European Journal of Clinical Nutrition (2010) 64, 887–893

& 2010 Macmillan Publishers Limited All rights reserved 0954-3007/10

www.nature.com/ejcn

ORIGINAL ARTICLE
Comparison of two malnutrition risk screening
methods (MNA and NRS 2002) and their association
with markers of protein malnutrition in geriatric
hospitalized patients
T Drescher1,2,4, K Singler2,4, A Ulrich2, M Koller3, U Keller1, M Christ-Crain1 and RW Kressig2

1
Division of Endocrinology, Diabetes and Clinical Nutrition, Department of Internal Medicine, University Hospital Basel, Basel,
Switzerland; 2Division of Acute Geriatrics, University Hospital Basel, Basel, Switzerland and 3Basel Institute for Epidemiology and
Biostatistics, University Hospital Basel, Basel, Switzerland

Background: Malnutrition occurs frequently in the elderly and is associated with increased morbidity and mortality. The mini-
nutritional assessment (MNA) has been used most frequently in the geriatric literature. The nutritional risk screening 2002 (NRS)
has been proposed as universal screening method for hospitalized patients. The aim of our study was to compare both tools as
they are correlated with protein malnutrition.
Methods: MNA, NRS, and markers of protein malnutrition were measured in 104 consecutive inpatients admitted to an acute
geriatric ward.
Results: The median age was 84 years (IQR: 78–89), 81 were females. The median body mass index was 23.1 kg/m2
(IQR: 20–27.3), the median upper-arm and calf circumferences were 25 cm (IQR: 23–29) and 33 cm (IQR: 29–36). According to
MNA, 23 patients were malnourished, 50 at risk of malnutrition, and 31 had a normal nutritional status. The NRS indicated that
35 were at moderate to severe risk of malnutrition and 69 at low risk. Serum prealbumin and retinol-binding protein
concentrations were inversely associated with the severity of malnutrition as indicated by the NRS (P ¼ 0.06 and o0.01,
respectively), whereas the MNA was not associated with these serum proteins. After adjustment for C-reactive protein and
creatinine clearance, only retinol-binding protein concentrations were consistently associated with both malnutrition scores.
Conclusions: The NRS seems to be superior compared with the MNA and serum proteins in identifying elderly patients at risk of
malnutrition during acute intercurrent illness.
European Journal of Clinical Nutrition (2010) 64, 887–893; doi:10.1038/ejcn.2010.64; published online 19 May 2010

Keywords: malnutrition screening; nutritional assessment scores; serum proteins; geriatric inpatients

