lable at ScienceDirect

Clinical Nutrition xxx (2013) 1e6

Contents lists avai

Clinical Nutrition

journal homepage: ht tp: / /www.elsevier .com/locate/c lnu

Original article

Mini Nutritional Assessment and short-form Mini NutritionalAssessment can predict the future risk of falling in olderadults e Results of a national cohort study

Alan C. Tsai a,b,*, Mei-Yen Lai a,c

aDepartment of Healthcare Administration, Asia University, 500 Liufeng Road, Wufeng, Taichung 41354, Taiwan, ROCbDepartment of Health Services Management, School of Public Health, China Medical University, Taichung 40402, Taiwan, ROCcNursing Department, Taichung Veterans General Hospital, Taichung, Taiwan, ROC

a r t i c l e i n f o

Article history:Received 24 June 2013Accepted 16 October 2013

Keywords:ElderlyFallingMalnutritionMini Nutritional AssessmentPredicting falls

Abbreviations: CC, calf circumference; MACMNA, Mini Nutritional Assessment; MNA-SF, shAssessment; T1, Taiwan version-1; T2, TaiwanLongitudinal Survey on Aging.* Corresponding author. Department of Healthcare

sity, 500 Liufeng Road, Wufeng, Taichung 41354, Tai3456x1943; fax: þ886 4 2332 1206.

E-mail addresses: atsai@umich.edu (A.C. Tsa(M.-Y. Lai).

0261-5614/$ e see front matter � 2013 Elsevier Ltd ahttp://dx.doi.org/10.1016/j.clnu.2013.10.010

Please cite this article in press as: Tsai AC, Lfuture risk of falling in older adults e Resul

s u m m a r y

Background & aims: Falling is a major issue in geriatric health. Tools that identify individuals at risk offalling can help reduce the risk of falling. The study aimed to determine whether the full and short-formMini Nutritional Assessment (MNAs) have the ability to predict the risk of falling in older adults.Methods: Subjects were 3118 � 53-year old Taiwanese who completed both the 1999 and 2003 “TaiwanLongitudinal Study on Aging” surveys. We rated these subjects with normalized versions of the MNA andapplied the standard cut-offs to define under-nutrition (�23.5 and �11 points, respectively). We usedmultivariate logistic regression analysis and receiver operating characteristics to evaluate the ability ofthese tools in predicting the risk of falling three years later.Results: Older adults rated as at risk of malnutrition with the full MNA (OR ¼ 1.87, 95% confidenceinterval ¼ 1.33e2.63, p < 0.001) or the short-form (1.39, 1.07e1.80, p ¼ 0.014) were associated withincreased risk of falling three years later. Both versions significantly predicted the risk of falling andperformed slightly better in �65-year old persons than in younger (53e64-year old) persons. The short-form performed relatively well compared to the full scale.Conclusions: Results suggest that the full and short-form MNAs, in addition to rating the risk ofmalnutrition, also predict the risk of falling in older adults. Although the short-form is slightly lesseffective than the full scale in predicting the future risk of falling, its simplicity, effectiveness and effi-ciency make it ideal as a multipurpose screening tool in clinical settings.

� 2013 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.

1. Introduction

Falling is a major issue in geriatric health. Falls can often lead tobodily injuries, physical functional disability, cognitive impairment,institutionalization or even death.1,2 Falls also often affect psycho-logical health. Elderly who experienced falling are less likely to goout and may become physically dependent. The cause of falling iscomplex andmulti-factorial.3 Peoplewho havemultiple risk factors

, mid-arm circumference;ort-form Mini Nutritionalversion-2; TLSA, Taiwan

Administration, Asia Univer-wan, ROC. Tel.: þ886 4 2332

i), bs88877@yahoo.com.tw

nd European Society for Clinical N

ai M-Y, Mini Nutritional Assets of a national cohort study,

such as physical functional impairment, poor visual acuity, cogni-tive impairment, frailty, general weakness, postural hypotension orpsychotropic medication have increased risk of falling.3e6 Poornutrition, especially protein-energy malnutrition (PEM), that con-tributes to general weakness or frailty, also increases the risk offalling.7,8

PEM is common in the elderly. It often accelerates loss of musclemass and weakens muscle strength or the ability to maintain bal-ance.9 PEM or malnutrition often develops inconspicuously. It maynot be noticed unless screened or examined. Thus, nutritionscreening is a necessary step in identifying those who are at risk ofmalnutrition for early intervention.

