Key words: Aged, hospital outcomes, length of stay, Mini Nutritional Assessment, nutritional status.Correspondence: Fabio Salvi, MD, U.O. Clinica di Medicina Interna, Azienda Ospedaliero-Universitaria “Ospedali Riuniti”, Via Conca, 60131Ancona, Italy.E-mail: f.salvi@univpm.itReceived March 21, 2007; accepted in revised form August 31, 2007.

Mini Nutritional Assessment (short form)and functional decline in older patients admittedto an acute medical ward

Aging Clinical and Experimental Research

Fabio Salvi1, Raffaella Giorgi1, Annalisa Grilli1, Valeria Morichi1, Emma Espinosa1,Liana Spazzafumo2, Maria Laura Marinozzi1, and Paolo Dessì-Fulgheri1

1Department of Internal Medicine, University “Politecnica delle Marche”, Ancona, 2Statistic and BiometryCentre, Department of Gerontological Research, Italian National Institute for Research and Care of Aging(INRCA), Ancona, Italy

ABSTRACT. Background and aims: Protein and/orenergy malnutrition is common in hospitalized olderpatients and is associated with poor outcomes. Amongrecommended nutritional screening tools, contrastingdata exist about the usefulness of the Mini Nutri-tional Assessment Short-Form (MNA-SF). We evalu-ated whether the MNA-SF, alone or integrated withserum albumin levels, is a reliable predictor of func-tional decline in older patients. Methods: We studied275 elderly patients (mean age 76.5, 60.7% males) ad-mitted to an acute medical ward of a tertiary-careteaching hospital over a 12-month period. In this ob-servational study, we evaluated nutritional status,with the MNA-SF alone or integrated with albumin.Data were collected at admission and related to lab-oratory and geriatric assessment features, and lengthof stay (LOS). Functional decline (defined as a loss≥10% in terms of Barthel Index score at dischargecompared with 2 weeks before admission) was con-sidered as outcome. Results: The MNA-SF estimated46% patients at risk of malnutrition. These subjectshad worse clinical features (lower total cholesteroland albumin levels), longer LOS (13.3 vs 11.2 days,p=0.014) and considerable functional decline (OR4.25, 95% CI 1.83-9.9, p=0.001). Integrating theMNA-SF with albumin values, we obtained an effectiveinstrument to detect older inpatients with protein-en-ergy malnutrition, at higher risk of undergoing func-tional decline (OR 16.19, 95% CI 4.68-56.03,p<0.0001). Conclusions: The MNA-SF is a usefulscreening tool for hospitalized elders at risk of mal-nutrition. It is associated with poor clinical outcomesand is able to predict functional decline. Together with

hypoalbuminemia, it better identifies patients with trueprotein-energy malnutrition.(Aging Clin Exp Res 2008; 20: 322-328)©2008, Editrice Kurtis

INTRODUCTIONProtein-energy malnutrition (PEM) is known to be fre-

quent in hospitalized elderly patients, and ranges from 30to 60% (1, 2). Reasons include poor recognition andmonitoring of nutritional status and inadequate intakeof nutrients for days at a time. PEM is associated withpoor clinical outcomes (3-6). Further deterioration in thenutritional status during in-hospital stay and after dis-charge has been described in hospitalized older patients(7). Nutritional interventions such as counseling, supple-mentation and enteral feeding, have been shown to im-prove nutritional status and to reduce mortality and com-plications (8). Therefore, the need for routine nutritionalassessment should be emphasized and implemented inhospital acute medical wards (9).

Unfortunately, there are no universally accepted indicesfor the diagnosis of PEM. Serum albumin and cholesterollevels or isolated anthropometrical measurements areparticularly prone to confounding factors (10, 11). More-over, currently recommended nutritional screening toolsare often too complex and time-consuming to be routinelyapplied in frail hospitalized elders (12). Indeed, their col-laboration is often difficult, especially when they are de-mented.

The Mini Nutritional Assessment (MNA) is a reliable as-sessment tool for nutritional problems (13). Poor nutri-tional status measured by MNA has been associated withincreased in-hospital mortality and longer length of stay

322 Aging Clin Exp Res, Vol. 20, No. 4

0ccoommssiiddere

isis obb--onall statusstatus,

g atedted withwith albumin.albuminmission andnd relaterelatedd toto lab-

c assessessmentsm featuresures, and lS)S).. Fununctionactionall decldecliinen (define(

% iinn teermsrms ooff BarBarthelhel IIndexndexpareparedd withwith 22 weeeks bedd asas ooutct omepatipa ent

on (PEEMM)) iiss kknownownn tto bebe freeleldederlrlyy paatients,tients andan raangen s fr

