Title: Malnutrition and associated factors in elderly hospital patients: a Belgian cross-
sectional, multi-centre study.
Authors: Katrien Vanderweea, Els Claysb, Ilse Bocquaerta, Micheline Gobertc, Bert
Folensd, Tom Defloora
a Nursing Science, Department of Public Health, Faculty of Medicine and
Health Sciences, Ghent University, Belgium
b Department of Public Health, Faculty of Medicine and Health Sciences,
Ghent University, Belgium
c Department of Public Health, Faculty of Medicine, Catholique University of
d Belgian Federal Public Service, Health, Food chain safety and Environment.
Short title: Belgian cross-sectional, multicentre study of malnutrition among elderly
Address for correspondence:
UZ Blok A – 2nd floor
De Pintelaan 185
Email address: firstname.lastname@example.org
Tel: +32 9 332 36 29
Fax: +32 9 332 50 02
Background & Aims: In Belgium, general data on the prevalence of malnutrition are
lacking. Prevalence rates are necessary to gain insight into the magnitude of
malnutrition and to establish a nutrition policy that takes the limited health care
resources into account. This study aimed to obtain insight into the prevalence of
malnutrition in Belgian elderly hospital wards and to identify factors associated with
the malnutrition prevalence.
Methods: A cross-sectional, multi-centre study in elderly wards of Belgian hospitals.
The nutritional status was assessed using the Mini Nutritional Assessment. A
standardised questionnaire was used to record demographic data and data on
potential factors associated with malnutrition.
Results: Out of 2329 elderly patients, 33% suffered from malnutrition. Almost 43% of
the patients were at risk of malnutrition and 24% were well-nourished. Having
swallowing difficulties, taste difficulties, and being transferred from a nursing home
were strongly associated with being malnourished.
Conclusion: The malnutrition prevalence in Belgian elderly hospitals wards is similar
to international figures. Elderly who have swallowing difficulties, taste difficulties, or
coming from a nursing home may need adequate nutritional care. Given the negative
impact of malnutrition on mortality and morbidity, an emphasis should be placed on
an effective nutritional policy.
Malnutrition, Prevalence, Elderly, Hospital, Mini Nutritional Assessment
In recent years we are increasingly faced with obesity in our Western society. This
health problem receives much attention. Conversely, malnutrition receives less
attention, although it is a frequently occurring health problem.
In the current study we focus on malnutrition in the elderly and use the definition of
malnutrition developed by Chen, Schelling & Lyder (2001): “Faulty or inadequate
nutritional status; undernourishment characterized by insufficient dietary intake, poor
appetite, muscle wasting and weight loss”.1
Malnutrition has a detrimental effect on physical and mental health which may lead to
a prolonged hospital stay and increased costs for patient and society 2;3. In Belgium,
the total health care expenditure as a proportion of gross domestic product (GDP)
was 9.3% in 2004 which is above the European average4. The growth in health
expenditure in Belgium is comparable to that in other Western European countries
and can be explained by several factors, such as the increasing number of elderly
people, higher expectations, growth in real GDP and increasing implementation of
new technology in the health care sector. Belgium had the fifth highest health care
expenditure per capita measured in purchasing power parity among European
countries. It is only higher in Luxembourg, the Netherlands, Germany and France.
Public sector funding as percentage of total expenditure on health care fluctuates
around 70% in Belgium and is expected to rise. One of the challenges for the future
Belgian health system is how to improve efficiency of the health system 4. The
Belgian government started with a National Food and Health Plan (NVGP-B) to
increase the health status of the population 5.
For the above-mentioned reasons, it is essential to invest in the prevention of
malnutrition. Screening for malnutrition is an essential measure in the prevention of
Internationally, the prevalence of malnutrition varies between 13 and 78% 3.
Recently, in a few Western European countries nationwide prevalence studies have
been conducted in hospitals 6;7. The overall prevalence rate of malnutrition was
estimated at 23.8% in the Netherlands (n=8028) and at 27.4% in Germany (n=1886).
These studies reported prevalence rates in hospitalised elderly patients ranging from
32.9% to 56.2%. Differences in methodology, definition of malnutrition, patient
characteristics, and setting among these studies may partially explain the wide range
of these figures.
Prevalence rates provide insight into the magnitude of the problem and help to
establish a nutrition policy that takes the limited health care resources into account.
General data on the prevalence of malnutrition in Belgium are lacking. These data
will round out the insight in malnutrition in Western Europe. Belgium is likely to be
representative for Western Europe as it located centrally in between the northern and
southern part of Western Europe.
