Nutritional status according to the Mini Nutritional Assessment (MNA®) and frailty in community dwelling older persons: A close relationship

Abstract

Objective

This study investigates the association between MNA results and frailty status in community-dwelling older adults. In addition the relevance of singular MNA items and subscores in this regard was tested.

Design

Cross-sectional study.

Setting

Community-dwelling older adults were recruited in the region of Nürnberg, Germany.

Participants

206 volunteers aged 75 years or older without cognitive impairment (Mini Mental State Examination >24 points), 66.0% female.

Measurements

Frailty was defined according to Fried et al. as presence of three, pre-frailty as presence of one or two of the following criteria: weight loss, exhaustion, low physical activity, low handgrip strength and slow walking speed. Malnutrition (<17 points) and the risk of malnutrition (17–23.5 points) were determined by MNA®.

Results

15.1% of the participants were at risk of malnutrition, no participant was malnourished. 15.5 % were frail, 39.8% pre-frail and 44.7% non-frail. 46.9% of the frail, 12.2% of the pre-frail and 2.2% of the non-frail participants were at risk of malnutrition (p<0.001). Hence, 90% of those at risk of malnutrition were either pre-frail or frail. For the anthropometric, dietary, subjective and functional, but not for the general MNA subscore, frail participants scored significantly lower than pre-frail (p<0.01), and non-frail participants (p<0.01). Twelve of the 18 MNA items were also significantly associated with frailty (p<0.05).

Conclusions

These results underline the close association between frailty syndrome and nutritional status in older persons. A profound understanding of the interdependency of these two geriatric concepts will represent the basis for successful treatment strategies.

Introduction

The Mini Nutritional Assessment (MNA®) is a validated screening tool for malnutrition in older persons (1) which has been widely used in clinical practice as well as in research (2, 3). In community-dwelling older adults a risk of malnutrition is notably prevalent: In a comprehensive review of the year 2009 Guigoz reported an average prevalence of 24% (4), and in a more recent review of Cereda an average prevalence of 27% was found (5). According to these reviews overt malnutrition was present on average in 2% (4) and 4% (5) of community-dwelling older persons.

Malnutrition can increase the age-associated loss of muscle mass and strength and therefore is seen to play an important role in the development of sarcopenia and consecutively also of physical impairment (6., 7., 8.), which both represent substantial elements of the frailty syndrome. Frailty is widespread in community-dwelling older adults: Santos-Eggimann et al. reported an average prevalence of frailty of 17.0% and of pre-frailty of 42.3% in a multinational study that included10 European countries (9). Frailty is strongly associated with negative health outcomes (10, 11), i.e. hospitalisation and loss of independence (12). The most widely accepted definition of frailty has been developed by Fried and co-workers (11).

Abellan van Kan recently reported a close association between frailty and the risk of malnutrition according to the MNA® in a small sample of older adults (13). The authors concluded that older persons at risk of malnutrition should be considered as frail (13). Up to now this has been the only report focusing on the association between the MNA® and frailty in community-dwelling older adults.

We hypothesized that certain MNA items or groups of items may be closer related to frailty than others. Therefore, the present study aimed to explore the association between the risk of malnutrition according to the Mini Nutritional Assessment (MNA®) and frailty in community-dwelling older adults in more detail by additionally looking at the MNA subscores and singular MNA items.

Methods

Design

From August 2009 to September 2010, a convenience sample of 206 community-dwelling volunteer older adults was recruited for this cross-sectional study in the region of Nürnberg, Germany. Inclusion criteria were age of 75 years or older, living independently at home and having neither mental impairment (Mini Mental State Examination: MMSE >24 out of 30 points (14)) nor acute disease.

Candidates were mainly approached by a newspaper advertisement and were screened for eligibility with the help of a telephone interview. Moreover eligible patients of a day-clinic and a rehabilitation centre were approached via personal contact. The assessments were conducted at the study site or, if participants were not able or willing to come there, were visited at their home.

The study was approved by the ethics committee of the Friedrich-Alexander-Universität Erlangen-Nürnberg and an informed consent form was signed by all subjects.