2). Reasononss iinnclucluddee popoor retoring ooff nunutritionnalal sstattatus a

off nnuutrriienenttss ffoorr daysda at a tippoooorr clinclinicaical outcomnun trriititional tchar

aa losslossat dischadischargrgee

dmission)dmission) waswa cocots:: ThThee MNA-SFMNA-SF es

k ofof mam lnutrinutritiiononclilinicnicalal featfeatuuresre (lllleveevelsls),), llodd

dequdequEMM isis associasso

). FFurthurtheerr deterideteriororata iuuriringng in--hhospiosp ttalal stast

beeenen ddese ccribribede in hohNutritioonanall iintenterrvevenntit

mmenenttaation anandd eproproveve nutnuppli

MNA-SF in acute medical inpatients

Aging Clin Exp Res, Vol. 20, No. 4 323

(LOS) (14, 15). It was recently recommended for screen-ing for malnutrition in older people (16), but its length lim-its its usefulness (17). A shorter version (MNA-SF) is stilluseful in identifying people at risk of malnutrition. It is eas-ier and faster to perform, although contrasting data ap-pear when it is applied to hospitalized older patients(17-19).

This study aimed at evaluating whether MNA-SF aloneor integrated with albumin is a valid screening tool for(protein-energy) malnutrition and a reliable predictor offunctional decline in older patients admitted to an acutemedical ward.

METHODSSubjectsThe study population consisted of 275 patients aged

65 years and more, admitted to the Department of In-ternal Medicine at the University Hospital of Ancona(tertiary-care teaching hospital), between January andDecember 2005. Of these, 124 (45%) came from theEmergency Department (ED); others were admitted forthird level procedures, such as angiography, coronaro-graphy, etc. Our 18-bed department is in the GeriatricsPost-Graduate Medical School centre and is oriented to-ward cardiovascular, cerebrovascular and metabolic dis-eases. The present study group represents 40% of the687 admissions to the Unit during the study period. Ex-clusion criteria were age ≤65 years, critical clinical con-ditions or very short in-hospital stay (preventing fromadministration of comprehensive geriatric assessment), andcachexia by known metastatic cancer. Each subject gavewritten informed consent to participate in the researchprotocol.

Clinical evaluationAnthropometrical measurements [height, without

shoes; weight, with light garments; body mass index(BMI)] were taken at admission, following standardized pro-cedures. LOS and number of drugs (n. drugs) were alsorecorded. BMI data were missing for 19 bedridden pa-tients (7%). Venous blood samples were collected after anovernight fast within 24 hours of admission. All laboratorydeterminations were carried out with standard commercialkits at the Ancona University Hospital laboratory. Hy-poalbuminemia was defined as albumin <35 g/L. Theblood urea nitrogen (BUN)/creatinine ratio was used as adehydration index; the presence of inflammation wasdescribed by erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) values.

Nutritional assessmentNutritional status was assessed with the MNA-SF ques-

tionnaire. It consists of 6 items extracted from the MNA:BMI, appetite, weight loss, mobility, current illness, andneuropsychological problems (0-14 points) (17). Ac-

cording to their MNA-SF score, patients were classified aswell nourished (≥12) or at risk of malnutrition (≤11).Then, we integrated MNA-SF with the presence of hy-poalbuminemia. We classified as malnourished patientssatisfying both conditions (12 of whom with missed BMI)and well nourished those not satisfying any condition. Pa-tients with MNA-SF scores <12 without hypoalbuminemiawere classified at risk of malnutrition. Therefore, we ex-cluded 14 patients with isolated hypoalbuminemia (3 ofwhom with missed BMI), probably due to severity of ill-ness. Thus, the final sample for this analysis resulted in261 patients.

Geriatric assessmentThe Geriatric Index of Comorbidity (GIC) and the Dis-

ease Burden (IDStot, as the sum of scores in each Indi-vidual Disease Severity) were used as comorbidity in-dices (20). The Mini Mental State Examination (MMSE)and 15-item Geriatric Depression Scale (GDS) describedcognitive and affective status (21, 22). Functional statuswas assessed by the Barthel Index (BI) (23), according toanamnestic data referring to 2 weeks before admission, atadmission and at discharge. Data about pre-hospitalfunctional status were obtained from the patient or aclose relative. Lastly, Instrumental Activities of Daily Liv-ing (IADL) (24) were recorded as number of lost IADL outof number of applicable IADL, expressed as a percentage.

Statistical analysisContinuous variables [expressed as means (SEM)] were

analyzed by univariate analysis of variance (ANOVA),adjusting data for age, comorbidity (as IDStot) and EDprovenance of patients. Admission from the ED wascharacterized by worse clinical features and a different clin-ical approach to the planned one. Therefore, it was in-troduced as a factor in statistical analysis. When ANOVAgroups were more than two, the multiple-comparisonsprocedure performed was by the Bonferroni method(post-hoc test). As we found no significant interaction be-tween gender and considered variables, it was not con-sidered as a factor in analysis. Functional decline duringhospitalization was defined as a loss ≥10% in terms of BIscore at discharge compared with 2 weeks before ad-mission. Six patients with a pre-hospital BI=0 were ex-cluded. Multiple logistic regression models were used toevaluate the association between functional decline andMNA-SF or MNA-SF integrated with hypoalbuminemia,adjusting for age, gender, comorbidity (as IDStot), and EDprovenance. Results are presented as adjusted odds ratio(OR) and 95% confidence intervals (95% CI). Categoricalvariables [expressed as frequency (percentage)] were an-alyzed by Pearson’s χ2 or Fisher’s Exact tests. A 2-tailedp-value <0.05 was considered statistically significant.Statistical analysis was performed with SPSS statisticalsoftware version 13.0 [SPSS Inc. Chicago, Ill., USA].

