Clinical Nutrition 30 (2011) 793e798

Contents lists available at ScienceDirect

Clinical Nutrition
journal homepage: http://www.elsevier.com/locate/clnu

Original article

Nutritional screening and mortality in newly institutionalised elderly: A

comparison between the Geriatric Nutritional Risk Index and the Mini Nutritional
Assessment
Emanuele Cereda a, *, Carlo Pedrolli b, Annunciata Zagami c, Alfredo Vanotti d, Silvano Piffer e,
Annalisa Opizzi f, Mariangela Rondanelli f, Riccardo Caccialanza a
a

Nutrition and Dietetics Service, Fondazione IRCCS Policlinico San Matteo, Viale Golgi 19, 27100 Pavia, Italy
Unit Operativa di Dietetica e Nutrizione Clinica, Ospedale S. Chiara, Azienda Provinciale per i Servizi Sanitari, Trento, Italy
Fondazione Bellaria Onlus, Appiano Gentile, Como, Italy
Servizio di Dietetica e Nutrizione Clinica, ASL Como, Como, Italy
e
Servizio Osservatorio Epidemiologico, Direzione per la Promozione e lEducazione alla Salute, Azienda Provinciale per i Servizi Sanitari, Trento, Italy
f
Servizio Endocrino-nutrizionale, Dipartimento di Scienze Sanitarie Applicate e Psicocomportamentali, Sezione di Nutrizione, Azienda di Servizi alla Persona di Pavia, Universit degli
Studi di Pavia, Pavia, Italy
b
c

a r t i c l e i n f o

s u m m a r y

Article history:
Received 27 January 2011
Accepted 27 April 2011

Background & aims: Several tools are available for nutritional screening. We evaluated the risk of
mortality associated with the Geriatric Nutritional Risk Index (GNRI) and the Mini Nutritional Assessment (MNA) in newly institutionalised elderly.
Methods: A prospective observational study involving 358 elderly newly admitted to a long-term care
setting. Hazard ratios (HR) for mortality among GNRI categories and MNA classes were estimated by
multivariable Coxs model.
Results: At baseline, 32.4% and 37.4% of the patients were classied as being malnourished (MNA <17)
and at severe nutritional risk (GNRI <92), respectively, whereas 57.5% and 35.2%, respectively, were
classied as being at risk for malnutrition (MNA 17e23.5) and having low nutritional risk (GNRI 92e98).
During a median follow-up of 6.5 years [25the75th percentile, 5.9e8.6], 297 elderly died. Risk for allcause mortality was signicantly associated with nutritional risk by the GNRI tool (GNRI<92
HR 1.99 [95%CI, 1.38e2.88]; GNRI 92e98 HR 1.51 [95%CI, 1.04e2.18]) but not with nutritional status
by the MNA. A signicant association was also found with cardiovascular mortality (GNRI <92 HR 1.79
[95%CI, 1.23e2.61]).
Conclusions: Nutritional risk by GNRI but not nutritional status by MNA was associated with higher
mortality risk. Present data suggest that in the nutritional screening of newly institutionalised elderly the
use of the GNRI should be preferred to that of the MNA.
! 2011 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.

Keywords:
Geriatric nutritional risk index (GNRI)
Mini nutritional assessment (MNA)
Elderly
Mortality
Outcome
Long-term care

1. Introduction
Malnutrition in long-term cares is reported to affect 60e80% of
residents.1e3 Nutritional derangements are of multifactorial origin
and aging, along with multiple comorbidities and related treatments, as well as poor nutritional care are now considered the most
important determinants in this setting.1,4,5 The importance of
malnutrition relates to the negative effects on the patients prognosis and the possibility to prevent complications by means of

* Corresponding author. Tel.: 39 0382 501615; fax: 39 0382 502801.

E-mail address: e.cereda@smatteo.pv.it (E. Cereda).

targeted nutritional intervention.5e7 Accordingly, early recognition

and treatment are now reccommended by existing guidelines and
nutritonal management should be included in the routine care of
every patients.7,8 Nowadays, several screening tools have been
introduced in clinical practice to identify patients candidate to
nutritional treatment.8,9 Among those available for the evaluation
of the institutionalised old patient, the use of the Mini Nutritional
Assessment is currently recommended by the European Society for
Clinical Nutrition and Metabolism (ESPEN).8 The advantages of
using this tool lie in the high sensitivity, the reliability among
different operators, its easiness that allows being performed by not
specialised staff in a short time, the low participation required to
the patient and in the capacity to address different geriatric-specic

0261-5614/$ e see front matter ! 2011 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.
doi:10.1016/j.clnu.2011.04.006