The aim of this study was to obtain insight into the prevalence of malnutrition in
Belgian elderly hospital wards and to identify factors associated with the prevalence
Materials and Methods
Setting and sample
The sample for this cross-sectional, multi-centre study consisted of elderly patients
from Belgian hospitals. All general and teaching hospitals with elderly wards in
Belgium were invited to participate in this study. The participation was completely
voluntarily. The participating centres did not receive any financial compensation.
An elderly ward was defined as “an acute hospital ward which provides medical
geriatric diagnosis, treatment and revalidation and assures nursing and caring for the
elderly patients in the most excellent conditions” (Royal Decree 23-10-1964 and
modified by decree 12-04-1984 Art. 1). In Belgium, there are 111 hospitals with
Patients were included in the study if they were 75 years or older and gave informed
consent to participate in the study.
To evaluate the nutritional status of the elderly, the short form of the Mini Nutritional
Assessment (MNA-SF) was used. The MNA is extensively validated and provides a
most comprehensive insight in the nutritional status of the elderly patient 8-11. This
instrument is recommended by the European Society of Parenteral and Enteral
Nutrition (ESPEN) for the detection of the presence of malnutrition and the risk of
developing malnutrition among the elderly in hospitals, nursing homes and home-
care programmes 8. The Belgian Health Authorities recommend to use the MNA for
The MNA contains 18 items which are divided into four categories: anthropometric
assessment, general assessment, dietary assessment, and subjective assessment.
However, as the completion of the full MNA tool is considered to be too time-
consuming in geriatric screening situations, MNA-SF was developed 13. The MNA-SF
is highly correlated with the full MNA 13;14. The MNA-SF consists of six items: food
intake, weight loss, mobility, psychological stress or acute disease,
neuropsychological problems and body mass index (BMI).
The maximum score on the MNA-SF amounts to 14. If a patient is identified as at risk
for malnutrition by the MNA-SF (score <12), the remaining 12 items of the MNA
need to be assessed 8;13.
Patients with an MNA score of <17 are considered as malnourished, patients with a
score between 17 and 23.5 are identified as at risk for malnutrition. A score of ≥ 12
on the MNA-SF or ≥24 on the full MNA indicate that the patient is well-nourished.
When it was not possible to measure the weight or length of immobile or bed bound
elderly, an estimation was used. To determine their weight, patients were asked to
estimate their own weight or the health care professional made a subjective
estimation. To determine the estimated length, the formula of Chumlea was used
which calculates the length of the patient from his knee height. 15 If the elderly had a
cognitive impairment and was not able to answer the questions, a family member or
a health care worker who knew the patient best was consulted.
In addition, a standardised questionnaire was developed and used to record
demographic data and data on potential factors associated with malnutrition. These
factors were selected based on a review of the literature 16;17. Data were recorded on
pathology, swallowing, chewing, taste, digestion (nausea, vomiting, diarrhoea), pain,
and mouth infection.
The content validity of this standardised questionnaire and also of the Dutch and
French translation of the MNA were evaluated by a double Delphi-procedure 18. This
Delphi-procedure was performed by an expert panel which consisted of 11 experts in
nutrition and/or the elderly. A clinical nurse specialist in nutrition, three head nurses
of elderly wards, an infection control nurse, a dietician, two professors of geriatrics, a
professor of gastroenterology, a gastroenterologist, and a lecturer in elderly care
participated in the expert panel. Four were French-speaking and seven Dutch-
speaking. The experts reached consensus after the second Delphi-round.
The software package TeleForm, version 10.1 (Cardiff Software, UK) was used to
develop the questionnaires.
Data collection procedure
In each participating ward one supervisor was assigned. These supervisors were
nurses, dieticians, or other health care professionals. Before the start of study, the
research team informed all supervisors and trained them to perform the survey. The
main purpose of this training was to ensure the correctness and uniformity of the use
of the MNA and the standardised questionnaire. For this purpose, patient cases were
used. Each supervisor obtained an information guide on both data collection
instruments. This guide supported the supervisors in instructing their colleagues who
assisted them with the data collection. All data were registered by a health care
professional. For practical reasons, the participating wards were allowed to choose
one day between May 16th and June 15th 2007 to perform the survey.
The study procedure was approved by the ethics committee (No B67020071952) of
the Ghent University Hospital and of each participating hospital. The aim of the study
was explained verbally to each patient. Written informed consent was obtained from
all patients. If the patient was unable to give consent, a relative was asked to give
their written informed consent.
Descriptive data were presented in frequencies and percentages or means and
standard deviations. For the univariate and multivariate analyses the independent
variable nutritional status was dichotomised in malnourished versus well-nourished or
at risk of malnutrition. Additionally, we did the analyses with nutritional status
dichotomised in malnourished or at risk of malnutrition versus well-nourished.