pprorott

CC

theheperrioiod.d. ExEx--

rrititicaicall clinicallinical concon-pital stay ((prevpreventientingng frofromm

ehehensnsiveiv geriatricca asseassessmess nt)ownn mmetaetastastatictic cacancncer. Each s

nformedrmed coconnsensentt ttoo pparticiparticococolol..

niniccalal evaluh

DS)DS) dessccrri2). FuFunctionalnctional statstat

ndex (BBII)) (23),(23), accordccordiinngg tnngg ttoo 22 weweeekksk befobe rere admadmissio

d at disscchargeharg . DaDatata abouab tnal statatuuss weree obtaobtainedined fro

clososee reellaatitiveve.. LLasastlt y, Instrt uminingg (IADL(IADL))LL (2(244)) werw e rofo nunummbber of

gavveeee rresearchsearch

metmetrriicalcal measurememeasuremhtht,, wiwithth llighighttkenken ata

papatiivvititieses ofo

asas nunumbermber ofof lostlosee IIADL,L, exexprpresseessedd

stiicacall anaanallyyssiissCConto inuououss vvav rii bb

aannalyalyzedzed byadadjd

F. Salvi, R. Giorgi, A. Grilli, et al.

324 Aging Clin Exp Res, Vol. 20, No. 4

RESULTSWe studied 275 patients (167 males, 60.7%) with a

mean age of 76.5±7.1 years (range 65-99). We ob-served high comorbidity [212 (77%) patients were inGIC class III or IV], as also suggested by the high numberof drugs used (6±3). Before admission, 125 (46%) patientswere independent in ADL (BI=100), whereas only 21(7.7%) had a severe disability (BI≤50). In contrast, only 39(31.7%) of the patients admitted from the ED were in-dependent, compared with 86 (57.7%) of those notcoming from the ED (p<0.0001).

MNA-SF and biological markersMNA-SF classified 127 patients (46.2%; 72 males, 55

females, p=0.22) at risk of malnutrition. We observed nodifference in ED provenance or comorbidity. Patients at risk

of malnutrition had lower BMI, serum albumin, total andLDL-cholesterol levels, and increased CRP. Moreover, therisk of developing pressure sores (as lower score on the Nor-ton scale) and levels of cognitive impairment, depression anddisability, as well as LOS, were increased (Table 1).

Integrating MNA-SF with hypoalbuminemia, 36 pa-tients (13.1%; 18 males, 18 females, p=0.25) were reallymalnourished, 91 (34.9%; 54 males) at risk of malnutri-tion, and 134 (51.3%; 87 males) well-nourished. Reallymalnourished patients had lower hemoglobin, serumiron, and total, HDL- and LDL-cholesterol levels. They al-so showed increased inflammatory indices, risk of devel-oping pressure sores (as lower score on the Nortonscale), depression and disability, and therefore appearedas frail elders. LOS was also significantly longer in thesepatients (Table 2).

Clinical variables MNA-SF ≥12 MNA-SF ≤11(n=148) (n=127) p

Age, yr 75.2 (0.5) 77.9 (0.7) 0.001BMI, Kg/m2 28.1 (0.4) 25.8 (0.4) <0.0001Hemoglobin, g/dL 13.1 (0.1) 12.9 (0.1) 0.24Creatinine, mg/dL 1.33 (0.04) 1.26 (0.05) 0.23BUN/creatinine ratio 20.3 (0.6) 20.6 (0.6) 0.67Total cholesterol, mmol/L 4.72 (0.09) 4.29 (0.1) 0.002HDL-cholesterol, mmol/L 1.07 (0.02) 1.02 (0.03) 0.14Triglycerides, mmol/L 1.46 (0.06) 1.33 (0.06) 0.12LDL-cholesterol, mmol/L 2.96 (0.07) 2.68 (0.08) 0.011Albumin, g/L 39.7 (0.4) 38.2 (0.4) 0.007Iron, μg/dL 76.3 (3.5) 70.2 (3.7) 0.24ESR, mm/h 29.7 (1.6) 33.6 (1.7) 0.1CRP, mg/dL 2.35 (0.61) 4.53 (0.65) 0.016

Geriatric assessment variables

ED, n (%)No 89 (60) 62 (51.2) 0.07*Yes 59 (40) 65 (48.8)

