.9 kg/m²) and obesity (≥30 kg/m²) [13].

HRQoL was measured by the EQ-5D, which is a standardized non-disease specific instrument consisting of two parts: the EQ-5D descriptive system and the EQ visual analogue scale (EQ VAS) [10]. The EQ-5D has been utilized in a number of studies, and the instrument is validated in acutely ill, elderly individuals [14] and community-living elderly women receiving medication (clodronate) for osteoporosis [15], but not in large populations of community-living elderly men and women. In a systematic literature review of self-assessed health instruments [16], the EQ-5D was one of the recommended generic health instruments for use in older people.

EQ-5D describes health in generic terms using five specific dimensions, which are important for elderly individuals: mobility, self-care, usual activities, pain/discomfort and anxiety/depression. Each dimension is divided into three levels of severity (no problems, some problems or extreme problems). Due to only a few participants reporting problems at the most severe level (extreme problems), this category was included with the individuals reporting some problems (second level) in the analyses of the various EQ-5D dimensions (Table 2). The EQ-5D instrument is designed for self-completion and was included as part of a self-administrated questionnaire. A single summary EQ-5D index with a maximum score of 1 is obtained by applying a scoring algorithm that assigns weights to each of the possible combinations of health, as described by the three levels within each of the five dimensions. In the current study, we applied the most widely used scoring algorithm, referred to as the UK time-trade-off tariff [17]. Subjects missing values from any of the five dimensions were excluded from the analyses.

In addition to this indirect health index assigned through a descriptive system, a direct method was used asking subjects to rate their health on a visual analogue scale (VAS) with a maximum score of 100. The endpoints were labeled as “worst imaginable health state” and “best imaginable health state.”

The number responding to the EQ VAS scale question (669 men and 637 women) was lower than the number responding to the EQ-5D health state descriptive system. When comparing responders to non-responders on the EQ VAS scale, both groups had a mean EQ-5D health index of 0.82. In the responders, the mean BMI and age were somewhat lower (0.4 kg/m² and 0.5 years, respectively). Although minor, these differences were statistically significant (BMI P = 0.02 and age P = 0.005).

Information on socio-demographics and smoking status (Table 1) was also obtained from self-administrated questionnaires.

We stratified the analyses when dealing with risk categories of malnutrition and HRQoL by gender (Tables 2, 3; Figs. 2, 3). However, for the relation between BMI and HRQoL, tests for interaction between BMI and sex were found to be non-significant both for the EQ-5D index (P = 0.57) and the EQ VAS score (P = 0.37). The combined results for men and women are presented (Figs. 4, 5). Analysis of covariance was used to obtain age-adjusted mean values with corresponding 95% confidence intervals for different risk categories of malnutrition and categories of BMI (Figs. 2, 3, 4, 5). Differences between groups were evaluated by the chi-square test or t-test when appropriate (Tables 1, 2). Binary logistic regression with adjustment for age was used to estimate the association between the risk category of malnutrition and the various EQ-5D dimensions (Table 3). The association between increasing risk of malnutrition and the EQ-5D- or EQ VAS score was assessed by linear regression in a multivariate model adjusting for age. We tested for a quadratic relation between BMI categories and HRQoL by also including the BMI categories squared in a multivariate linear regression model (Figs. 4, 5).

The importance of the differences in HRQoL scores between risk groups of malnutrition may be examined by calculating their effect size as the mean difference divided by the standard deviation (SD) of the control group [18]. We evaluated the detected differences against the criteria introduced by Cohen [19] using the SD of the low-risk category of malnutrition. Effect size values of 0.2–<0.5, 0.5–<0.8 and ≥0.8 were characterized as small, medium and large differences, respectively.