794

E. Cereda et al. / Clinical Nutrition 30 (2011) 793e798

conditions (nutritional and health status, independence, quality of

life, cognition, mobility and subjective health).10 However, a recent
literature review suggested a disadvantage in the limited prediction
of outcome, particularly of mortality.11
A new tool, the Geriatric Nutritional Risk Index (GNRI), has been
recently proposed.12 Previous studies have supported its use in
reason of the signicant association with most nutritional parameters and both short and long-term outcomes.13e16 Moreover, based
on objective parameters that can be collected even in non cooperative patients it could provide reliable assessments and appears
feasible in every healthcare setting.16,17
As the validation of a screening tool to clinical use depends on
the association with outcome as well as on the comparison with
other commonly applied tools we designed the present study. The
objective of our research will be the association of the GNRI with
both nutritional variables and mortality when compared to the
MNA. In regard with this, attention was focused only on newly
institutionalised elderly as previous studies have suggested that
prevalence and determinants of malnutrition may be different
according to setting and the patients provenience.16,18
2. Materials and methods
2.1. Study design

#24, good nutritional status.10 The tool was generally completed by

the same well-trained dietitians because only in few cases a direct
participation by the patient was possible. Tube feeding (n 4) was
not counted as a prescription drug use and allowed assigning the
highest scores to the questions concerning the number of meals,
protein and fruit/vegetables intakes, uid consumption, appetite and
mode of feeding.
Risk of nutrition-related health complications was assessed by
the GNRI. This tool requires the evaluation of serum albumin and
the calculation of ideal body weight by Lorentzs formula. The score
is derived according to the following equation:

GNRI 1:489 % albumin; g=L 41:7

% weight=ideal body weight

When weight exceeded ideal body weight, a score 1 is

assigned the weight/ideal body weight ratio. Then, patients were
categorized as being: at severe/moderate risk, GNRI <92; at low
risk, GNRI 92e98; at no risk, GNRI >98.12,16
The GNRI was rst validated and introduced in clinical practice
in 2005 (three years after the start of the study).12 Therefore the
GNRI has been retrospectively derived for those patients included
before the 2005 and prospectively calculated for those entered after
its rst description.

In the present research paper we analysed the data pertaining

a prospective cohort study designed in 4 long-term care facilities in
the provinces of Como (n 1), Pavia (n 1) and Trento (n 2). The
recruitment phase started in May 2002 and ended in May 2007.
Further information on the present study are detailed elsewhere.19
In brief, every year, for 2 weeks, all the subjects newly admitted to
the facilities, aged >65 years and agreeing to participate were
assessed for eligibility. Patients were excluded in case of established terminal diseases and known neoplastic disorder.

2.3. Other covariates

2.2. Nutritional assessment

2.4. Outcome

At baseline,
parameters:

patients

were

assessed

for

the

following

2.2.1. Anthropometry
Body weight (by a calibrated at scale or a chair scale or a hoist
provided weighting device), height, knee-height, mid arm circumference and triceps skinfold thickness were measured in agrseement with established procedures.20 Height was estimated from
knee-height in non-ambulatory patients or in case of abnormal
spinal curvature.21 The body mass index (BMI) and the arm muscle
area were calculated according to standard equations.20,22
2.2.2. Biochemistry
Eight to 12-h fasting venous blood samples were also drawn for
the evaluation of haemoglobin, total lymphocytes count, serum
albumin, transthyretin, transferrin, total cholesterol and creatinine.
2.2.3. Nutritional screening tools
The MNA was used to grade the nutritional status. The tool is
made of 18 questions grouped in 4 rubrics: anthropometry (BMI,
weight loss, mid-upper arm and calf circumference), general state
(medications, mobility, presence of pressure ulcers, lifestyle, presence of psychological stress or neuropsychological problems), dietary assessment (autonomy of feeding, quality and number of meals,
uid intake) and self perception about health and nutrition. The total
score ranges from 0 to 30 points. Nutritional status is dened
according to the following threshold values: MNA <17, proteincalorie malnutrition; MNA 17e23.5, at risk of malnutrition; MNA

After the consultation of medical records or having the patients

interviewed and physically examined, other information collected
at baseline were: age, gender, admission diagnosis, major comorbidities (diabetes and hypertension). Main admission diagnosis
groups used in the analyses were selected on the basis of the
literature, in reason of an established association with nutritional
status and/or survival.

Vital status (up to November 30, 2010) was ascertained by active

follow-up (Como and Pavia) or by means of record linkage with
centralized provincial registry (Trento). Causes of death were coded
according to the International Classication of Diseases, 10th
Revision (ICD-10) and classied as cardiovascular (CV), respiratory,
neoplasms or others/not reported.
The study was performed in adherence to the principles of the
Declaration of Helsinki and protocol was approved by local Institutional Ethics Commettee.
2.5. Statistical analyses
Data were presented as mean and standard deviation, median
and interquartile range [25the75th percentile] or counts and
pecentage. Comparisons of groups were performed using one way
ANOVA followed by post-hoc test (continuous variables) or chisquare test (categorical variables).
A receiver operating characteristic (ROC) curve was initially
generated for each nutritional screening tool; the area under the
ROC curve (AUC) indicated the probability of predicting the
outcome. The cutoff risk point of nutrition for each tool was then
dened from the highest sensitivity - (1 - specicity) value in the
ROC curve. Then, associations between non collinear (Pearsons
statistic) variables and all-cause and cause-specic mortality were
assessed by the Cox proportional hazards regression model. Only
variables with a P < .25 at univariate analysis were retained in the
nal multivariable models. Results were presented as hazard ratio
and 95% condence interval (HR, [95%CI]).