Introduction and assessment tools have been proposed, including anthro-

Malnutrition is a frequent condition in the elderly and is
associated with increased morbidity, mortality, hospitaliza-
tions and a lower quality of life (McWhirter and Pennington,
1994; Sullivan, 1995; Flodin et al., 2000). Especially among
hospitalized or institutionalized elderly persons, malnutri-
tion affects a considerable part up to 40% and even worsens
during hospitalization (Norman et al., 2008). Several screening
Correspondence: Dr T Drescher, Abteilung Endokrinologie, Universitätsspital
Basel, Petersgraben 4, Basel CH-4031, Switzerland.
E-mail: dreschert@uhbs.ch
4 for a considerable number of patients when history is
These two authors contributed equally to this work.
Received 6 December 2009; revised 19 March 2010; accepted 6 April 2010;
published online 19 May 2010
pometric and biochemical parameters.
The mini-nutritional assessment (MNA) is the most
established nutritional screening tool for older adults
developed in 1994 (Guigoz et al., 1994; Guigoz, 2006). Its
screening part consists of six items, whereas further assess-
ment of the nutritional status is based on 12 additional
items, including dietary history and anthropometric measure-
ments. It was shown to predict mortality (Persson et al.,
2002) and the functional status in geriatric patients (Ödlund
Olin et al., 2005; Cereda et al., 2008). However, administering
the MNA is quite time consuming and not applicable
not available because of dementia or other problems of
communication (Bauer et al., 2005).
 Malnutrition screening in geriatric patients
T Drescher et al
888
The nutritional risk screening 2002 (NRS) has been described
by Kondrup et al. (2003b). It has been proposed as universal
screening tool for malnutrition in hospitalized patients by
assessing body mass index (BMI), weight loss, appetite and
severity of disease. It allows a more rapid and simple
identification of patients requiring nutritional support and
reflects especially the severity of acute comorbidities. In
contrast to the MNA, the NRS is applicable in nearly all
patients and takes much less time to perform (Bauer et al., 2005;
Kyle et al., 2006); in addition, it is better validated with
outcome data in various patient populations (Kondrup et al.,
2003b; Sorensen et al., 2008). However, the NRS was not
specifically developed for geriatric patients. The European
Society of Parenteral and Enteral Nutrition recommends the
use of MNA for the non-hospitalized elderly and the NRS for
hospitalized patients (Kondrup et al., 2003a). The MNA and
NRS have not been compared in their ability to predict markers
of protein malnutrition, that is serum levels of visceral proteins.
A number of serum proteins have been described to be
related to malnutrition with inconsistent results (Corti et al.,
1994; Mühlethaler et al., 1995; Omran and Morley, 2000a, b;
Sergi et al., 2006). Many of them are influenced by factors
other than nutrition, especially inflammation, and are,
therefore, not specific for malnutrition (Devoto et al.,
2006). The aim of our study was to compare two scores for
malnutrition (MNA and NRS) in their ability to predict the
degree of protein malnutrition, especially the serum con-
centrations of visceral proteins, that is albumin, prealbumin,
and retinol-binding protein. As acute inflammatory diseases
and renal dysfunction may influence serum protein levels,
we also analysed the predictive potency of the two tools in
patients without and with inflammatory diseases and
dependent on glomerular filtration rate (GFR).
Materials and methods
We performed a single centre cross-sectional study in a
sample of 104 consecutive acute geriatric patients enrolled
during a 5-month study period (August 2007–December
2007). The study was performed at the Acute Geriatric ward
of the University of Basel Hospital, Switzerland. Admitted
patients (average age: 84 years) were considered to have a
high degree of frailty and active multiple pathologies,
requiring a holistic geriatric approach. The project was
approved by the local ethics committee.
After receiving informed consent from each patient, blood
samples were obtained within 24 h after admission for
determination of serum albumin, prealbumin, retinol-binding
protein, blood lymphocytes, creatinine and C-reactive
protein. Proteins were measured on a Beckman-Coulter
nephelometry system (Beckman-Coulter, Brea, USA) using
antibodies against transthyretin, albumin (Beckman-Coulter)
and retinol-binding protein (Dako, Glostrup, Denmark).
The GFR was estimated with the Cockroft–Gault formula
(Cockroft and Gault, 1976).
European Journal of Clinical Nutrition
Calf and upper-arm circumferences as parameters of
muscle mass were measured and MNA and NRS were assessed
within the first 3 days after admission.
The MNA covers 18 items including anthropometric
assessment (BMI, calf and upper-arm circumference), general
assessment (medication, acute illness, psychological pro-
blems and mobility), nutritional assessment (fluid intake,
number of daily meals and composition of food intake) as
well as self-assessment of the nutrition and health status. The
maximal score is 30, a score from 17 to 23.5 indicates a
patient ‘at risk of malnutrition’, a patient with a score o17