GIC, n (%)Class I - -Class II 41 (27.7) 22 (17.3) 0.07*Class III 58 (39.2) 49 (38.6)Class IV 49 (33.1) 56 (44.1)

n. drugs (at admission) 5.8 (0.2) 6.2 (0.2) 0.27Norton plus scale 16 (0.3) 14 (0.3) <0.0001MMSE (adjusted for scholarity) 24.7 (0.3) 23 (0.4) 0.002GDS 3.3 (0.3) 4.8 (0.3) <0.0001Barthel Index (before admission) 91.8 (1.6) 81.4 (1.7) <0.0001

Independent, n (%) 85 (58.2) 40 (31.7) <0.0001*Barthel Index (at admission) 82.1 (2) 63 (2.2) <0.0001Barthel Index (at discharge) 90.8 (1.7) 74.4 (1.9) <0.0001IADL lost, % 12 (2.1) 32.6 (2.2) <0.0001LOS, days 11.2 (0.6) 13.3 (0.6) 0.014

For abbreviations, see text. Results of univariate analysis of variance adjusted for age, comorbidity (as IDStot) and provenance (ED yes/no) of patients, exceptfor age (Student’s t-test); *results of Fisher’s Exact Test or Pearson’s χ2. Values are as adjusted means (SE), except for age, which is mean (SE).

Table 1 - Characteristics of study population by MNA-SF score.

IrIroon,n,ESR,ESR,CR

tis1

Kur127)27 p

77.7.99 (0.7)(0.7)25.88 (0.4)(0.4)

(0.1) 12.9 (0 11.331.33 (0.04)(0.04) 120.320.3 (0.6)0.64.724.72 (0.09)(0 01.071 (0

mmol/Lol/Lg/L

μgμg//dLdL//mm/hmm/hmgmg/d/dLL

5)5)0.0.66 (0.6))

4.294.29 (0.1)(01.021.02 (0(0

0.06)062.962.96 (0.07)(0.07)39.739.7 (0.4)(0.4)76 376

OR PFORssmssmentent varivariableabless

MNA-SF in acute medical inpatients

Aging Clin Exp Res, Vol. 20, No. 4 325

MNA-SF and functional outcomeForty-two patients (15.3%) underwent a ≥10% func-

tional decline. Thirty-three of them (78.6%) resulted at riskof malnutrition by MNA-SF and 14 (36%) were “really mal-nourished” according to the combined criterion. Adjustingfor age, gender, comorbidity and ED provenance, anMNA-SF ≤11 conferred a high risk (OR 4.25, 95% CI1.83-9.9; p=0.001) of developing a significant function-al decline during their in-hospital stay (Table 3). We ob-tained even more striking results integrating hypoalbu-minemia. We observed a significantly greater risk of func-tional decline in really malnourished patients (OR 16.19,95% CI 4.68-56.03; p<0.0001) and ones at risk of mal-nutrition (OR 4.27, 95% CI 1.53-11.97; p=0.006) thanin well-nourished ones (Table 4). In both multiple logistic re-

gression models, ED provenance and comorbidity (as ID-Stot) also conferred a higher risk of undergoing function-al decline, whereas age and gender did not.

DISCUSSIONForty-six percent of elderly patients admitted to an

acute medical ward are estimated at risk of malnutrition bythe MNA-SF. This is in keeping with previous data (1, 2,19, 25). These patients had worse clinical features,longer LOS, and underwent functional decline duringtheir in-hospital stay, confirming the negative impact ofthis condition (3-7, 17). We found that the MNA-SFscores were related to albumin and cholesterol levels,inflammatory indices, geriatric assessment variables suchas cognitive and affective status, disability, and LOS

Clinical variables Well-nourished At risk Malnourished(n=134) (n=91) (n=36) p

Age, yr 75 (0.5)* 76.4 (0.7)* 81.7 (1.3)‡ <0.0001BMI, Kg/m2 28.2 (0.4)‡ 25.9 (0.5)* 25.7 (0.9)* <0.0001Hemoglobin, g/dL 13.3 (0.1)* 13 (0.2)* 12 (0.3)‡ 0.001Creatinine, mg/dL 1.3 (0.05) 1.29 (0.05) 1.32 (0.09) 0.97BUN/creatinine ratio 19.6 (0.5) 21.1 (0.6) 19.8 (1.1) 0.17Total cholesterol, mmol/L 4.8 (0.09)* 4.56 (0.11)* 3.42 (0.19)‡ <0.0001HDL-cholesterol, mmol/L 1.1 (0.03)* 1.06 (0.03)* 0.86 (0.05)‡ <0.0001Triglycerides, mmol/L 1.46 (0.06) 1.4 (0.07) 1.19 (0.12) 0.15LDL-cholesterol, mmol/L 3.03 (0.08)* 2.88 (0.09)* 2.01 (0.16)‡ <0.0001Albumin, g/L 40.5 (0.3)* 40.3 (0.4)* 32.3 (0.6)‡ <0.0001Iron, μg/dL 80.8 (3.7)* 76 (4.3)* 45.2 (7.5)‡ <0.0001ESR, mm/h 27.7 (1.7)* 30.4 (2)* 45.8 (3.4)‡ <0.0001CRP, mg/dL 1.61 (0.59)* 2.53 (0.69)* 10.78 (1.23)‡ <0.0001