795

E. Cereda et al. / Clinical Nutrition 30 (2011) 793e798

Statistical analyses were performed using MEDCALC" for

Windows Version 11.3.0.0 (MedCalc Software, Mariakerke,
Belgium). A two-sided P-value <0.05 was considered statistically
signicant.
3. Results
In total, 358 patients of the population initially assessed were
included in the present analyses (n 14, lost to follow-up; n 161,
excluded for incomplete data on MNA score). The features of the
present cohort (Table 1) were almost similar to those of the original
population19 in terms of age, gender, anthropometry, biochemistry
and prevalence of main admission diagnoses.
Provenience of patients were: hospital (76.9%), home (19.2%)
and other institution (3.9%).
According to the MNA, 32.4% and 57.5% of the patients were
classied as being malnourished (MNA <17) and at risk for
malnutrition (MNA 17e23.5), respectively, whereas using the GNRI
tool, 37.4% and 35.2% of the population were scored at severe (GNRI
<92) and at low nutritional risk (GNRI 92e98), respectively. The
features of the population by nutritional status (MNA) and nutritional risk (GNRI) are presented in the Table 1. Both the MNA and
the GNRI were associated with most of nutrition-related variables.
During a median follow-up of 6.5 years [25the75th percentile,
5.9e8.6], 297 (83%) elderly died. The most frequent causes were
those CV (48.9%; coronary artery disease 23.4%, heart failure 38.3%,
stroke 30.3%, others or unspecied 8.0%). The other events were
due to respiratory diseases (15.1%, 87% by pneumonia), neoplasm
(4.5%), or micellaneous causes (31.5%). The baseline features of
survivors and dying patients are summarized in the Table 2.
Particularly, surviving patients presented with higher scores of
both MNA and GNRI. However, only grading for nutritional risk
appeared associated with fatal outcome. In agreement with this,
ROC analyses conrmed that the highest accuracy for outcome
prediction by the GNRI was found for a cutoff '98,7 (AUC 0.627
[95%CI, 0.574e0.677]; sensitivity, 0.811 [95%CI, 0.762e0.854];
specicity, 0.426 [95%CI, 0.300e0.559]), which is similar to the
reccommended threshold value for excluding nutritional risk.12

Table 2
Baseline clinical and demographic characteristics of the population by outcome.
Characteristic

Male, %
Age, mean (SD), years
Body mass index, mean (SD), kg/m2
GNRI, mean (SD)
<92, %
92e98, %
MNA, mean (SD)
<17,%
17e23.5, %
Arm muscle area, mean (SD), cm2
Albumin, mean (SD), g/L
Transthyretin, mean (SD), mg/dL
Transferrin, mean (SD), mg/dL
Haemoglobin, mean (SD), g/dL
Lymphocytes count, mean (SD),/mm3
Total cholesterol, mean (SD), mg/dL
Creatinine, mean (SD), mg/dL
All-type dementia, %
COPD, %
Heart disease, %
Hip fracture, %
Stroke, %
Other admission diagnoses, %
Diabetes, %
Hypertension, %

Dead
(n 297)

27.9
86.1 (7.5)
23.6 (4.5)
92.8 (7.4)
40.1
36.0
18.5 (4.1)
34.3
55.9
36.5 (16.2)
35.7 (4.2)
19.6 (7.8)
198 (40)
12.1 (2.5)
1892 (811)
183 (42)
1.12 (.56)
42.5
4.4
21.2
12.1
11.4
25
8.8
21.7

Survivors
(n 61)

13.1
77.3 (8.2)
24.9 (4.8)
96.1 (6.8)
24.6
31.1
19.8 (3.8)
23.0
65.6
35.0 (10.4)
37.4 (4.0)
21.5 (5.0)
212 (41)
13.0 (1.4)
1919 (584)
223 (51)
1.02 (.24)
45.9
0
6.6
0
9.8
23
4.9
8.2

P Value

0.015
<0.001
0.043
0.001
0.004
0.024
0.223
0.538
0.004
0.043
0.041
0.026
0.869
0.009
0.515
0.671
0.136
0.011
0.004
0.826
<0.001
0.325
0.006

Abbreviations: GNRI, Geriatric Nutritional Risk Index; MNA, Mini Nutritional

Assessment; COPD, chronic obstructive pulmonary disease.
a
Data are reported as mean (SD) or percentages (%). Percentages are calculated
within single groups.
b
Continuous and categorical variables were compared between groups with one
way ANOVA or the Fisher exact test, respectively.