has to be classified as ‘malnourished’.
The NRS consists of five items, of which one is the age of
the patient (470 years), one the BMI, one the appetite of the
patient, one is accidental weight loss, and one is the
consideration of the severity of acute illness of the patient.
The maximal score is 7. A score of 0 indicates patients
without risk for malnutrition, a score of 1–2 indicates low
and a score of 3–7 moderate to severe risk of malnutrition.
The mini-mental state was evaluated to assess cognitive
function (Folstein et al., 1975).
For statistical analyses, categorical data were compared
using w2 tests. For multigroup comparisons of normally
distributed data, one-way analysis of variance with least
square difference for post hoc comparisons was applied. For
skewed data, the Mann–Whitney U-test was used for two
groups, and the Kruskal–Wallis one-way analysis of variance
test for multiple group comparisons. Correlation analyses
were performed using Spearman’s rank correlations.
A P-value below 0.05 was considered statistically significant.
We used multiple linear regression analysis to assess the
association of the three protein markers, albumin, prealbu-
min, and retinol-binding protein, with each the MNA and
NRS in separate models. We used the original (that is non-
categorized) malnutrition score of each patient as outcome
and the protein markers as covariates in the linear regression
models. Owing to the susceptibility of these protein markers
to inflammation and possibly to renal function, we addi-
tionally adjusted the models for CRP and GFR. We graphi-
cally checked the assumptions of the linear models by
plotting residuals vs covariate values.
Finally, we performed cross-classification analyses of the
two scores after categorizing the individual MNA assess-
ments as normal, at risk of malnutrition, malnourished and
NRS assessment as low and moderate to severe risk of
malnutrition (Table 2). The Spearman’s rank correlation was
used to test the null hypothesis of independence of the
two scores.
Results
Baseline characteristics
Of the 122 consecutive patients included, there were missing
data in 18 patients, leaving complete data in 104 patients.
The median age of the 104 patients was 84 (IQR: 78–89)
 Malnutrition screening in geriatric patients
T Drescher et al
889
years, 81 were females. A total of 24 patients lived in nursing Table 1 Characteristics of the 104 patients
homes before admission to the hospital, 5 patients died
Age (years) 84 (78–89)
during hospitalization. The most frequent causes of admis- Female sex 81 (78%)
sion were infectious or inflammatory diseases with predo- Nursing home residents 24 (23%)
minance of urinary and respiratory tract infections, followed Death during hospitalization 5 (4.8%)
by cardiac and cerebrovascular diseases, falls, malignancies,
Primary cause of admission to hospital
dementia or depression and others. BMI, upper-arm and calf Infectious/inflammatory 32 (31%)
circumferences and other patient characteristics are shown Cardiac disease 16 (15%)
in Table 1. Cerebrovascular disease 9 (8.5%)
Fall (including fracture) 16 (15%)
Dementia/delirium/depression 8 (7%)
Malignancy 7 (6%)
Nutritional assessment scores Others 16 (15%)
According to the MNA, 23 patients (22%) were malnour-
ished, 50 patients (48%) were at risk of malnutrition Comorbidities
Dementia 33 (32%)
and 31 (30%) had a normal nutritional state. The NRS Diabetes mellitus 24 (23%)
showed that 35 patients (34%) were at moderate to severe Cardiac disease (including atrial fibrillation) 39 (38%)
risk of malnutrition and 69 patients (66%) at low risk Renal failure (creatinin clearance o60 ml/min) 67 (64%)
(Tables 1 and 2). Although both instruments were associated
BMI (kg/m2) 23.1 (20–27.3)
with each other (P ¼ 0.001), they showed substantial dis- Arm circumference (cm) 25 (23–29)
crepancies. Table 2 shows the cross-classification of the two Calf circumference (cm) 33 (29–36)
assessment scores regarding the normal, at risk and poor
nutritional state classification. The MNA classified fewer MNA 21 (17–24)
o17 points; poor nutritional status 23 (22%)
patients as malnourished than the NRS (23 vs 35, respec- 17–23.5 points; at risk for malnutrition 50 (48%)
tively). Of the 23 patients categorized as malnourished by 423.5 points; normal nutritional status 31 (30%)
the MNA, the NRS concordantly classified 11 as at severe risk
and 12 as at low risk of malnutrition. The NRS, on the other NRS 2 (1–3)
1–2 points; low risk for malnutrition 69 (66%)
hand, classified 35 patients as at severe risk of malnutrition. 3–7 points; moderate to severe risk for malnutrition 35 (34%)
Of these, 18 patients were judged as at risk and 6 patients as
normal with the MNA. In 60% of all observations, the MMS (maximum 30 points) 26 (21–28)
classifications of the two scores disagreed, mostly concerning
Laboratory analysis
the results ‘malnourished’ or ‘at risk of malnutrition’. Albumin (g/l) 31 (28–33)
Certainly, a normal nutritional status according to the Prealbumin (mg/l) 145 (102–177)
MNA can be compatible with severe risk of getting Retinol-binding protein (mg/l) 37.5 (32–48)
malnourished according to the NRS. Lymphocyte count (/ml) 1250 (900–1700)
C-reactive protein (mg/l) 24 (7–58)
GFR (Cockroft–Gault, ml/min) 52 (39–69)