Geriatric assessment variables

ED, n (%)No 81 (60.4) 50 (54.9) 12 (33.3) 0.015†Yes 53 (39.6) 41 (45.1) 24 (66.7)

GIC, n (%)Class I - - -Class II 40 (29.9) 18 (19.8) 4 (11.1) 0.1†Class III 50 (37.3) 35 (38.5) 14 (38.9)Class IV 44 (32.8) 38 (41.8) 18 (50)

n. drugs (at admission) 5.6 (0.3) 6.6 (0.3) 6.3 (0.6) 0.06Norton plus scale 16.7 (0.2) 14.5 (0.3) 12.1 (0.5) <0.0001#

MMSE (adjusted for scholarity) 24.6 (0.3) 23.3 (0.4) 22.7 (0.8) 0.017GDS 3.4 (0.3)* 5.1 (0.3)* 3.8 (0.6) <0.0001Barthel Index (before admission) 94 (1.6) 83.7 (1.9) 73.9 (3.3) <0.0001#

Independent, n (%) 81 (61.4) 36 (40) 4 (11.1) <0.0001†

Barthel Index (at admission) 85.6 (2) 67.4 (2.4) 47.1 (4.2) <0.0001#

Barthel Index (at discharge) 93 (1.7) 78.2 (2) 58.1 (3.8) <0.0001#

IADL lost, % 10.4 (2.1) 29.1 (2.5) 44.7 (4.4) <0.0001#

LOS, days 10.7 (0.6)* 12.1 (0.7)* 16.6 (1.3)‡ <0.0001

For abbreviations, see text. Results of univariate analysis of variance adjusted for age, comorbidity (as IDStot) and provenance (ED yes/no) of patients;†: results of Pearson’s χ2; *: significant difference between two groups; */‡: ‡value is significantly different from *values; #: all comparisons are significant. Val-ues are adjusted means (SE), except for age, which is mean (SE).

Table 2 - Characteristics of study population according to combined criteria (MNA-SF and hypoalbuminemia).

artIIADADLLLOLOSS

0.010.015566.7)

--44 ((111.1.11))

38.5) 1414 ((38.9)338 (4411..8)8) 18 (56.66.6 (0.3)0.3

) 14.5.5 (0.3)(0.3)4.64.6 (0.3)3) 23.32 (0 433..44 ((0.3)3)*

onn)) 9494 ((11.6)6)8181 ((61.4)

(atat aadmidmisssision)on) 85ell IndIndexex (at(at didiscscharharge)ge)lloost,st, %%daaysys

evi

0.6)6)12.11 (0.5)(0.5)22.77 (0.8)(0.8

3) 383.7.7 (1.9)366 (4(400)

67.467.4 (2(233 ((11..7)7)

10.410.4 ((22.1)1)10.710.7 ((00

R PEtextext.. ResuResultslts ofo univnivara ia

soon’sn’s χχ22;; *:* sisignignificaficantnt dif

Fmeansmeans ((SESE),), exce

F. Salvi, R. Giorgi, A. Grilli, et al.

326 Aging Clin Exp Res, Vol. 20, No. 4

(Table 1), all suggesting that the MNA-SF is a good,rapid screening tool for risk of malnutrition in hospitalizednon-surgical older patients. Moreover, patients with anMNA-SF ≤11 had a higher risk of developing considerablefunctional decline during hospitalization, with ED prove-nance and comorbidity also involved (Table 3).

There is a complex interrelationship among nutri-tional status, illness severity, and clinical outcomes inhospitalized elders (26). A certain proportion of theseproblems may be avoided by timely recognition of mal-nutrition and starting early nutritional supplementation (8,27). An overt problem is that there is no clinical standardfor the diagnosis of PEM.

Serum albumin is not so useful in patients admitted foracute disease (10), and inflammatory components maypredispose to low cholesterol levels (28). A detailed nu-tritional assessment, even if recommended, is often im-possible, requiring more time than that available. Thus,having a reliable screening tool that can identify per-sons who are malnourished or at risk of malnutrition iscrucial. Unfortunately, current tools do not perform verywell or are often too long to administer (12, 17). Recently,the MNA-SF was proposed as a very quick screening toolin 408 pre-operative ambulatory older patients (29), butno association was found between MNA-SF scores andboth laboratory data and clinical outcomes in 73 elderlypatients with hip fracture (19). Our data indicate MNA-SFas a valid screening tool for risk of malnutrition, since itcan be correlated with poor nutritional habits and func-tional decline during hospitalization.