However, according to the MNA the best accuracy was found for
a cutoff '19 (AUC 0.593 [95%CI, 0.540e0.645], P 0.436 vs. GNRI
tool; sensitivity, 0.546 [95%CI, 0.487e0.603]; specicity, 0.672 [95%
CI, 0.540e0.787]).
Accuracy (AUC) of the established cutoffs for MNA and GNRI were:
MNA <17, 0.557 [95%CI, 0.504e0.609] (sensitivity, 0.317 [95%CI,

Table 1
Baseline clinical and demographic characteristics of the population by Geriatric Nutritional Risk Index and Mini Nutritional Assessment categories.
Characteristic

Male, %
Age, mean (SD), years
Body mass index, mean (SD), kg/m2
GNRI, mean (SD)
MNA, mean (SD)
Arm muscle area, mean (SD), cm2
Albumin, mean (SD), g/L
Transthyretin, mean (SD), mg/dL
Transferrin, mean (SD), mg/dL
Haemoglobin, mean (SD), g/dL
Lymphocytes count, mean (SD),/mm3
Total cholesterol, mean (SD), mg/dL
Creatinine, mean (SD), mg/dL
All-type dementia, %
COPD, %
Heart disease, %
Hip fracture, %
Stroke, %
Other admission diagnoses, %
Diabetes, %
Hypertension, %

Overall population
(n 358)

25.4
84.6 (8.2)
23.8 (4.0)
93.4 (6.8)
18.7 (2.0)
35.2 (2.9)
36.0 (2.5)
19.9 (6.9)
200 (39)
12.2 (2.3)
1896 (778)
186 (40)
1.11 (.55)
43.0
3.7
18.7
10.0
11.2
13.4
8.9
21.8

GNRI< 92
(n 134)

23.1
85.6 (8.1)
21.3 (4.2) c
85.9 (5.1) c
16.7 (4.1) c
31.1 (9.1) c
32.4 (3.3) c
15.7 (5.4) c
186 (46) c
11.7 (2.2) c
1801 (926)
168 (40) c
1.06 (.50)
34.3
3.7
22.4
15.7
1.2
12.7
11.2
28.4

GNRI 92e98
(n 126)

23.8
84.2 (8.0)
25.1 (4.3)
94.6 (1.7) c
19.2 (3.7) c
35.1 (9.6)
36.1 (1.8) c
21.6 (9.4)
200 (35) c
12.2 (2.4)
1868 (674)
194 (40)
1.10 (.38)
44.4
5.6
17.5
7.1
14.2
11.2
10.3
24.6

GNRI > 98
(n 98)

30.6
83.8 (8.8)
25.5 (4.1)
101.9 (3.2) c
20.8 (3.2) c
35.9 (13.2)
40.7 (2.2) c
22.3 (9.2)
214 (34) c
12.8 (2.4)
1995 (723)
211 (40)
1.27 (.83)
53.1
1.0
15.3
6.1
7.1
17.4
4.1
9.2

P Value
0.380
0.200
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
0.016
0.382
<0.001
0.109
0.016
0.197
0.357
0.023
0.242
0.379
0.137
0.002

MNA< 17
(n 116)

24.1
86.5 (7.6)
20.6 (3.7) c
89.2 (7.6) c
13.9 (2.4) c
30.1 (9.3) c
34.9 (4.1) c
18.1 (9.8) c
191 (45) c
12.2 (2.3)
1716 (677)
179 (41)
1.05 (.38) c
44.0
2.6
21.6
12.9
11.2
7.7
6.0
25.0

MNA 17e23.5
(n 206)

23.8
84.2 (8.2)
25.1 (4.2)
94.9 (6.3)
20.3 (1.8) c
37.2 (11.8)
36.4 (4.0)
20.8 (6.1)
205 (37)
12.2 (2.3)
1970 (835)
190 (45)
1.10 (.47)
45.6
3.4
16.0
6.3
10.2
18.5
8.7
18.9

Abbreviations: GNRI, Geriatric Nutritional Risk Index; MNA, Mini Nutritional Assessment.
a
Data are reported as mean (SD) or percentages (%). Percentages are calculated within single groups.
b
Continuous and categorical variables were compared between groups with one way ANOVA or the Chi-square test, respectively.
c
Signicantly different versus all the other groups by post-hoc test.

MNA #24
(n 36)

38.9
81.0 (9.1) c
26.6 (3.9)
97.6 (6.8)
25.1 (1.2) c
38.2 (9.4)
37.6 (4.6)
20.5 (4.8)
200 (42)
11.4 (3.3)
2038 (627)
189 (43)
1.37 (1.04)
25.0
8.3
25.0
22.2
16.7
2.8
7
10.2

P Value
0.149
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
0.021
0.026
0.334
0.119
0.220
0.036
0.068
0.263
0.282
0.006
0.640
0.004
0.047
0.294

796

E. Cereda et al. / Clinical Nutrition 30 (2011) 793e798

Table 3
Predictors of all-cause and cardiovascular mortality according to Cox proportional hazard regression models.
Characteristic

All-cause mortality
Hazard Ratio

Gender (Male)
Age (years)
GNRI categories
<92
92e98
>98
MNA classes
<17
17e23.5
#24
All-type dementia
COPD
Heart disease
Hip fracture
Stroke
Diabetes
Hypertension

(95% CI)