Serum proteins Abbreviations: BMI, body mass index; GFR, glomerular filtration rate; MMS, mini-
Data of serum albumin, prealbumin, retinol-binding protein, mental status; MNA, mini-nutritional assessment; NRS, nutritional risk score.
Values are median (with interquartile ranges in brackets) or numbers (%),
blood lymphocytes and C-reactive protein concentrations respectively.
are listed in Table 1. The values for the normal range
especially of prealbumin and retinol-binding protein vary
according to the assay condition. Using a lower reference
limit for prealbumin of 160 mg/l (Vellas et al., 2000), 43 Table 2 Cross-classification
patients had prealbumin concentrations within the normal
NRS
range; with a lower limit of 200 mg/l (in-house laboratory
lower limit; Brugler et al., 2002); however, only 15 patients Severe risk Low risk Row total
were within the normal range. Using a lower limit for
retinol-binding protein of 30 mg/l (Corti et al., 1994), 86 MNA
Poor 11 12 23
patients were within the normal range, whereas with a
At risk 18 31 49
lower limit of 60 mg/l (in-house laboratory, unpublished Normal 6 26 32
data), only 11 patients had concentrations within the
normal range. Column total 35 69 104
Markers of protein malnutrition were similar in the Abbreviations: MNA, mini-nutritional assessment; NRS, nutritional risk score.
different groups of MNA scores (Figure 1a). Specifically, the Number of patients classified into three or two risk categories according to
group with the lowest MNA score o17 (n ¼ 23), indicating a MNA and NRS, respectively.
poor nutritional state, had a median serum albumin of 30 g/l

