Background
Mental health problems are among the most prevalent conditions in elderly people. Anxiety and depression, often seen as co-morbid conditions with overlapping symptoms [
1], are the two most frequent mental health disorders [
2]. Malnutrition is also relatively common in elderly individuals and may be associated with mental health, particularly depression [
3].
While several studies have found mental disorders to be a risk factor for involuntary weight loss/malnutrition in geriatric inpatients and outpatients [
4], little population-based research has been done on the relationship between risk of malnutrition and mental health in this age group. A study from Sweden found depressive symptoms to predict malnutrition in community living elderly [
5], whereas a German study of nursing home residents found no significant difference in the mean malnutrition score between residents with and without depression [
6]. Furthermore, studies of the relationship between body mass index (BMI) and depressive symptoms in elderly individuals have yielded conflicting results. In a study from the US, depression in men was found to be inversely associated with body weight [
7]. A later study of a multiethnic elderly population found an increased risk of depression with increasing BMI, but the most adverse impact of obesity on depression was found in African Americans [
8]. Neither of these studies examined the lower BMI categories in more detail.
In the current study, we therefore aimed to investigate the associations between mental health and both the risk of malnutrition and BMI in a large sample of community-living elderly men and women. We hypothesised that there is a positive relationship between impaired mental health and risk of malnutrition and low BMI.
Discussion
In this study, we found that mental health symptoms were strongly associated with the risk of malnutrition in elderly individuals. Both the risk of malnutrition and mental health symptoms were more prevalent in women than in men. To our knowledge, this is the largest population-based study that explored the relationship between the risk of malnutrition and mental health in elderly individuals.
Some previous studies in this area have utilised the Geriatric Depression Scale (GDS) and the Mini Nutritional Assessment (MNA) instrument for the assessment of the relationship between depression and malnutrition. A Swedish study of 579 community-living elderly people found that depressive symptoms were predictive of malnutrition [
5]; this was observed to a larger extent in men than in women. The relationship between depression and malnutrition in nursing home residents was investigated in a German study, and no differences was found in the mean MNA score between subjects who had depression and those who did not. However, a modest association was demonstrated between malnutrition and depression in a regression analysis [
6]. A study of 267 community-living elderly in Brazil [
16] showed a positive relationship between malnutrition and depression.
We believe the 'MUST' tool used in the current study has an advantage over the MNA with regards to the associations explored. The MNA has been validated in a number of studies of elderly individuals, but it includes information about both neuropsychological problems and psychological stress [
17]. A positive correlation between the MNA risk score and the symptoms of depression could therefore be anticipated. The 'MUST' tool does not include any component that explores mental health. This is the first study to use either the 'MUST' tool or the SCL-10 the assessment of the relationship between risk of malnutrition and mental health.
Increased risk of malnutrition (combining medium and high risk) was found in 7.1% of the individuals in the current sample. In previous studies of community-living elderly individuals, prevalence rates for the risk of malnutrition varied from 2.5% to 21% [18-21]. This variation in prevalence may reflect the use of different criteria both to define malnutrition and differences in sample selections.
In accordance with former studies on adult and elderly individuals, we found that women had more mental health symptoms than men [
22]. This gender difference is not fully understood but may to some extent be explained by an underreporting of depressive symptoms by male individuals [
23].
Mental health may be assessed by both a categorical approach, which considers diagnoses that are based on a distinct cut-off, and a dimensional approach, which considers symptoms along a continuum. The latter approach also takes into account subthreshold symptoms of anxiety and depression, which may also adversely affect daily life [
24,
25]. The present study revealed statistically significant associations using both a categorical and a more dimensional approach.
Somatic diseases, especially stroke, myocardial infarction and cancer, represent risk factors for depressive symptoms in elderly individuals [
26]. Somatic diseases may also increase the risk of malnutrition [
21]. However, adjusting for the history of these three important somatic diseases did not affect the conclusions of the current study.