The Mini Nutritional Assessment (MNA) is a simple, non-invasive and efficient tool designed for assessing/screening therisk of malnutrition in elderly adults.10,11 It consists of 18 items, andrates anthropometric, dietary, global and self-viewed aspects ofnutrition. MNA has a simplified short-form (SF) consisting of 6 keyMNA items.12,13 Both scales are multifunctional, and have been

utrition and Metabolism. All rights reserved.

ssment and short-form Mini Nutritional Assessment can predict theClinical Nutrition (2013), http://dx.doi.org/10.1016/j.clnu.2013.10.010

A.C. Tsai, M.-Y. Lai / Clinical Nutrition xxx (2013) 1e62

reported to predict hospital length of stay, hospitalization outcome,physical functional status and mortality.14e16 Both scales includeitems that assess mobility, weight loss, and dementia that areassociated with the risk of falling. Thus, in theory, the scale shouldbe able to predict the risk of falling. Recent studies also have shownthat elderly who have poorer nutritional status have higher risk offalling.1,8,17,18 However, the ability of the MNAs to predict the risk offalling has not been robustly examined in a large prospective study.Thus, the present study was undertaken to examine the abilities ofthe MNA and the MNA-SF in predicting the future risk of falling in alarge sample of older Taiwanese.

2. Materials and methods

2.1. Source of data

The study analyzed data sets of the “Taiwan Longitudinal Surveyon Aging” (TLSA), a population-based cohort study conducted bythe Bureau of Health Promotion of Taiwan.19 The survey employed amulti-stage sampling process to draw a population-representativesample of 4412 � 60-year old Taiwanese men and women fromnon-institutionalized citizens, and 4049 of these participantscompleted the initial survey in 1989 (Fig. 1). The cohort was sur-veyed every 3 or 4 years. In 1996, a second sampling of 2462 per-sons, 50e66 years old, drawnwith the samemethods, was added tothe cohort to extend the age range of the cohort. Subjects in thecombined cohort were interviewed in 1999 and 2003. The 1999interview had expanded survey on dietary intake and includeditems in the MNA. Thus, the 1999 survey was chosen as baseline forthe present study and the 2003 survey as end point.

In each survey, trained interviewers conducted face-to-face in-terviews using a structured questionnaire at respondents’ resi-dence. Of the combined cohort, 4440 persons completed theinterview in 1999, and 3778 persons completed the interview in2003. Among these subjects, 3656 completed both surveys (Fig. 1).

The TLSA survey from which the present study is based uponwas conducted according to the guidelines laid down in theDeclaration of Helsinki and all procedures involving human sub-jects/patients were approved by government-appointed represen-tatives. All participants signed a written consent before taking partin the study. The detail of survey design and procedure has beendescribed elsewhere.20

1989Survey design--Using a multi-stage sampling psample of 4,412 Taiwanese, ≥60 years old. Amo

Attrition1996

(a) 2,669 of the original cohort, ≥67 years old, c(b) 2,462 persons of a new cohort of 3,041 perso

process) completed the survey. The combine

Attrition1999 (baseline of the present study)

4,440 of the combined cohort, ≥53 years old, co

Attrition2003 (end point of the present study)

3,778 persons completed the 2003 survey, and 3246 who missed BMI data, 119 missed fall data3,118 for the 1999-2003 longitudinal regressio

Fig. 1. Flowchart of the Taiwan Longitudinal Survey on

Please cite this article in press as: Tsai AC, Lai M-Y, Mini Nutritional Assefuture risk of falling in older adults e Results of a national cohort study,

2.2. The outcome measure

The outcome measure was the incidence of falling during the12-month period prior to the time of the questionnaire interviewaccording to the question “Did you have any fall during the past 12months?” Those who answered “yes” were considered havingfallen for that year regardless of the number of falls.

2.3. Rating the baseline nutritional status

We used a normalized and slightly modified MNA (Taiwanversion-2, T2) and its short-form (SF), to rate the baseline nutri-tional status of the participants.21,22 MNA-T2 was derived fromTaiwan version-1 (T1), which was normalized from the originalMNA by adopting the population-specific anthropometric cut-offpoints and dietary features.23 MNA-T2 further replaced CC andMAC for BMI item in full MNA by transferring 1 of the 3 BMI pointsto the MAC item and 2 to the CC item.23,24 Therefore, the MNA-T2 isbased on 17, rather than 18 items but has the same total score (30points). The MNA-T2 has been shown to perform at least as well asthe MNA-T1.24 With the exception of fluid intake, all items in theMNA were available in the questionnaire. Thus, grading with theMNA-T2 was based on 16 items and the total score was propor-tionately adjusted to a 30-point basis. A total score <17 wasconsidered as malnourished; 17e23.5, as at risk of malnutrition;�24, as normal. The MNA-T2-SF is similar to the original MNA-SF,except that item F (BMI) was replaced by item R (CC). A totalscore �7 was considered as malnourished; 8e11, as at risk ofmalnutrition, and �12, as normal.12