In order to explain the reduced usefulness of MNA-SF inclinical pathways, some authors hypothesize that it does notidentify more precisely patients at risk of or with PEM,and/or worsened clinical outcomes, probably because it doesnot incorporate laboratory data. Although our data showedthat MNA-SF per se is a good screening tool, it appears stillmore effective in identifying frail elderly patients when it isintegrated with hypoalbuminemia, probably because it bet-ter identifies patients with true PEM (Tables 2 and 4). In thisfield, Formiga et al. (19) state that a low MNA-SF score does

not represent a risk for normo-albuminemic patients, hy-pothesizing the absence of hypoalbuminemia as the resultof good medical care. In our opinion, it is difficult to relatethe absence of hypoalbuminemia in acute inpatients to a cor-rect medical management. In our study, albumin data werecollected within the first 24 hours, when the effect of suchmanagement was limited, whereas that of illness severity wasgreater. Moreover, we could not detect significant differencesin laboratory values between patients with MNA-SF ≤11without hypoalbuminemia and well-nourished patients. Incontrast, our findings regarding the risk of functional declineduring in-hospital stays suggest that a low MNA-SF score isa risk factor even in normo-albuminemic patients, at least inthose with acute medical problems. Lastly, hypoalbumine-mia conferred a further higher risk of undergoing large-scalefunctional decline during in-hospital stays (Table 4).

Another important finding was the correlation betweenMNA-SF and LOS. This matches the results of recent largestudy on 908 hospitalized older patients in France (30),where malnutrition, evaluated by MNA-SF, conferred a2.5-fold higher risk of longer LOS. In addition, it was not-ed that hypoalbuminemia does not increase the efficacy ofthe MNA-SF alone in predicting LOS, suggesting thatthe relationship is independent of illness severity.

Some limitations of our study should be considered.First, we do not have information about other anthro-pometric measures, such as mid-arm circumference, par-ticularly in the 19 patients with missed BMI. Nevertheless,some authors have shown that anthropometric mea-sures have lower correlation than other MNA subscoreswith laboratory indices of malnutrition in older patients(15). Moreover, 12 of our patients with missed BMIwere “malnourished” according to both criteria, sug-gesting that the BMI value was not determinant in clas-sifying them. Indeed, in these subjects, MNA-SF waspositive independently of BMI availability (that is, a score

Factors OR 95% CI p

MNA-SF ≤11 4.25 1.83-9.9 0.001ED yes 2.85 1.21-6.72 0.016Gender (female) 1.81 0.82-3.99 0.14Age 1.04 0.98-1.11 0.15Comorbidity (IDStot) 1.10 1.02-1.19 0.01

Logistic regression with dependent variable: functional decline during hos-pitalization (defined as loss ≥10% in terms of BI score at discharge comparedwith 2 weeks before admission). Age and IDStot are non-categorical variables.Nagelkerke r-square =0.32. For abbreviations, see text.

Table 3 - Clinical predictors of functional decline during hospi-talization.

Factors OR 95% CI p

Combined criteria:Well-nourished 1At risk of PEM 4.27 1.53-11.97 0.006Really malnourished 16.19 4.68-56.03 <0.0001

ED (yes) 4.03 1.55-10.49 0.004Gender (female) 1.61 0.69-3.77 0.27Age 1.02 0.96-1.09 0.56Comorbidity (IDStot) 1.11 1.03-1.2 0.009

Logistic regression with dependent variable: functional decline during hos-pitalization (defined as loss ≥10% in terms of BI score at discharge comparedwith 2 weeks before admission). “Well nourished”: normo-albuminemic pa-tients with MNA-SF≥12; “At risk of PEM”: normo-albuminemic patients withMNA-SF≤11; “Really malnourished”: hypoalbuminemic patients with MNA-SF≤11. Age and IDStot are non-categorical variables. Nagelkerke r-square=0.4. For abbreviations, see text.

Table 4 - Clinical predictors of functional decline during hospi-talization.

iiddenenaannaaa d/od/nnnn

erlylye MNA-SFMNA-SF

alnutritalnutri ion,on, sincesince ittutritionalal habitshabits andand ffuncn -

ospitpitalalizatiz ion.plainain tthehe rereduceducedd uusefse uf lneu ss of M

pathp waywayh ssyy ,, ssoomem autauthohorsrs hypohypothtifyify momorere prepreciseciselyly patientoror worworsenesenerrr dd clinicalccorporporatea lNA

inemicmic ppatieatienntsts,, hhypoalbumuminemiainemia asas thethe rresues

nn ourour opiopinion,nion, itit is diffdifficulticult tohypoalbubul minemineuuu miamia iinn acacuucc teteu inpati

edical mamannaagemeentnt. IInn oour studycolollleecteedd wwiithihinn tthhee first 224 houmamannaaa aagga eemementn wawassaa limitedgreatgreatererttt . MMoreoinin labo