1.73 (1.34e2.23)
1.05 (1.03e1.07)
0.78 (0.67e0.90)
1.63 (1.21e2.18)
1.20 (1.00e1.62)
1 (reference)
0.83 (0.68e1.01)
1.21 (0.80e1.82)
0.87 (0.59e1.29)
1 (reference)
0.76 (0.57e1.01)
1.47 (.73e2.98)
1.19 (.84e1.69)
3.56 (2.27e5.58)
1.49 (1.00e2.32)
1.59 (.99e2.55)
2.04 (1.47e2.84)

Cardiovascular mortality
b

P-value

Hazard Ratio

<0.001
<0.001
<0.001c
0.001
0.023

1.72 (1.28e2.33)
1.03 (1.01e1.05)
0.71 (0.59e0.85)
1.99 (1.38e2.88)
1.51 (1.04e2.18)
1 (reference)
e
e
e
e
1.14 (0.84e1.76)
e
1.29 (.84e1.89)
2.50 (1.53e4.07)
1.63 (1.03e2.57)
1.41 (.85e2.35)
1.34 (.93e1.94)

0.059c
0.367
0.491
0.057
0.285
0.217
<0.001
0.047
0.054
<0.001

(95% CI)

P-value

Hazard Ratio

<0.001
0.006
<0.001
<0.001
0.029

1.53 (1.03e2.26)
1.03 (1.00e1.06)
0.71 (0.56e0.91)
1.79 (1.23e2.61)
1.28 (.83e2.01)
1 (reference)
e
e
e
e
0.83 (0.48e1.47)
e
1.65 (1.05e2.59)
3.01 (1.68e5.69)
2.76 (1.59e4.80)
1.52 (.86e2.70)
2.03 (1.33e3.08)

e
e
e
0.204
e
0.218
<0.001
0.036
0.181
0.117

(95% CI)

P-value
0.032
0.035
0.007
0.003
0.242

e
e
e
0.534
e
0.030
<0.001
<0.001
0.152
<0.001

Abbreviations: GNRI, Geriatric Nutritional Risk Index; MNA, Mini Nutritional Assessment; COPD, chronic obstructive pulmonary disease.
a
Risk of death at univariate analysis.
b
Risk of death at multivariable analysis (variables retained in the nal model were those having a P-value <0.25 at univariate analysis).
c
Linear increase in risk over categories assumed (checked with likelihood ratio test).

0.264e0.373]); MNA <24, 0.509 [95%CI, 0.455e0.561] (sensitivity,

0.902 [95%CI, 0.863e0.934]); GNRI <92, 0.577 [95%CI, 0.524e0.629]
(sensitivity, 0.401 [95%CI, 0.344e0.459]); GNRI '98, 0.602 [95%CI,
0.549e0.653] (sensitivity, 0.805 [95%CI, 0.755e0.848]).
Finally, multivariable Cox regression models were used to
compute HRs for all-cause and cause-specic mortality (Table 3).
Risk for all-cause mortality was signicantly associated only with
nutritional risk by the GNRI tool (GNRI <92 HR 1.99 [95%CI,
1.38e2.88], P < 0.001; GNRI 92e98 HR 1.51 [95%CI, 1.04e2.18],
P 0.029) but not with nutritional status by the MNA. Other
predictor of mortality were male gender, age, previous hip fracture
and stroke. Interestingly, and in agreement with previous ndings,15 the association of the GNRI with all-cause mortality was
superior to that of albumin. Particularly, univariate HRs for reference cut-offs of albumin (vs. no risk [>38 g/L]) were: <35 g/L (high
risk), HR 1.47 [95%CI, 1.12e1.94], P 0.006; 35e38 g/L (low risk),
HR 0.97 [95%CI, .72e1.29], P 0.818.
Risks for cause-specic mortality were also computed. A significant association was also found between GNRI and cardiovascular
mortality (GNRI <92 HR 1.79 [95%CI, 1.23e2.61], P 0.003). No
relationship was found with the other causes of death.
4. Discussion
With the present study, we conrmed that the prevalence of
nutritional derangements is consistently high in newly institutionalised elderly, being up to 90% and 70% by the MNA and the
GNRI, respectively. Both these tools showed a signicant association with the great part of anthropometric and biochemical
markers of nutritional status. However, the GNRI showed a higher
prognostic value. All these ndings could be reasonably discussed
in view of how the same tools are structured and the population
recruited as in non newly institutionalised elderly the MNA was
demonstrated a valuable tool signicantly associated with health
complications and mortality.14
The MNA is a multidimensional tool that addresses several
features of the patient, particularly anthropometry, dietetic habits
and the general state.10 Most of the questions on which is based are
likely to receive a low score in patients newly admitted to a rehabilitative or long-term care as this last is generally the destination
of those discharged from an acute care setting. Malnutrition is