European Journal of Clinical Nutrition

 Malnutrition screening in geriatric patients
T Drescher et al
890
Prealbumin Retinol-binding Protein
400 100
P = 0.53 P = 0.90

300 75

mg/I

mg/I
200 50

100 25

0 0
MNA <17 MNA 17-23.5 MNA >23.5 MNA <17 MNA 17-23.5 MNA >23.5

Albumin Lymphocyte count

P = 0.33
50 P = 0.06 3000

40
2000
30

/uI
g/I

20
1000
10

0 0
MNA <17 MNA 17-23.5 MNA >23.5 MNA <17 MNA 17-23.5 MNA >23.5

Prealbumin Retinol-binding Protein

400 P = 0.06 100 P < 0.01

300 75
mg/I

mg/I

200 50

100 25

0 0
NRS 1-2 NRS 3-7 NRS 1-2 NRS 3-7

Albumin Lymphocyte count

P = 0.45 P = 0.96
50 3000
40
2000
30
g/I

/uI

20
1000
10

0 0
NRS 1-2 NRS 3-7 NRS 1-2 NRS 3-7

Figure 1 MNA (a) and NRS (b) in its three or two outcomes, respectively, referring to serum prealbumin, retinol-binding protein, albumin and
blood lymphocytes. Data are shown as medians with scatterplots representing all values. Shaded areas denote in-house normal ranges.

(IQR: 27–32), a median prealbumin of 156 mg/l (IQR: 92–185)
and a median retinol-binding protein of 37 mg/l (IQR:
29.5–47.0). In the intermediate risk group with an MNA
score of 17–23.5 (n ¼ 50), median serum albumin was 30 g/l
(IQR: 28–33), median prealbumin 143 mg/l (IQR: 101–174)
and mean retinol-binding protein 37.5 mg/l (IQR: 34.0–46.5).
The group with MNA423.5 (n ¼ 31) had a median
serum albumin concentration of 32 g/l (IQR: 29.5–35.0), a
median prealbumin of 151 mg/l (IQR: 110–192) and a
median retinol-binding protein of 38 mg/l (IQR: 31.0–50.5).
No significant difference could be found between these
groups. Similarly, the lymphocyte count was not different in
the three groups of MNA scores (Figure 1a).
Markers of protein malnutrition prealbumin and retinol-
binding protein gradually declined with increasing malnu-
trition as indicated by the NRS (Figure 1b). In patients with a
high risk of malnutrition (NRS: 3–7; n ¼ 35), median serum
albumin was 31 g/l (IQR: 26–33), median prealbumin