F iinnitit doesdoes notnot

k ooff oror wwitithh PEPEMMes,es, pprroorrr bablbabloo yylll becausebecause

ryry datda a. AltAlthohouugghh op rr sese iiss aa goodgood scrreeeeee inin idenidentifyingtifyin ffhh hyph

iinnenn thehe effef e

eaeassaa ththatat ooff illnillneessss secocouldu nonottooo detdetectectt ssigniignifif

valueluess bebetweentween ppatieatt hyypopoaalbumlbumininememiaia

cocontratran sts , ourr ffiifff ndndinddururiningg inin-haa rii k

MNA-SF in acute medical inpatients

Aging Clin Exp Res, Vol. 20, No. 4 327

≤8). In our sample, the BMI of the “really malnourished”group (Table 2) was higher than the conventional limit ofoverweight, and similar to that of subjects at risk of mal-nutrition. Despite this, the MNA-SF seems to be able topredict worse clinical outcomes even in cases of over-weight, suggesting that, at least in our high cardiovascu-lar risk patients, a little weight excess represents neitherbetter nutritional status nor a protective factor for unfa-vorable hospital outcomes. Second, all clinical parametersare affected by illness severity, and we are unable to iso-late the effect of malnutrition from the influence of the dis-ease on clinical outcomes, but this is a common problemto all studies about PEM in hospitalized acute patients.Nevertheless, in order to take this aspect into considera-tion, we adjusted data for both ED provenance and co-morbidity.

Thus, the MNA-SF alone or when integrated withserum albumin levels is a useful screening tool to identifypatients at risk of (protein-energy) malnutrition among hos-pitalized elders admitted to acute medical wards. Be-cause of its correlation with clinical features and functionaldecline, it may represent a simple, rapid instrument toidentify patients needing early nutritional support.

Further studies are needed to confirm our results inhospitalized older patients, especially for other than med-ical and cardiovascular reasons, such as surgery, injury, orspecific pathologies.

CONCLUSIONSIn this observational study, we investigated whether

MNA-SF alone or integrated with serum albumin levels is areliable predictor of functional decline in older patientshospitalized in an acute medical ward. The results indicat-ed that the MNA-SF is a valid instrument, able to identifypatients at risk of malnutrition with worse clinical features,longer LOS and functional decline during hospitalization. In-tegrating MNA-SF with albumin values, we obtained an ef-fective tool to detect older inpatients with a high probabil-ity of having true PEM. These data suggest that the MNA-SF is useful in screening hospitalized elders at risk of mal-nutrition with poor clinical features and functional out-come, identifying patients needing early nutritional support.

ACKNOWLEDGEMENTSThis study was supported by research grants from the University of

Ancona “Politecnica delle Marche” (FS and PDF), without any role inthe design, execution, analysis or interpretation of data, or writing ofthe study.

tritional status and adverse outcomes in older hospitalized medi-cal patients. J Am Geriatr Soc 1999; 47: 532-8.

4. Potter J, Klipstein K, Reilly JJ, Roberts M. The nutritional statusand clinical course of acute admissions to a geriatric unit. AgeAgeing 1995; 24: 131-6.

5. Naber TH, Schermer T, de Bree A, et al. Prevalence of malnu-trition in nonsurgical hospitalized patients and its associationwith disease complications. Am J Clin Nutr 1997; 66: 1232-9.

6. Sullivan DH, Walls RC, Bopp MM. Protein-energy undernutritionand the risk of mortality within one year of hospital discharge: afollow-up study. J Am Geriatr Soc 1995; 43: 507-12.

7. Volkert D, Hübsch S, Oster P, Schlierf G. Nutritional support andfunctional status in undernourished geriatric patients during hos-pitalization and 6-month follow-up. Aging Clin Exp Res 1996; 8:386-95.

8. Milne AC, Avenell A, Potter J. Meta-analysis: protein and energysupplementation in older people. Ann Intern Med 2006; 144:37-48.

9. Kruizenga HM, Van Tulder MW, Seidell JC, Thijs A, Ader HJ,Van Bokhorst-de van der Schueren MA. Effectiveness and cost-effectiveness of early screening and treatment of malnourished pa-tients. Am J Clin Nutr 2005; 82: 1082-9.

10. Courtney ME, Greene HL, Folk CC, Helinek GL, Dmitruk A.Rapidly declining serum albumin values in newly hospitalizedpatients: prevalence, severity, and contributory factors. JPEN JParenteral Enteral Nutr 1982; 6: 143-5.

11. Onder G, Landi F, Volpato S, et al. Serum cholesterol levels andin-hospital mortality in the elderly. Am J Med 2003; 115: 265-71.

12. Omran ML, Morley JE. Assessment of protein energy malnutri-tion in older persons, part I: history, examination, body compo-sition, and screening tools. Nutrition 2000; 16: 50-63.

13. Guigoz Y, Vellas B, Garry P. Mini Nutritional Assessment: apractical assessment tool for grading the nutritional state of elderlypatients. Facts Res Gerontol 1994; 2: 15-59.

14. Van Nes MC, Herrmann FR, Gold G, Michel JP, Rizzoli R.Does the Mini Nutritional Assessment predict hospitalizationoutcomes in older people? Age Ageing 2001; 30: 221-6.

15. Kagansky N, Berner Y, Koren-Morag N, Perelman L, Knobler H,Levy S. Poor nutritional habits are predictors of poor outcome invery old hospitalized patients. Am J Clin Nutr 2005; 82: 784-91.

16. Kondrup J, Allison SP, Elia M, Vellas B, Plauth M. ESPENguidelines for nutritional screening 2002. Clin Nutr 2003; 22:415-21.