2.4. Other variables

All other variables were derived from the 1999 data sets. Alcoholdrinking status was classified as non-drinker, drinking<1 time/wk,and >1 time/wk; cigarette smoking as never smokers, currentsmokers and past smokers; and betel nut chewing was classified asno chewers, current chewers and past chewers. Routine physicalactivity was classified according the number of times performingphysical exercise >30 min/day. Body mass index (BMI) was calcu-lated according to kg (weight)/m2 (height). The status of chronicdiseases/conditions including hypertension, diabetes, heart dis-ease, stroke, osteoarthritis, gout, hip fracture, lower-back pain,

rocess to draw a population-representative ng them, 4,049 completed the initial survey.

(1,047) and non-response (333)

ompleted the survey.ns (50-66 years old drawn with the same d cohort totaled 5,131 persons.

(540) and non-response (151)

mpleted the survey.

(315) and non-response (347)

,656 completed both surveys. After excluding and 173 missed other data, the net N wasn analysis.

Aging (TLSA), the data source of the present study.

ssment and short-form Mini Nutritional Assessment can predict theClinical Nutrition (2013), http://dx.doi.org/10.1016/j.clnu.2013.10.010

A.C. Tsai, M.-Y. Lai / Clinical Nutrition xxx (2013) 1e6 3

visual acuity, hearing ability and incontinence was based on self-reported but physician confirmed answers to each of the specificdiseases/conditions.

2.5. Statistical analysis

All statistical analyses were performed with the StatisticalPackage for the Social Sciences (SPSS version 18.0. Chicago, IL).Simple statistics were used to compute descriptive data. Multivar-iate logistic regression analysis was performed to determine theability of the baseline nutritional status rated with the MNA andMNA-SF to predict the risk of falling three years later, controlled forconfounding factors as indicated in footnote to Table 3. The analysisincluded 3118 persons after excluding those whowere without BMIdata (246), fall data (119) and other data (173) (Fig. 1). The analysiswas weighting-adjusted according to sampling ratio of each sam-pling. The ability of baseline nutritional status rated with the MNAorMNA-SF in predicting falling three years later was evaluatedwithreceiver operating characteristic (ROC) analysis. Statistical signifi-cance for all analyses was set at a ¼ 0.05.

3. Results

Table 1 shows the characteristics of subjects at baseline. Thesample included slightly more men (52.9%) than women; 61.6%were �65 years old; 70.0% had �6 years of formal education; 81.2%were BMI 19e27 kg/m2; 20% were current smokers; 25.3% drank�1/week; 59.5% exercised�3 times/wk; and 18.0% experienced oneor more falls during the past 12 months.

Table 1Characteristics of subjects at baseline (%).

Item N ¼ 3118

SexMen 52.9Women 47.1

Age (y)53e64 38.465e74 38.3�75 23.3

Formal education (y)0e6 70.07e12 18.1�13 11.9

BMI (kg/m2)<19 7.519e27 81.2>27 11.3

Smoking statusNo 80.0Yes 20.0

Drinking > once/wkNo 74.7Yes 25.3

Betel nut quid chewingNo 95.1Yes 4.9

Physical exercise (times/wk)0 34.70e2 5.8�3 59.5

Fell during the past 12 monthsNo 82.0Yes 18.0

Please cite this article in press as: Tsai AC, Lai M-Y, Mini Nutritional Assefuture risk of falling in older adults e Results of a national cohort study,

Table 2 shows subjects’ MNA item-score patterns. About 6.3% ofsubjects had poor appetite; 11.4% lost weight unexpectedly; 7.6%had psychological stress; 11.6% had pressure sore/skin lesion; 10.8%had inadequate intake of protein-rich foods; 5.3% had inadequatefruit/vegetable intake; 13.4% had poor self-rated nutritional status;and 37.6% had poor self-rated health. According to MNA-T2, 6.5%were at risk of under-nutrition (MNA score <24) and 93.5% werenormal (MNA score �24). According to MNA-T2-SF 13.6% were atrisk of under-nutrition (MNA-SF score �11) and 84.6% were normal(MNA score �12).