European Journal of Clinical Nutrition


138 mg/l (IQR: 84–172) and median retinol-binding protein
34 mg/l (IQR: 30–41). In the group with a low risk of
malnutrition (NRS: 1–2; n ¼ 69), the results were for albumin
31 g/l (IQR: 28–33), prealbumin 156 mg/l (IQR: 110–185) and
retinol-binding protein 42 mg/l (IQR: 34–51). The respective
P-values for serum albumin, prealbumin and retinol-binding
protein were 0.45, 0.06 o0.01. Lymphocyte count was
similar in both groups of NRS scores.
Prealbumin was significantly correlated with serum albu-
min (r ¼ 0.51, Po0.0001), with retinol-binding protein
(r ¼ 0.72, Po0.0001), lymphocyte count (r ¼ 0.25, Po0.05)
and inversely with C-reactive protein (r ¼ 0.55, Po0.0001).
There was no correlation with GFR, BMI, upper-arm and calf
circumference. Retinol-binding protein correlated with ser-
um albumin (r ¼ 0.39, Po0.0001), lymphocyte count
(r ¼ 0.26, Po0.01) and inversely with C-reactive protein
(r ¼ 0.43, Po0.0001) and likewise with GFR (r ¼ 0.3,
P ¼ 0.002). Retinol-binding protein did not correlate with
the anthropometric parameters BMI, upper-arm and calf
circumference. Serum albumin did not correlate with GFR.
In the univariate analysis, the association of C-reactive
protein and GFR with NRS did not reach significance
(P ¼ 0.11 and 0.06, respectively). There was also no associa-
tion of CRP with MNA (P ¼ 0.9), but a borderline significant
association of GFR with MNA (P ¼ 0.06).
Patients with and without inflammatory diseases and renal
dysfunction
The subgroup of patients without laboratory evidence of a
systemic inflammatory condition (C-reactive protein
o10 mg/l (n ¼ 33)) did not differ from the whole patient
cohort with respect to demographic and anthropometric
parameters. In multivariable for C-reactive protein and GFR-
adjusted models, retinol-binding protein was highly asso-
ciated with the NRS (P ¼ 0.002), whereas prealbumin and
albumin had no association with the NRS. The same markers
showed a different pattern for association with the MNA.
Prealbumin was not associated with the MNA, but albumin
and retinol-binding protein independently predicted MNA
scores (P ¼ 0.03 and 0.04, respectively).
Patients with low and high BMI and calf circumference,
respectively
A total of 24 patients had a BMIo20 kg/m2, 62 patients
were between 20 and 30 kg/m2 and 18 patients had a
BMI430 kg/m2. The median MNA score in the subgroup
with BMIo20 kg/m2 was 15.75 (IQR: 13.5–18.1), for BMI
20–30 kg/m2, median MNA score was 22 (IQR: 18.1–24.5)
and for BMI430 kg/m2, median MNA score was 25
(IQR: 21–26.4) (Po0.0001). The median NRS scores in the
three BMI groups were 2 (IQR: 1.75–3), 2 (IQR: 1–3) and 1
(IQR: 1–2), respectively (P ¼ 0.005).
The calf circumference was obtained in all 104 patients, 67
were above the threshold of 30.5 cm, 37 patients were below.
Malnutrition screening in geriatric patients
T Drescher et al
891
With the exception of albumin, which was significantly
higher in patients with high calf circumference (P ¼ 0.03), all
laboratory parameters were similar in both groups, whereas
both MNA and NRS scores were higher in patients with
higher compared with lower calf circumference (Po0.0001,
P ¼ 0.001, respectively).
Discussion
The main finding of this study performed in newly admitted
patients of an academic acute geriatric ward was that,
although NRS and MNA were highly associated with each
other, the NRS identified more patients with or at risk of
malnutrition than did the MNA. We identified the retinol-
binding protein as the only serum protein studied with a
consistent association with both NRS and MNA scores. We
conclude, therefore, that in acutely ill geriatric inpatients,
the NRS might be the appropriate tool for identifying most
of the patients with or at risk of malnutrition, and, therefore,
better reflects malnutrition than MNA scores or the studied
serum protein levels.
The use of the MNA has been propagated in geriatric
patients (Kondrup et al., 2003a), and it has been repeatedly
used in acute hospitalized geriatric patients and in long-term
care (Hudgens and Langkamp-Henken, 2004; Soini et al.,
2004). However, it has further drawbacks, as it is rather time
consuming (approximately 20 min in the present inpatient
population). In addition and in agreement with an earlier
report (Bauer et al., 2005), it could not be completed in a
substantial part of patients because of cognitive impairment
and communication problems.
The absence of a correlation between MNA and serum
prealbumin was in agreement with one (Langkamp-Henken
et al., 2005), but in contrast to another earlier report (Vellas
et al., 2000). However, this latter study included younger and
healthier outpatients than our frail inpatients, which may
explain the different findings.
The calf circumference represents an anthropometric
parameter of muscle mass; it reflects disability and self-
reported physical function (Rolland et al., 2003) and has
been correlated with serum albumin concentrations in
hospitalized older patients (Bonnefoy et al., 2002). We found
that the calf circumference correlated with both, MNA and
NRS, suggesting that the calf circumference represents a valid
parameter of malnutrition. This simple parameter—similar
to BMI—is probably at least as good as more sophisticated
anthropometric or bioelectric measurements in estimating
nutritional status (Omran and Morley, 2000a).
Biochemical markers are an attractive option in assessing
the nutritional status because they are easy to be determined
and to be standardized in clinical practice. A traditional
marker is serum albumin, which, however, is thought to
reflect inflammation rather than nutrition (Don and
Kaysen, 2004). In addition, its value is limited by the long
half-life (14–20 days), hydration, posture, hepatic and renal
European Journal of Clinical Nutrition
 Malnutrition screening in geriatric patients
T Drescher et al
892