Individuals with BMI < 20.0 kg/m
2 had a two to three times higher prevalence of significant mental health symptoms (table
3) and the corresponding adjusted OR was 2.0 and of borderline significance (p = 0.06) (table
3). Obesity (BMI >30.0) was not associated with more mental health symptoms. Previous studies have reported both a decreased [
7] and an increased risk [
8] of depression in obese elderly individuals. However, the lower BMI categories were not specifically examined in these two studies.
The Tromsø study included participants from both urban and rural areas although the majority live in the city centre. Our results may not be generalised to all other elderly populations as both living conditions and health care organisation differ between countries. However, we believe that it is likely that similar relationships are present in other similar community living elderly Western populations.
As discussed above, this study has several strengths as well as some potential limitations. First, the SCL-10 captures symptoms of both anxiety and depression, although depression is more influential in the relationship with nutritional status. However, considerable overlap exists between anxiety and depression, which often appear as co-morbid disorders [
1,
27].
Second, eating disorders were not assessed in this study. In a recent review of eating disorders in the elderly, depression was described as the most important co morbid condition. However, the prevalence of eating disorders is low in the elderly population [
28].
Third, the study sample that exhibited valid values for the SCL-10-score and the 'MUST' score represented 52% of the target population. Thus, selection bias may be a concern. However, it is likely that the elderly men and women who did not complete the survey or omitted key information were frailer, more cognitive impaired and more prone to both malnutrition and impaired mental health than the persons who were included in the study sample.
Fourth, by using 1.85 as the cut-off for the SCL-10 score yielded significant mental health problems of 4.2% in men and 9.8% in women, which may be an underestimation. In elderly people, the prevalence of major depression is 1 to 4%, the prevalence of minor depression is 4 to 13% [
26] and the prevalence of anxiety is 3.2% to 14.2% [
29]. The cut-off of 1.85 for the SCL-10 score was adopted from previous studies that describe the SCL-10 [
12] and has not been compared to clinical diagnostic interviews in community-living elderly men or women. However, the main purpose of the current study was not to describe the prevalence of mental health problems but to determine the relationship between impaired mental health and nutritional status.
Fifth, there was no screening of cognitive decline in this study. Mild cognitive impairment can be present a long time before dementia is identified and this might be associated with malnutrition and symptoms of anxiety and depression. However, participants had to both independently visit a research centre and accomplish a detailed self administrated questionnaire. This reduces the risk of cognitive impairment among participants included in the study population.
The current study also demonstrated a significant association between subthreshold mental health symptoms and the risk of malnutrition. Several reports have described other adverse health effects that are related to subthreshold mental health symptoms in elderly individuals [
30,
31]. The cut-off for the SCL-10 used in the current study identified 13.6% of men and 22.4% of women with subthreshold symptoms. This corresponds well with the 20.2% of older women identified with subthreshold depression in a recent study that used the Center for Epidemiological Studies Scale for Depression (CES-D) [
32].
The cross-sectional design hampers conclusions about the directionality of the associations. The most important is probably the influence of depression on appetite and food intake. This can lead to weight loss and increase the risk of malnutrition. In the Diagnostic and Statistical Manual of Mental Disorders [
33], both weight gain and weight loss are among the diagnostic criteria for depression. In contrast, malnutrition may also be associated with micronutrient deficiencies that adversely affect mental health. Inadequate intake of nutrients and energy may lead to deficiency of folic acid, thiamine or cobalamin [
34] which might worsen mental health symptoms. A recent study that evaluated the impact of weight change alone in elderly people found that weight loss predicted an increase in depressive symptoms [
35]. Hence, a bidirectional relationship between the risk of malnutrition and mental health symptoms may be present and result in a vicious circle over time in affected individuals.
Authors' contributions
JMK, JF, OG and BKJ were responsible for the initial design of the study. JMK did the analyses and wrote the first draft of the paper. BKJ contributed to the analyses, interpretation of the results and the review of the drafts. All authors contributed to the interpretation of the data and review of the manuscript for important intellectual content. All authors read and approved the final manuscript.