Sample characteristics

Sample characteristics were assessed in personal standardized interviews by trained personnel. Age was calculated from birth date. Height was obtained using a tape measure with participants standing upright at a wall without shoes. Weight was measured in light clothing without shoes. If height or weight could not be measured due to physical impairments, they were obtained from a recent physical examination or documented as reported by participants. BMI was calculated as weight [kg] / height² [m²]. Self-reported living situation was categorized as “living alone” or “not living alone”. Level of education was defined as “high” for participants with university entrance diploma or higher degrees, “medium” for those who attended a secondary school and “low” for participants with only elementary school or no degree. The Geriatric Depression Scale (GDS; max. 15 points (15)) was used to examine emotional status with scores >6 points denoting depressive mood. Dependency in everyday life was determined with the questionnaire on instrumental activities of daily living (IADL, 8 questions, max. 8 points) of Lawton and Browdy (16) and the Barthel index of activities of daily living (ADL, 10 questions, max. 100 points) (17). A higher level of dependency is represented by a lower score. Self-reported medication use was classified as “more than three medications” or “less than three medications”. To assess comorbid conditions the Charlson Comorbidity Index was used, which is a sum score of 19 weighted diseases (1, 2, 3 or 6 points) that have been found to increase mortality. A higher score indicates a higher mortality risk (18).

Mini Nutritional Assessment (MNA®)

The MNA® is a screening tool for identifying overt malnutrition or the risk of malnutrition in older people (1). The MNA® consists of 18 items (A-R) which are rated with 0 up to 3 points. We merged the single MNA items into subscores based on the proposition of Guigoz et al. (4) and calculated the following subscores: “anthropometric” (weight loss [B], BMI [F], mid-arm- [Q] and calf-circumference [R]; max. 8 points), “general” (stress or acute disease [D], neuropsychological problems [E], medication [H], pressure sores or skin ulcers [I]; max. 6 points), “dietary” (anorexia [A], number of meals [J], intake of beverages [M], protein rich foods [K] and fruits [L]; max. 7 points), “subjective” (self-perception of nutritional state [O] and health state [P]; max. 4 points) and “functional” (mobility [C], living independently [G], difficulties with feeding [N]; max. 5 points).

A maximum total score of 30 points can be achieved with a cut-off of ≥ 24 points indicative of being well-nourished, 17-23.5 points for risk of malnutrition and <17 points for being malnourished.

Frailty assessment

Frailty was diagnosed according to Fried et al. (11) based on five criteria: weight loss (> 4.5 kg in the last year), exhaustion (self-reported feeling that everything was an effort or that one could not get “going” more than 2 times a week), low grip strength (Jamar dynamometer, men ≤ 29-32 kg, women ≤ 17-21 kg stratified by BMI), low walking speed (depending on gender and height > 6-7 sec/ 4.57 m) and low physical activity (Minnesota Leisure Time Activities Questionnaire (19)) (men < 383 kcal/ week, women < 270 kcal/ week). Participants with 3 or more positive citeria were defined as frail, with 1-2 as pre-frail and participants without positive criteria as non-frail.

Statistics

Sample characteristics are presented as median (min.-max.) for continuous variables and as percentage for categorial variables. Single MNA items were dichotomised into maximum versus less than maximum score. Differences in the distribution of categorial sample characteristics, MNA categories and single MNA items between frail, pre-frail and non-frail participants were tested for significance by chi2 testing. For continuous variables among the sample characteristics the Kruskall-Wallis test and for MNA subscores the Mann-Whitney U test was used for comparison of frailty groups. An acceptable level of significance was established a priori as p<0.05.For data analysis the statistical software package sPss 19 was used.

Results

Sample characteristics in the three frailty groups are presented in table 1. Study participants had a median age of 83 (75-96) years, 66.0 % were female. Thirty-two (15.5%) of the participants were frail, 82 (39.8%) pre-frail.

Table 1.

Main characteristics of the population in three frailty groups

Characteristics	non-frail (n=92)	pre-frail (n=82)	frail (n=32)	p
Female sex [%]	57.6	67.1	87.5	0.009 a
Age [years] b	82 (76-91)	84 (76-94)	86 (75-96)	0.001 c
BMI [kg/m2]	26.7 (21.0-35.0)	28.1 (20.9-35.3) 26.4 (18.6-36.1)	0.118
Living alone (%)	50.6	59.5	80.0	0.019
Educational level
Low [%]	42.3	39.2	53.3
medium [%]	23.5	31.6	46.7
high [%]	34.1	29.1	0.0	0.005
MMSE [points]	29 (25-30)	29 (24-30)	29 (25-30)	0.585
GDS [points]	0 (0-4)	2 (0-5)	3 (0-9)	<0.001
IADL [points]	8 (5-8)	8 (1-8)	7 (2-8)	<0.001
ADL [points]	100 (90-100)	100 (55-100)	95 (40-100)	<0.001
CCI [points]	0 (0-4)	2 (0-5)	3 (0-9)	0.002
At risk of MN [%]	2.2	12.2	46.9	<0.001