a common comorbidity in hospitalised patients, signicantly

related to age and multimorbidity.1,6,23,24 Moreover, during the
hospital stay patients more frequently experience psychological
stress, develop pressure ulcers and lose weight due to severe
catabolic stress and/or caloric decits often persisting after
discharge.2,6,25 In agreement with this evidence and our ndings,
previous studies have reported lower prevalence of well-nourished
(MNA < 24) patients and lower mean MNA scores in post-acute
rehabilitative care setting than in nursing homes.14,18 Along with
this, it seems that in nursing home malnutrition is inversely associated with the duration of stay (unpublished data) with
well-nourished patients being likely to live longer and those
malnourished deceasing early.
However, as discussed in a recent review, it appears that as it is
structured, the GNRI is able to reect the patients course of disease
through the different settings. Based on serum albumin and the
discrepancy between ideal weight and real weight it may account
for both acute and chronic causes of malnutrition associated with
underlying disease and age-related factors.16 Albumin, as marker of
inammation, is able to mirror the severity of acute diseases26 and
previous studies have a shown a signicant association with inhospital mortality.27 However, this biochemical parameter could
also decrease due to prolonged impaired energy balance and
depleted protein mass stores28,29 which can frequently occur in
a post-acute setting.1e4 Moreover, although the GNRI formula was
structured to give a higher weight to albumin than to weight, also
the real-ideal weight factor has demonstrated to improve the
prognostic role of the tool.15 Given the general tendency of the
elderly to badly adapt to malnutrition and to lose lean body mass,
this ratio should be intended as nutritional reserve and capacity to
cope with diseases. Previous studies have demonstrated have found
that there is little difference between calculating the ideal weight
with the Lorentz formula or deriving it from the height and a BMI of
22 kg/m2.30,31 Similarly, there appears to be no effect on the calculation of the GNRI.31 Our recent study conrmed in a similar series
the association between BMI '21 kg/m2 and mortality.19 Moreover,
GNRI threshold values have been established to predict health
complications also according to the cutoff values for weight loss.12
Finally, it is intriguing that nutritional risk was signicantly
associated with cardiovascular mortality but not with other cause
of death, such as infectious diseases. In this respect, it is possible

E. Cereda et al. / Clinical Nutrition 30 (2011) 793e798

that the investigation of other causes has been unpowered by the

reduced number of events. However, the relationship between
nutritional risk and death due to cardiovascular diseases could be
probably explained in view of the interplay between changes in the
inammatory and hormonal status ([ in glucocorticoids and catecholamines) and the disregulation of the autonomic nervous
system occurring with malnutrition and aging.32e35 Indeed, this
last issue deserves further investigation.
Indeed, both the MNA and the GNRI tools offer the clinician the
advantages of being rapid and easy to perform and of providing
objective assessments, even in non cooperative patients.11,16 The
MNA does not require any biochemical test, and so any additional
cost, to be completed. However, serum albumin assessment is now
included in patients evaluation as routine examination and it is of
limited costs.
It is worth of mention that there is still confusion between an
index of nutritional status and one of nutritional risk as the second
takes into account nutritional status-related complications.12 The
use of screening tools should also be considered in view of their
association with outcome.
Recent comprehensive literature overviews have highlighted
that, although the MNA is very sensitive in identifying malnorished
patients, as dened by different reference methods,10 it shows
limitations in predicting outcome.11 In this respect, the GNRI
appears to work better.14,16
A high sensitivity is required for screening purposes while
a high specicity is required for diagnosis. The MNA has been
associated with a high risk overdiagnosis11 and the advantages of
a positive screening and treatment still need to be evaluated and
balanced in both terms of outcome improvement and costeffectiveness.
There is now evidence that nutritional support is able to reduce
the rate of complications only when delivered to elderly patients
who are undernourished or at risk of malnutrition.36 In this respect,
several studies have been performed using the MNA as screening
tool,10 whereas no one is available yet for the GNRI. Moreover, the
MNA has been demonstrated sensible to changes associated with
nutritional repletion.37 Indeed, in the process of screening tools
validation the gap that still remains in scientic literature is that of
intervention studies comparing the value of different tools in
guiding effective nutritional treatment.
The points of strenght of our study are the multivariable
assessment of risk, the good sample size, the limited number of
patients lost to and the length of follow-up. Moreover, we highlight
that potential for bias has been limited through the homogeneous
assessment performed by well-trained operators.38
However, also some limitations should be taken into account for
a correct interpretation of data. First, given the observational nature
of our study, the effect of follow-up variables has not been
considered. Along with this, we can not exclude the effect of
different healthcare practices among the institutions involved on
the patients course of disease. Finally, although according to our
study design it is likely that patients with acute complications have
been not recruited, it can be hypothesized that further adjustments
for factors better accounting for an inammatory background (e.g.
C-reactive protein; not available) would have probably improved
our ndings.
5. Conclusion
In newly institutionalised elderly patients, the prevalence of
nutritional derangements is high, independently of the tool used
for nutritional screening. However, nutritional risk by GNRI but not
nutritional status by MNA was associated with higher mortality
risk. Present data suggest that in the the nutritonal screening of