17. Rubenstein LZ, Harker JO, Salvà A, Guigoz Y, Vellas B.Screening for undernutrition in geriatric practice. Developingthe short-form Mini-Nutritional Assessment (MNA-SF). J GerontolA Biol Sci Med Sci 2001; 56: M366-72.

18. Ranhoff AH, Gjøen AU, Mowé M. Screening for malnutrition inelderly acute medical patients: the usefulness of MNA-SF. J NutrHealth Aging 2005; 9: 221-5.

19. Formiga F, Chivite D, Mascaró J, Ramón JM, Pujol R. No cor-relation between Mini-Nutritional Assessment (short form) scaleand clinical outcomes in 73 elderly patients admitted for hipfracture. Aging Clin Exp Res 2005; 17: 343-6.

20. Rozzini R, Frisoni GB, Ferrucci L, et al. Geriatric Index ofComorbidity: validation and comparison with other measuresof comorbidity. Age Ageing 2002; 31: 277-85.

21. Folstein MF, Folstein SE, McHugh PR. “Mini Mental State”: apratical method for grading the cognitive state of patients for theclinician. J Psichiatr Res 1975; 12: 189-98.

22. Yesavage JA, Brink TL, Rose TL, et al. Development and vali-

REFERENCES1. Constans T, Bacq Y, Bréchot JF, Guilmot JL, Choutet P, Lamisse

F. Protein energy malnutrition in elderly medical patients. J AmGeriatr Soc 1992; 40: 263-8.

2. Sullivan DH, Sun S, Walls RC. Protein-energy undernutritionamong elderly hospitalized patients. JAMA 1999; 281: 2013-9.

3. Covinsky KE, Martin GE, Beyth RJ, Justice AC, Sehgal AR,Landefeld CS. The relationship between clinical assessments of nu-

hohospspeded ththp

y,y oorr

ststudyudy,, we inveeststigigatedate whethor inntegrtegratateded wiwithth serserumr albumilb

ppredictordictor ofof functionalunctional declineecpiittalializezedd iinn ann acuacuttee mem dicalaatt ththee MNA-MNA-SF iststs aatt risk fLO

urishe

C, Helinenekk GLGL,, DDmitrmitrukukumin vvala uueses inin newlnew yy hosphospitalizita e

,, seseveveririty,t andand conco tritributob rryy factfactors JPteral Nuuttrr 1198982; 66:: 14143-3-5.5

er G, LaLannddii F, Volplpaatoto S,S, eet al.l Serin-hhospospiti alal mortmortala ity in tthehe elderl7171..

1212. OOmrmranan ML,M Motition in

is aaerer papattieientntss

Thehe rresultsesults indicatindicannstrumstrumentent,, ablablee ttoo iid

trtritiitiono wwithith worsworsee clliffuncuncff titiccc onalonal decldec inee durd

A-SFA-SF wiwithth albalbumdedetect

Medd 2200003

Assessssmentment ofof pprroteoteininonsns,, papart I: hihisstotory,ry eexxa

sscrecreeeniningng ttoooollss. Nutrit ittGuigoz YY,, VVellasellas B,B GaGprpractical assessmassespapatitienents

141

MNA-SF in acute medical inpatients

dation of a geriatric depression screening scale: a preliminary re-port. J Psychiatr Res 1983; 17: 37-49.23. Mahoney FI, Barthel DW. Functional evaluation: the BarthelIndex. Md State Med J 1965; 14: 61-5.

24. Lawton MP, Brody EM. Assessment of older people: self-main-taining and instrumental activities of daily living. Gerontologist1969; 9: 179-86.

25. Gazzotti C, Albert A, Pepinster A, Petermans J. Clinical usefulnessof the Mini Nutritional Assessment (MNA) scale in geriatricmedicine. J Nutr Health Aging 2000; 4: 176-81.

26. Sullivan DH, Bopps MM, Roberson P. Protein-energy undernu-trition and life-threatening complications among the hospital-ized elderly. J Gen Intern Med 2002; 17: 923-32.

27. Persson M, Hytter-Landahl A, Brismar K, Cederholm T.Nutritional supplementation and dietary advice in geriatric patientsat risk of malnutrition. Clin Nutr 2007; 26: 216-24.

28. Ettinger WH Jr, Harris T, Verdery RB, Tracy R, Kouba E.Evidence for inflammation as a cause of hypocholesterolemia inolder people. J Am Geriatr Soc 1995; 43: 264-6.

29. Cohendy R, Rubenstein LZ, Eledjam JJ. The Mini NutritionalAssessment-Short Form for preoperative nutritional evaluation ofelderly patients. Aging Clin Exp Res 2001; 13: 293-7.

30. Lang PO, Heitz D, Hedelin G, et al. Early markers of prolongedhospital stays in older people: a prospective, multicenter study of908 inpatients in French acute hospitals. J Am Geriatr Soc2006; 54: 1479-80.

328 Aging Clin Exp Res, Vol. 20, No. 4