Note: “low”= elementary school or no degree; “medium”= secondary school; “high”= university entrance diploma or higher degrees; MMSE= Mini Mental State Examination (14); GDS= Geriatric Depression Scale (15); IADL= Instrumental Activities of Daily Living ((16)); ADL= Activities of Daily Living (17); CCI= Charlson Comorbidity Index (18); ^a Chi2 Test (for all ordinal variables); ^b Median (min.-max.) (all continuous variables); ^c Kruskall-Wallis Test (all continuous variables)

The most prevalent frailty criterion was “low grip strength” which was found in 39.1% of the participants, followed by “low walking speed” and “exhaustion” each found in 21.9%. “Low physical activity” was reported by 20.3% and weight loss by 8.3%.

The frailty groups differed significantly with regard to gender (p<0.01) and age (p<0.01): With increasing numbers of frailty criteria participants were significantly older and more often female. Frail older adults lived alone more often (p<0.05), had a lower educational level (p<0.01), higher depression scores (p<0.001) and lower ADL (p<0.001) and IADL (p<0.001) scores than the pre-frail and non-frail groups. Frail participants scored significantly higher on the Charlson Comorbidity Index than pre-frail and non-frail ones (p<0.01).

Twenty seven (15.1%) were identified as being at risk of malnutrition. None of the men and women of our sample were malnourished. Risk of malnutrition was found in 2.2% of the non-frail, in 12.2% of the pre-frail and in 46.9% of the frail participants (p<0.001). This association is displayed in absolute numbers in figure 1.

Figure 1.

Histogram of the distribution of participants at risk of malnutrition and well-nourished individuals according to MNA® in the three frailty groups by absolute numbers

Table 2 shows the prevalence of the single MNA items in non-frail, pre-frail and frail participants. There was a significant association between frailty status and 12 of the 18 MNA items. For “neuropsychological problems”, “eating dependency” and “mid-arm circumference <22 cm” this association was also evident but not significant. No association was found for the items “BMI <23 kg/m²“, “<3 full meals /day” and “<3 markers for protein intake”.

Table 2.

Prevalence of single MNA items below maximum score in non-frail, pre-frail and frail older adults [%]

MNA items	Letter	MNA subscore	non-frail n=(92)	pre-frail n=(82)	frail n=(32)	pa
Moderate or severe anorexia	A	dietary	3.3	13.4	31.3	<0.001
Weight loss >1kg /3 months or unknown	B	anthropometric	8.7	17.1	31.3	0.009
Impaired mobility	C	functional	0.0	1.2	9.4	0.003
Stress or acute disease	D	general	14.1	35.4	40.6	0.001
Neuropsychological problems	E	functional	2.2	7.3	12.5	0.075
BMI <23 kg/m2	F	anthropometric	7.6	6.1	15.6	0.240
Not living independently	G	functional	2.2	8.5	21.9	0.002
> 3 drugs /day	H	general	34.8	53.7	71.9	0.001
Pressure sores or skin ulcers	I	general	2.2	11.0	28.1	<0.001
<3 full meals /day	J	dietary	16.3	7.3	18.8	0.127
<3 markers for protein intake	K	dietary	60.9	53.7	59.4	0.617
<2 servings of fruit /day	L	dietary	9.8	17.1	31.3	0.016
<5 cups of fluid /day	M	dietary	2.2	9.8	21.9	0.002
Eating dependency	N	functional	0.0	3.7	6.3	0.091
Feels malnourished or is uncertain of nutritional state	o	subjective	3.3	1.2	21.8	<0.001
Considers health status not better than peers	P	subjective	28.3	40.2	68.8	<0.001
Mid arm circumference <22cm	Q	anthropometric	0.0	2.4	6.6	0.080
Calf circumference <31cm	R	anthropometric	1.1	2.4	12.5	0.007

^a Chi-square Test

Figure 2 displays the boxplots of the five MNA subscores in the three frailty groups. In all subscores frail participants scored significantly lower than pre-frail (p<0.05) and non-frail participants (p<0.01) except for the “general assessment”. Here the median scores of pre-frail did not differ significantly from those of frail participants, but those of non-frail did differ significantly from those of both pre-frail and frail participants (p<0.01)

Figure 2.