797

newly institutionalised elderly the use of the GNRI should be

preferred to that of the MNA.
Funding/support
The study was supported by the Fondazione IRCCS Policlinico
San Matteo, Pavia, Italy.
Statement of authorship
All authors signicantly contributed to the work, read and
approved the nal version of the manuscript. Dr Cereda had full
access to all of the data in the study and takes responsibility for the
integrity of the data and the accuracy of the data analysis. Study
concept and design: Cereda, Pedrolli, Vanotti. Acquisition of data:
Cereda, Pedrolli, Piffer, Opizzi, Rondanelli, Zagami. Analysis and
interpretation of data: Cereda, Pedrolli. Drafting of the manuscript:
Cereda, Pedrolli. Critical revision of the manuscript for important
intellectual content: Cereda, Caccialanza, Pedrolli, Rondanelli.
Conict of interest statement
The Authors certify that there are no afliations with or
involvement in any organization or entity with a direct nancial
interest in the subject matter or materials discussed in the
manuscript.
Acknowledgments
The authors are grateful to all doctors, dietitians and nurses of
the Institutions involved for their assistance with data collection in
this study.
References
1. Thomas DR, Ashmen W, Morley JE, Evans WJ. Nutritional management in longterm care: development of a clinical guideline. Council for Nutritional Strategies in Long-Term Care. J Gerontol A Biol Sci Med Sci 2000;55:M725e34.
2. Arvanitakis M, Beck A, Coppens P, De Man F, Elia M, Hebuterne X, et al.
Nutrition in care homes and home care: how to implement adequate strategies
(report of the Brussels Forum (22-23 November 2007). Clin Nutr
2008;27:481e8.
3. Salva A, Coll-Planas L, Bruce S, De Groot L, Andrieu S, Abellan G, Vellas B,
Andrieu S, Bartorelli L, Berner YN, Bruce S, Corman B, Domingo A, Egger TP, de
Groot L, Guigoz Y, Imedio A, Planas M, Porras C, Rovira JC, Salv A, Serra JA,
Vellas B, Task Force on Nutrition and Ageing of the IAGG and the IANA.
Nutritional assessment of residents in long-term care facilities (LTCFs):
recommendations of the task force on nutrition and ageing of the IAGG
European region and the IANA. J Nutr Health Aging 2009;13:475e83.
4. Muscaritoli M, Anker SD, Argils J, Aversa Z, Bauer JM, Biolo G, et al. Consensus
denition of sarcopenia, cachexia and pre-cachexia: joint document elaborated
by Special Interest Groups (SIG) "cachexia-anorexia in chronic wasting
diseases" and "nutrition in geriatrics". Clin Nutr 2010;29:154e9.
5. Norman K, Pichard C, Lochs H, Pirlich M. Prognostic impact of disease-related
malnutrition. Clin Nutr 2008;27:5e15.
6. Caccialanza R, Klersy C, Cereda E, Cameletti B, Bonoldi A, Bonardi C, et al.
Nutritional parameters associated with prolonged hospitalization in ambulatory adult patients. CMAJ 2010;182:1843e9.
7. Lochs H, Pichard C, Allison SP. Evidence supports nutritional support. Clin Nutr
2006;25:177e9.
8. Kondrup J, Allison SP, Elia M, Vellas B, Plauth M. Educational and clinical
practice Committee, European Society of Parenteral and Enteral nutrition
(ESPEN). ESPEN guidelines for nutrition screening 2002. Clin Nutr
2003;22:415e21.
9. Donini LM, Savina C, Rosano A, Cannella C. Systematic review of nutritional
status evaluation and screening tools in the elderly. J Nutr Health Aging
2007;11:421e32.
10. Guigoz Y. The Mini Nutritional Assessment (MNA) review of the literatureeWhat
does it tell us? J Nutr Health Aging 2006;10:466e85. discussion 485e487.
11. Beck AM, Holst M. Rasmussen HH Efcacy of the Mini Nutritional Assessment
to predict the risk of developing malnutrition or adverse health outcomes for
old people. e-SPEN 2008;3:e102e7.
12. Bouillanne O, Morineau G, Dupont C, Coulombel I, Vincent JP, Nicolis I, et al.
Geriatric Nutritional Risk Index: a new index for evaluating at-risk elderly
medical patients. Am J Clin Nutr 2005;82:777e83.