Using univariate binary logistic regression analysis, the odds ratio and related 95%
confidence interval for each variable was calculated. To evaluate the independence
of the observed associations, the variables with a value p<0.100 were simultaneously
entered in a multivariate binary logistic regression analysis. Multicollinearity and
interaction effects were tested. Multicollinearity was observed between the variables
swallowing and chewing. The variable with the highest odds ratio in the univariate
analysis (swallowing) was included in the multivariate model. No interaction effects
were perceived. In the multivariate analysis a value of p<0.05 was considered to
indicate statistical significance.
All statistical analyses were performed with the software package SPSS version 15.0
(SPSS, Inc, USA)
Eighty-one percent (n=90) of the hospitals with elderly wards in Belgium participated
in the study. Five teaching hospitals and 85 general hospitals took part in the study
representing 4834 potential participants. Figure 1 presents the flow chart of the
included patients. In total, 2667 patients of 178 elderly wards were screened. The
data of 336 patients were incomplete. Only patients with complete MNA-SF data
were included in the analysis. In total, 2329 elderly patients were analysed in this
study. There were no significant differences between included and excluded patients
related to the demographic variables (see Table 1).
The majority of the patients (91.9%) were recruited from general hospitals. Almost
70% (n=1629) of the patients were female. Their mean age was 83.8 years (SD 5.19)
and their mean BMI was 24.7 kg/m² (SD 5.20). The most frequently occurring
pathologies were dementia (15.7%), diabetes mellitus (11.9%), decompensate heart
failure (11.3%) and depression (8.0%) (see Table 2).
Demographic data of the total sample of patients admitted in all elderly wards in
Belgium during the study period were provided by the Belgian Federal Public Service
Health, Food Chain Safety and Environment. The mean age of the total sample was
82.5 years and 68.2% of the patients were female. The occurrence of most
pathologies was similar to that of the study sample: dementia 14.7%, decompensate
heart failure 15.3%, and depression 7%. Only the occurrence of diabetes mellitus
(24%) was higher in the total sample.
Prevalence of malnutrition
Nutritional assessment based on the MNA revealed that 33% (n=768) of all elderly
patients suffered from malnutrition. Almost 43% percent (n=997) of the patients were
at risk of malnutrition and 24.2% (n=564) were well-nourished. Table 3 presents the
characteristics of patients according to their nutritional status. There was no
significant difference in nutritional status between patients admitted to teaching
hospitals and patients admitted to general hospitals and between male and female
elderly. The mean age, length of hospital stay, BMI, and living in a nursing home
were significantly different between the three nutritional categories.
Table 4 shows associations between being malnourished and several patient
factors. This univariate analysis demonstrates that being 85 years or older, coming
from a nursing home, having chronic obstructive pulmonary disease (COPD),
malignant disease (oncology), pneumonia, urinary infection, decompensate heart
failure, delirium, dementia, depression, taste difficulty, swallowing difficulty, digestion
problem, pain, mouth infection, chewing difficulty, and a longer hospital stay were
associated with a significantly higher prevalence of malnutrition.
The same univariate analysis with being malnourished or being at risk for malnutrition
as dependent variable was performed. The results of this analysis were very similar.
Only decompensate heart failure was no longer a significant associative factor.
To confirm associations found in univariate analysis, a multivariate binary logistical
regression analysis was performed (see Table 5). Patients with swallowing difficulties
were almost five times more malnourished. Elderly patients coming from a nursing
home were almost three times more likely to be malnourished than patients coming
from home or a service flat. Patients with taste difficulties had two and a half times
more risk to be malnourished. Patients with digestion problems had approximately
85% more chance of being malnourished and patients who are 85 years or older
35%. Some pathologies (COPD, malignant disease, pneumonia, dementia, major
abdominal surgery, depression, and delirium) increased the chance of being
malnourished between 42% and 86%. Finally, patients with a longer hospital stay
also had a higher risk to be malnourished.
The same multivariate analysis with being malnourished or being at risk for
malnutrition as dependent variable was carried out. The results were fairly similar.
Major abdominal surgery, decompensate heart failure, and taste difficulty were no
longer significant independent factors. On the contrary, urinary incontinence became
a significant associated factor.