Boxplots of subscores of the MNA® in the three frailty groups. The boxes represent the interquartile range with the bold horizontal lines denoting median score. The whiskers show the highest and lowest values within the 1.5-fold interquartile range. The circles represent outliers and asterisks extreme outliers

Discussion

This is the first study to analyse in detail the association between the risk of malnutrition according to the MNA® and frailty according to Fried. We found that about half of the frail participants were at risk of malnutrition and that over 90 % of those at risk of malnutrition were frail or pre-frail. The strong association between frailty and a risk of malnutrition is considerably due to a relevant overlap in contents of the Fried criteria and the MNA. The most obvious overlap is the mutual assessment of weight loss as discussed below. Additionally the MNA items low mid-arm- and low calf-circumference are indicators of decreased muscle mass, which is closely related to the frailty criteria of low gait speed and low handgrip strength. Furthermore, the frailty criterion “exhaustion” is assessed by self-perception of “everything was an effort” or “I could not ‘get going’“, which may show some overlap with depressive conditions as assessed in the neuropsychological item of the MNA®.

For the assessment of frailty, in terms of comparability, we used the original definition of Fried et al. which focuses on the physical phenotype of frailty and which has already been applied in several large epidemiological studies (11, 20., 21., 22., 23.).

In accordance with Fried et al. (11) frail participants in the present study were older than pre-frail and non-frail ones, and female gender and lower educational level dominated in frail individuals. Although there were significant differences in ADL and IADL scores between frailty groups (table 2), even the frail men and women were quite autonomous in their daily living, indicated by high median scores. Remarkably, frail older adults tended to live alone more often than pre-frail and non-frail. Although frail participants had the highest depression scores of the three groups, according to the GDS their median score was still lower than the cut-off for depression of 6 points.

Risk of malnutrition was prevalent in 15.1% of our community-dwelling seniors and no participant was malnourished. Similar results were found in another sample of German community-dwelling men and women (11.0% at risk, 0% malnourished) (24) and by Johansson et al. in a Swedish sample of older adults living at home (16.0% at risk, 1.6% malnourished) (25). Our prevalence rates are somewhat lower than those reported in the reviews of Guigoz (24.0% at risk, 2.0% malnourished) (4) and Cereda (27.4% at risk, 4.2% malnourished). It has to be considered, however, that these values are calculated average values that were derived from studies with noticeably higher and lower prevalence rates (at risk 0-76%, malnourished 0-26%), with different mean age and health status of participants. In addition the respective studies came from different geographic areas. Secondly, our data is derived from a convenience sample that lacks representativeness for the German older population in general.

In the present study the association between the MNA® and frailty has been documented on three levels: MNA total score categories, MNA-subscores and single MNA items.

For the MNA category “at risk of malnutrition” we found a significantly increased prevalence in pre-frail and frail community-dwelling older adults compared to non-frail. This is in line with the conclusions of Guigoz (4) and of Cereda (5), who both showed for different settings in their reviews on the MNA®, that the prevalence of risk of malnutrition in older adults increases with the level of dependency. This is underlined by the observation, that frailty (table 1) as well as the MNA® (8, 26) are strongly associated with ADL.

Our results are also generally in accordance with the short report of Abellan van Kan and Vellas (13), on the higher prevalence of a risk of malnutrition in frail older adults compared to pre-frail and non-frail. In our sample we also detected two non-frail individuals being at risk of malnutrition. Abellan van Kan and Vellas found overt malnutrition in some of their participants (all of them frail) (13) which was not the case in our sample. These results illustrate the fact that the MNA® has only limited value to identify frailty.

In addition to the overall risk of malnutrition we also found all MNA subscores to be associated with frailty.

In the anthropometric assessment nearly all pre-frail and non-frail individuals reached the maximum score whereas approximately half of the frail men and women scored lower. Calf circumference was the most distinctive item of this subscore. The reduced calf circumference in many pre-frail and frail participants indicates a loss of muscle mass and therefore increased prevalence of sarcopenia. Since weight loss is a known risk factor for frailty (28) and one of the Fried criteria, the MNA item “weight loss” was associated with frailty as expected. While 8.3% of our participants reported weight loss according to the frailty definition (>4.5 kg in the last year), the respective MNA criterion was positive in 15.5% (>1kg in the last three months or does not know).