798

E. Cereda et al. / Clinical Nutrition 30 (2011) 793e798

13. Cereda E, Limonta D, Pusani C, Vanotti A. Geriatric nutritional risk index:

a possible indicator of short-term mortality in acutely hospitalized older
people. J Am Geriatr Soc 2006;54:1011e2.
14. Cereda E, Pusani C, Limonta D, Vanotti A. The ability of the Geriatric Nutritional
Risk Index to assess the nutritional status and predict the outcome of homecare resident elderly: a comparison with the Mini Nutritional Assessment.
Br J Nutr 2009;102:563e70.
15. Cereda E, Zagami A, Vanotti A, Piffer S, Pedrolli C. Geriatric Nutritional Risk
Index and overall-cause mortality prediction in institutionalised elderly:
a 3-year survival analysis. Clin Nutr 2008;27:717e23.
16. Cereda E, Pedrolli C. The geriatric nutritional risk index. Curr Opin Clin Nutr
Metab Care 2009;12:1e7. Erratum in: Curr Opin Clin Nutr Metab Care.
2009;12:683.
17. Cereda E, Limonta D, Pusani C, Vanotti A. Feasible use of estimated height for
predicting outcome by the Geriatric Nutritional Risk Index in long-term care
resident elderly. Gerontology 2007;53:184e6.
18. Kaiser MJ, Bauer JM, Rmsch C, Uter W, Guigoz Y, Cederholm T, et al. Mini
Nutritional Assessment International Group. Frequency of malnutrition in
older adults: a multinational perspective using the mini nutritional assessment. J Am Geriatr Soc 2010;58:1734e8.
19. Cereda E, Pedrolli C, Zagami A, Vanotti A, Piffer S, Opizzi A, et al. Body mass
index and mortality in institutionalised elderly. J Am Med Dir Assoc
2011;12:174e8.
20. World Health Organization. Physical status: the use and interpretation of
anthropometry. Report of a WHO Expert Committee. World Health Organ Tech
Rep Ser 1995;854:1e452.
21. Cereda E, Bertoli S, Battezzati A. Height prediction formula for middle-aged
(30-55 y) Caucasians. Nutrition 2010;26:1075e81.
22. Frisancho AR. New norms of upper limb fat and muscle areas for assessment of
nutritional status. Am J Clin Nutr 1981;34:2540e55.
23. Lucchin L, DAmicis A, Gentile MG, Battistini NC, Fusco MA, Palmo A, et al,
PIMAI group. An Italian investigation on nutritional risk at hospital admission:
the PIMAI (Project: Iatrogenic MAlnutrition in Italy) study. e-SPEN, Eur
e-Journal Clin Nutr Metab 2009;4:e199e202.
24. Lucchin L, DAmicis A, Gentile MG, Battistini NC, Fusco MA, Palmo A, et al,
PIMAI group. A nationally representative survey of hospital malnutrition: the
Italian PIMAI (Project: Iatrogenic MAlnutrition in Italy) study. Med J Nutr Metab
2009;2:171e9.

25. Sullivan DH, Walls RC, Lipschitz DA. Protein-energy undernutrition and the risk
of mortality within 1 y of hospital discharge in a select population of geriatric
rehabilitation patients. Am J Clin Nutr 1991;53:599e605.
26. Ballmer PE. Causes and mechanisms of hypoalbuminaemia. Clin Nutr
2001;20:271e3.
27. Knaus WA, Wagner DP, Draper EA, Zimmerman JE, Bergner M, Bastos PG, et al.
The APACHE III prognostic system. Risk prediction of hospital mortality for
critically ill hospitalized adults. Chest 1991;100:1619e36.
28. Cereda E, Vanotti A. Short dietary assessment improves muscle dysfunction
identication by Geriatric Nutritional Risk Index in uncomplicated institutionalised patients over 70 years old. Clin Nutr 2008;27:126e32.
29. Baumgartner RN, Koehler KM, Romero L, Garry PJ. Serum albumin is associated
with skeletal muscle in elderly men and women. Am J Clin Nutr
1996;64:552e8.
30. Shah B, Sucher K, Hollenbeck CB. Comparison of ideal body weight equations
and published height-weight tables with body mass index tables for healthy
adults in the United States. Nutr Clin Pract 2006;21:312e9.
31. Yamada K, Furuya R, Takita T, Maruyama Y, Yamaguchi Y, Ohkawa S, et al.
Simplied nutritional screening tools for patients on maintenance hemodialysis. Am J Clin Nutr 2008;87:106e13.
32. Papanicolaou DA, Wilder RL, Manolagas SC, Chrousos GP. The pathophysiologic
roles of interleukin-6 in human disease. Ann Intern Med 1998;128:127e37.
33. Avraham Y, Hao S, Mendelson S, Berry EM. Hypothalamic-pituitary-adrenal
responses to weight loss in mice following diet restriction, activity or separation stress: effects of tyrosine. Nutr Neurosci 2002;5:327e35.
34. Venihaki M, Dikkes P, Carrigan A, Karalis KP. Corticotropin-releasing hormone
regulates IL-6 expression during inammation. J Clin Invest 2001;108:1159e66.
35. Yeh SS, Schuster MW. Geriatric cachexia: the role of cytokines. Am J Clin Nutr
1999;70:183e97.
36. Milne AC, Avenell A, Potter J. Meta-analysis: protein and energy supplementation in older people. Ann Intern Med 2006;144:37e48.
37. Salas-Salvado J, Torres M, Planas M, Altimir S, Pagan C, Gonzalez ME, et al.
Effect of oral administration of a whole formula diet on nutritional and
cognitive status in patients with Alzheimers disease. Clin Nutr 2005;24:390e7.
38. Kaiser R, Winning K, Uter W, Lesser S, Stehle P, Sieber CC, et al. Comparison of
two different approaches for the application of the mini nutritional assessment
in nursing homes: resident interviews versus assessment by nursing staff.
J Nutr Health Aging 2009;13:863e9.