impairment, and possibly ageing itself (Omran and Morley,
2000b). The usage of serum albumin levels in frail elderly
even without inflammation has been questioned (Kuzuya
et al., 2007). Nevertheless, it has considerable prognostic
impact (Phillips et al., 1989; Ferguson et al., 1993; Corti et al.,
1994).
Prealbumin has a shorter half-life (2–3 days) and is,
therefore, assumed to be a more sensitive marker for acute
nutritional changes. However, its function in nutritional
screening and in predicting mortality is controversial
(Ferguson et al., 1993; Brugler et al., 2002; Langkamp-Henken
et al., 2005; Devoto et al., 2006; Carriere et al., 2008).
Robinson et al. (2003) found prealbumin to be a significant
predictor of malnutrition, whereas serum albumin and
retinol-binding protein were not. Retinol-binding protein
behaves similarly to prealbumin, but has an even shorter
half-life (6–12 h). In our study, prealbumin levels showed a
more gradual decline with increasing NRS as compared with
serum albumin levels. Similar to serum albumin and retinol-
binding protein, prealbumin levels are influenced by
systemic inflammatory diseases, limiting their use in acutely
ill patients (Lopez-Hellin et al., 2002; Devoto et al., 2006).
This might explain why prealbumin levels in our study
declined with increasing NRS scores, which were more
influenced by comorbidities compared with the MNA score.
Total lymphocyte count has been proposed as a useful
indicator of nutritional state, decreasing with progressive
malnutrition and being an indicator of a poor prognosis
(Omran and Morley, 2000b). However, we found no correla-
tion between lymphocyte count and MNA or NRS, in
agreement with Kuzuya et al. (2005).
The cross-classification of MNA and NRS revealed sub-
stantial discrepancies in classification of the patients. A
similar analysis was performed by Bauer et al. (2005), who
only distinguished between the two categories as normal and
malnourished/at risk. They found that in approximately 50%
of all observations, MNA and NRS resulted in discrepant
classifications. In contrast to our results, the MNA classified a
much greater percentage of patients as at risk or mal-
nourished than the NRS did. In our cross-classification,
analysed in the same matter, that is using only two
categories, only about one-third of all observations showed
disagreeing classifications. Moreover, off-diagonal classifica-
tions were similarly distributed with both scores. The NRS
tends to classify more patients as malnourished or at risk
than the MNA. The impact of acute disease on the NRS is
certainly the most important factor.
A recent study from Brazil compared the NRS and MNA for
predicting clinical outcomes in hospitalized patients (Raslan
et al., 2010). In 169 elderly patients, the MNA classified
much more patients as at risk of malnutrition than the NRS
did (73 vs 42%, respectively), but the NRS was equal in
predicting complications, prolonged hospital stay and death.
The main limitation of our study is the lack of a gold
standard for the diagnosis of malnutrition in general.
Therefore, we could only compare two assessment scores,
European Journal of Clinical Nutrition
with anthropometric and biochemical parameters reflecting
different aspects of malnutrition (Covinsky et al., 2002;
Soeters et al., 2008).
In conclusion, in our cohort, the NRS reflected the degree
of biochemical protein malnutrition better than the MNA,
even in the presence of an acute inflammatory disease. As
acute disease is one of the most important risk factors for
malnutrition, especially in hospitalized patients, it is a
strength of the NRS to strongly weight acute disease. In
addition, the NRS was feasible in nearly all patients, whereas
the MNA was not applicable in 11% of the patients. On the
other hand, the MNA included a scoring of the patient’s
functional state. Therefore, the two methods have their own
advantages and disadvantages, the NRS being superior in
identifying hospitalized geriatric patients who are at risk of
malnutrition because of acute disease.
Conflict of interest
The authors declare no conflict of interest.
Acknowledgements
TD was responsible for data analysis and drafted the manu-
script. KS developed the study design and data collection. AU
was involved in the data collection. MK assisted with
statistical analysis. UK contributed to critical revision of
the manuscript. MCC coordinated the data analysis and
significantly supported the drafting of the manuscript. RWK
helped with the study design and the revision of the
manuscript. All authors have read and approved the final
manuscript.
References
Bauer J, Vogl T, Wicklein S, Trögner J, Mühlberg W, Sieber C (2005).
Comparison of the mini nutritional assessment, subjective global
assessment, and nutritional risk screening (NRS 2002) for nutri-
tional screening and assessment in geriatric hospital patients.
Z Gerontol Geriat 38, 322–327.
Bonnefoy M, Jauffret M, Kostka T, Jusot J (2002). Usefulness of calf
circumference measurement in assessing the nutritional state of
hospitalized elderly people. Gerontol 48, 162–169.
Brugler L, Stankovic A, Bernstein L, Scott F, O’Sullivan-Maillet J
(2002). The role of visceral protein markers in protein calorie
malnutrition. Clin Chem Lab Med 40, 1360–1369.
Carriere I, Dupuy A, Lacroux A, Cristol J, Delcourt C (2008).
Biomarkers of inflammation and malnutrition associated with
early death in healthy elderly people. J Am Geriatr Soc 56, 840–846.
Cereda E, Valzolgher L, Pedrolli C (2008). Mini nutritional assess-
ment is a good predictor of functional status in institutionalised
elderly at risk of malnutrition. Clin Nutr 27, 700–705.
Cockroft D, Gault M (1976). Prediction of creatinine clearance from
serum creatinine. Nephron 16, 31–41.
Corti M, Guralnik J, Salive M, Sorkin J (1994). Serum albumin level
and physical disability as predictors of mortality in older persons.
JAMA 272, 1036–1042.