This study was the first Belgian large-scale cross-sectional multi-centre study
focusing on malnutrition in elderly patients. Based on the MNA, 33% of the elderly
hospital patients were malnourished. This prevalence figure is consistent with recent
European prevalence figures in elderly hospital patients 6;7;19. In the annual national
Dutch survey, the prevalence of malnutrition in geriatric hospital wards was similar
(32.9%). However, in that study, malnutrition was defined according to one of the
following criteria: BMI less than 18.5 kg/m², unintentional weight loss (6 kg in the
previous 6 months or 3 kg in the previous month), or BMI between 18.5 and 20 kg/m²
in combination with no nutritional intake for 3 days or reduced intake for more than 10
days 6. In the German study, the prevalence in the geriatric departments was higher.
In total, 56.2% of the patients suffered from moderate or severe malnutrition. The
subjective global assessment (SGA) and anthropometric measurements were used
to assess the nutritional status 7.
In Belgium, malnutrition is a substantial problem in elderly hospital wards. Besides
the 33% malnourished elderly, nearly 43% was at risk for malnutrition. Almost four
out of five elderly patients were at risk for malnutrition or suffered from malnutrition
and only 24% was well-nourished.
As mentioned above, this was the first large-scale study focusing on the prevalence
of malnutrition in Belgium. However, there have been some other studies, often small
in size, in which the nutritional status of elderly patients in Belgian hospitals was
investigated. 20-24 These studies focused, for example, on the prevention of
malnutrition, a nutritional care program, a minimum geriatric screening tool to detect
geriatric problems including nutrition, and the use of the MNA. In two of these studies
the prevalence of malnutrition was explored. 20;23 In the study of Gazzotti et al. 20,
studying the clinical usefulness of the MNA in geriatric medicine, 22% of the
hospitalised elderly patients were malnourished and 49% were at risk of malnutrition.
This prevalence figure is lower than in our study, however, the sample included only
175 patients. Next, in the study of Pepersack 23, 1139 patients consecutively
admitted to 12 geriatric wards were studied concerning the outcomes of a continuous
process improvement of a nutritional care program. For this purpose the median
score of the MNA was evaluated, however no real prevalence figures of malnutrition
were given. The median value on the MNA was 18 points which indicates also a high
prevalence of poor nutritional status among hospitalised elderly.
Several independent patient factors were found to be associated with malnutrition or
being at risk for developing malnutrition. These identified factors provide valuable
information on those patients with a high risk to develop malnutrition and may
consequently facilitate the detection of patients with a poor nutritional status.
Three factors were rather strongly associated with being malnourished, namely
swallowing difficulties, taste difficulties, and being transferred from a nursing home.
Among these factors, having swallowing difficulties had the most powerful
relationship with being malnourished.
Both swallowing and taste difficulties are known to be associated with age and
disease as well as treatment factors 3. Patients with swallowing difficulties may have
a decreased food intake which makes them more vulnerable for malnutrition.
Suominen et al. 19 found also a quite strong association between swallowing
difficulties (OR 3.03; 95%CI 2.10-4.37) and malnutrition (MNA <17) in nursing home
residents in Helsinki. Furthermore, having taste difficulties was also relatively highly
associated with a poor nutritional status. Elderly patients with taste difficulties may
have a poor appetite and consequently an insufficient food intake. In literature, taste
difficulties are often cited as a causal factor of malnutrition 3;25. Next, coming from a
nursing home was an important associated factor. In the Belgian study of Gazzotti et
al. 20 was also found that patients originating from nursing homes had a poorer
nutritional status than those living at home. Elderly coming from a nursing home are
generally highly dependent in their functioning. If they need to be admitted to the
hospital, their general condition will be poor which might make them more
susceptible for malnutrition. On the other hand, these elderly may already have been
malnourished in the nursing home as several studies have revealed that elderly
residing in nursing homes often suffer from malnutrition 19. In order to gain insight in
the problem of malnutrition in Belgian nursing homes, it would be interesting to
perform a prevalence study in this setting.
Various medical conditions such as COPD, pneumonia, and malignant disease may
reduce the appetite and have a negative influence on the nutritional intake 26. Several
studies revealed also an association between respiratory diseases and malnutrition.
In the Dutch annual prevalence study, also a multivariate logistic regression was
performed which indicated that COPD was independently associated with
malnutrition (OR 1.58 95% CI 1.40-1.80) 6. However, as stated above, malnutrition
was defined differently than in the current study. Pirlich et al.7and Meijers et al.
6,identified in their nationwide studies, respectively in Germany and the Netherlands,
malignant disease as an independent risk factor for malnutrition in hospital patients
(Germany: 1.51; 95%CI 1.18-1.93 and the Netherlands: OR 2.74; 95% CI 2.39-3.15).
In addition, pathologies such as dementia, depression, and delirium may be
associated with a poor appetite. It must be noted that the factors dementia and