In accordance with the frailty concept (11) and with other study results (21, 29, 30) health related issues as evaluated by the general assessment were associated with frailty. The poorer health status of the frail older adults of our sample is also indicated by the higher prevalence of comorbidity (table 1) and high medication use (table 2). The closest association with pre-frailty and frailty in this subscore was observed for the item “pressure sores or skin ulcers” (table 2).

In the dietary assessment anorexia increased significantly with frailty level, which is in line with Morley et al.’s conclusion, that anorexia can be regarded as an early risk factor of frailty (31). In line with Bartali et al.’s findings, that a low intake of vitamins is associated with frailty (20), we observed that two servings of fruit per day were consumed by the frail participants least often. The association between the MNA item “low fluid intake” and frailty might be explained by a subclinical state of dehydration which might be related to exhaustion and reduced physical performance.

The subjective assessment showed that the self-perception of nutritional state and health state of our participants may be interpreted as realistic. Frail individuals, who scored lower in the dietary assessment and in the general assessment, rated themselves as being at worse health and nutritional status than their peers accordingly.

The strong association between frailty and the functional subscore can be explained by the physical focus of Fried et al.’s frailty definition. This is also indicated by the increasing dependency in daily activities in pre-frail and frail participants. Additionally, low mobility has been found to be a risk factor for frailty (33). Though, it has to be mentioned, that apart from outliers, a lower score was only found in the frail and not in pre-frail or non-frail individuals.

At single item level, it was shown, that BMI – contrary to the results of Hubbard et al. (27) – was not associated with frailty. This may be due to the high median BMI even of our frail participants (table 1) and to the rather high BMI cut-off of 23 kg/m² used in our analysis (table 2). Furthermore, contrary to Bartali et al. (20) and Beasley et al. (32) we did not find an association of frailty with the amount of protein ingested. An explanation may be the only very rough estimation of protein intake provided by the MNA®. Additionally the MNA items <3 full meals/day and mid arm circumference <22cm were also not significantly more prevalent in frail participants. That means, four of the ten MNA items directly focusing on aspects of food intake and nutritional status were not associated with frailty. Taken together with the strong association between the frailty-concept and the MNA® as found in this investigation, these observations raise the question, if not at least parts of the MNA®, in fact measure frailty. Additionally, they also pin-point that the MNA® is a screening tool to identify risk groups that may benefit from nutritional interventions and not a diagnostic tool for malnutrition.

It may be regarded as a strong point of our study that it applied a multi-level approach when investigating the association between frailty and the MNA® in categories, subscores and single items to profoundly describe boundary points of the two concepts. This is also the first approach investigating a functional subscore of the MNA® in the setting of community dwelling older adults. Another advantage of this study is the use of standardized, valid and widely accepted definitions of frailty and malnutrition, which will facilitate the comparison to previous and future investigations.

A clear limitation of the study is the relatively small sample size. Furthermore, our sample of volunteers lacks representativeness. Nevertheless, the prevalence rates of frailty and malnutrition of our sample were comparable to larger studies. The overlap of the frailty concept of Fried et al. and the MNA® may also be seen as a limitation. Therefore, for future research it may be indicated to compare the frailty concept according to Fried to other malnutrition screening tools or, the other way round, compare the MNA® with other frailty tools.

The present study focused on the MNA long-form which is rather time-consuming and has often been replaced in clinical practice by the recently adapted MNA short-form (24). Future evaluations should therefore also investigate the association between frailty and the MNA short-form.

In conclusion, the presented results underline the interconnection and overlap of the concepts of frailty and malnutrition. Our results also suggest, that the MNA category “at risk of malnutrition” may not be seen as a “grey” category of neither being malnourished, nor being well nourished. In fact the at risk category may be regarded as a specific third group between those two groups that is characterized by a high degree of functional impairment which may jeopardize adequate nutrition. It is therefore crucial for medical personnel involved in caring for older individuals to, on the one hand, pay particular attention to nutritional status in frail men and women and on the other hand to be aware of a decline in functionality in individuals with low MNA scores. Only a profound understanding of the interdependency of these two geriatric concepts, malnutrition and frailty, builds the basis for successful treatment strategies.

Funding and conflicts of interest

This work was partly supported by Nestlé HealthCare, Vevey, Switzerland and the Theo und Friedl Schöller-Foundation, Nürnberg, Germany. JB, RD, MJK, KV, JMB, WU, CCS and DV declare that they have no conflict of interest.