Covinsky KE, Covinsky MH, Palmer R, Sehgal A (2002). Serum
albumin concentration and clinical assessments of nutritional
status in hospitalized older people: different sides of different
coins? J Am Geriatr Soc 50, 631–637.
Devoto G, Gallo F, Marchello C, Racchi O, Garbarini R, Bonassi S et al.
(2006). Prealbumin serum concentration as a useful tool in the
assessment of $ malnutrition in hospitalized patients. Clin Chem
52, 2281–2285.
Don BR, Kaysen G (2004). Serum albumin: relationship to inflam-
mation and nutrition. Semin Dial 17, 432–437.
Ferguson R, O’Connor P, Crabtree B, Batchelor A, Mitchell J, Coppola
D (1993). Serum albumin and prealbumin as predictors of clinical
outcomes of hospitalized elderly nursing home residents. J Am
Geriatr Soc 41, 545–549.
Flodin L, Svensson S, Cederholm T (2000). Body mass index as a
predictor of 1-year mortality in geriatric patients. Clin Nutr 19,
121–125.
Folstein MF, Folstein S, Mc Hugh P (1975). ‘Mini-mental state’: a
practical method for grading the cognitive state of patients for the
clinician. J Psychiatr Res 12, 189–198.
Guigoz Y (2006). The mini nutritional assessment (MNA) review
of the literature—what does it tell us? J Nutr Health Aging 10,
466–487.
Guigoz Y, Vellas B, Garry P (1994). Mini nutritional assessment: a
practical assessment tool for grading the nutritional state of
elderly patients. Facts Res Gerontol 4 (Suppl 2), 15–59.
Hudgens J, Langkamp-Henken B (2004). The mini nutritional
assessment as an assessment tool in elders in long-term care.
Nutr Clin Pract 19, 463–470.
Kondrup J, Allison S, Elia M, Plauth M (2003a). ESPEN-guidelines for
nutrition screening 2002. Clin Nutr 22, 415–421.
Kondrup J, Rasmussen H, Hamberg O, Stanga Z (2003b). Nutritional
risk screening (NRS 2002): a new method based on an analysis of
controlled clinical trials. Clin Nutr 22, 321–336.
Kuzuya M, Izawa S, Enoki H, Okada K, Iguchi A (2007). Is serum
albumin a good marker for malnutrition in the physically
impaired elderly? Clin Nutr 26, 84–90.
Kuzuya M, Kanda S, Koike T, Suzuki Y, Iguchi A (2005). Lack of
correlation between total lymphocyte and nutritional status in the
elderly. Clin Nutr 24, 427–432.
Kyle UG, Kossovsky M, Karsegard V, Pichard C (2006). Comparison of
tools for nutritional assessment and screening at hospital admis-
sion: a population study. Clin Nutr 25, 409–417.
Langkamp-Henken B, Hudgens J, Stechmiller J, Herrlinger-Garcia K
(2005). Mini nutritional assessment and screening scores are
associated with nutritional indicators in elderly people with
pressure ulcers. J Am Diet Assoc 105, 1590–1596.
Lopez-Hellin J, Baena-Fustegueras J, Schwartz-Riera S, Garcia-Arumi E
(2002). Usefulness of short-lived proteins as nutritional indicators
in surgical patients. Clin Nutr 21, 119–125.
McWhirter J, Pennington C (1994). Incidence and recognition of
malnutrition in hospital. BMJ 308, 945–948.
Malnutrition screening in geriatric patients
T Drescher et al
893
Mühlethaler R, Stuck A, Minder C, Frey B (1995). The prognostic
significance of protein-energy malnutrition in geriatric patients.
Age Ageing 24, 193–197.
Norman K, Pichard C, Lochs H, Pirlich M (2008). Prognostic impact
of disease-related malnutrition. Clin Nutr 27, 5–15.
Ödlund Olin A, Koochek A, Ljungqvist O, Cederholm T (2005).
Nutritional status, well-being and functional ability in frail elderly
service flat residents. Eur J Clin Nutr 59, 263–270.
Omran M, Morley J (2000a). Assessment of protein energy malnutri-
tion in older persons, part I: history, examination, body composi-
tion, and screening tools. J Nutr 16, 50–63.
Omran M, Morley J (2000b). Assessment of protein energy malnutri-
tion in older persons, part II: laboratory evaluation. Nutr 16,
131–140.
Persson M, Brismar K, Katzarski K, Nordenström J, Cederholm T
(2002). Nutritional status using mini nutritional assessment and
subjective global assessment predict mortality in geriatric patients.
J Am Geriatr Soc 50, 1996–2002.
Phillips A, Shaper A, Whincup E (1989). Association between serum
albumin and mortality from cardiovascular disease, cancer, and
other causes. Lancet 2, 1434–1436.
Raslan M, Gonzales M, Goncalves Dias M, Nascimento M, Castro M,
Marques P et al. (2010). Comparison of nutritional risk screening
tools for predicting clinical outcomes in hospitalized patients.
Nutr (in press).
Robinson MK, Trujillo E, Mogensen K, Rounds J, McManus K,
Jacobs D (2003). Improving nutritional screening of hospitalized
patients: the role of prealbumin. J Parenter Enteral Nutr 27,
389–395.
Rolland Y, Lauwers-Cances V, Cournot M, Nourhashemi F, Reynish
W, Riviere D et al. (2003). Sarcopenia, calf circumference, and
physical function of elderly women: a cross-sectional study. J Am
Geriatr Soc 51, 1120–1124.
Sergi G, Coin A, Enzi G, Volpato S, Inelmen E, Buttarello M et al.
(2006). Role of visceral proteins in detecting malnutrition in the
elderly. Eur J Clin Nutr 60, 203–209.
Soeters PB, Reijven P, van Bokhorst-de van de Schueren M, Schols J,
Halfens R, Meijers J et al. (2008). A rational approach to nutritional
assessment. Clin Nutr 27, 706–716.
Soini H, Routasalo P, Lagström H (2004). Characteristics of the mini-
nutritional assessment in elderly home-care patients. Eur J Clin
Nutr 58, 64–70.
Sorensen J, Kondrup J, Prokopowicz J, Schiesser M, Krähenbühl L,
Meier R et al. (2008). EuroOOPS: an international, multicentre
study to implement nutritional risk screening and evaluate
clinical outcome. Clin Nutr 27, 340–349.
Sullivan DH (1995). The role of nutrition in increased morbidity and
mortality. Clin Geriatr Med 11, 661–674.
Vellas B, Guigoz Y, Baumgartner M, Garry P, Lauque S, Albarede J-L
(2000). Relationships between nutritional markers and the mini-
nutritional assessment in 155 older persons. J Am Getriatr Soc 48,
1300–1309.
European Journal of Clinical Nutrition