Table 3 shows the ability of the nutritional status rated with theMNA and MNA-SF in predicting falling three years later in olderTaiwanese according to multivariate binary logistic regressionmodels, respectively. Results showed that subjects who were ratedas at risk of malnutrition (malnourished þ at risk of malnutrition)at baseline were 87% more likely to have falling three years lateraccording to the full-MNA (OR ¼ 1.87, 95% confidenceinterval ¼ 1.33e2.63, p < 0.001), and 39% more (1.39, 1.07e1.80,p ¼ 0.014) according to MNA-T2-SF. Both models were controlledfor the same socio-demographic, lifestyle and health-relatedvariables.

Figure 1 shows the flowchart of the Taiwan Longitudinal Surveyon Aging (TLSA). Figure 2 shows the receiver operating character-istic (ROC) curves for the MNA-T2 and MNA-T2-SF in predictingfalling in 3118 � 53-year old Taiwanese stratified by age (53e64 vs.�65-year old). For subjects �65 years old (Chart A), the area underthe curve (AUC) and 95% confidence interval (95%CI) ¼ 0.635(0.609e0.660), p< 0.001, sensitivity¼ 0.659 and specificity¼ 0.548for MNA-T2; and AUC 95%CI ¼ 0.597 (0.570e0.624), p < 0.001,sensitivity ¼ 0.504 and specificity ¼ 0.680 for MNA-T2-SF. Forsubjects <65 years old (Chart B), the respective values were0.637(0.588e0.685), p < 0.001, 0.720 and 0.498 for MNA-T2; 0.566(0.514e0.618), p ¼ 0.010, 0.329 and 0.806 for MNA-T2-SF.

4. Discussion

4.1. Predicting the risk of falling with the MNA

Results show that both the MNA and MNA-SF significantlypredicted the risk of falling 3 years later in older Taiwanese. Severalstudies have suggested an association between nutritional statusand the risk of falling in older persons. In a cross-sectional study,Neyens et al.18 found that malnutrition is associated with anincreased risk of falling and impaired physical activity in Dutchlong-term care residents. Bauer et al.25 found that malnutritionwasassociated with falls in the acute care setting. Salvi et al.26 observedthat the MNA-SF could identify those with poor clinical outcomeand predicted functional decline which is a risk factor of falling inolder persons.27,28 Salva et al.1 analyzed the risk factors for falls incommunity-dwelling adults with dementia and found that fallershad worse nutritional status according to the MNA and worse basicactivities of daily-living dependency. Andre et al.17 graded thenutritional status of community-dwelling elderly and found thatthose who had a fall history had lower MNA score compared withthose who did not (MNA score 18.3 vs. 21.0). However, Isenringet al.29 failed to find an association in a sample of community-dwelling elderly persons. The study observed a low prevalence ofmalnutrition, probably because it excluded persons who had oculardisease, recent or recurrent history of musculoskeletal injury andsurgery, walking aid use, and neurological problems, the conditionsthat are more likely to be associated with malnutrition. These re-sults generally suggest an association between nutritional status(as rated with the MNA) and the risk of falling in older persons inboth clinical settings and community settings. The present study

ssment and short-form Mini Nutritional Assessment can predict theClinical Nutrition (2013), http://dx.doi.org/10.1016/j.clnu.2013.10.010

Table 2Subjects’ MNA-item scoring patterns at baseline (N ¼ 3118).

Item (item score) Assigned score n (%)

A. Appetite status over past 3 m (2)Severely poor 0 194(6.2)Moderately poor 1 2(0.1)No loss of appetite 2 2922(93.7)

B. Weight loss during the last 3 m (3)Weight loss >3 kg 0 357(11.4)Does not know 1 3(0.1)Between 1 and 3 kg 2 0(0)No weight change 3 2758(88.5)

C. Mobility (2)Bed or wheelchair-bound 0 9(0.3)Can get out of bed/chair but can’t go out 1 59(1.9)Can move around 2 3050(97.8)

D. Have suffered psychological stress (2)Yes 0 237(7.6)No 2 2881(92.4)

E. Neuropsychological problems (2)Severe dementia or depression 0 1(0.0)Mild dementia 1 21(0.7)No psychological problem 2 3096(99.3)F.BMI status (not included in T2)

G. Can live independently (1)No 0 19(0.6)Yes 1 3099(99.4)

H. Taking >3 prescribed medicines (1)Yes 0 78(2.5)No 1 3040(97.5)

I. Pressure sore or skin lesions (1)Yes 0 361(11.6)No 1 2757(88.4)

J. Whole meals eaten per day (2)1 0 1(0.0)2 1 117(3.8)3 2 3000(96.2)

K. Consumption of protein-rich foods (1)0 or 1 “yes” 0 36(1.2)2 “yes” 0.5 299(9.6)3 “yes” 1 2783(89.3)

L. Consumes �2 servings of fruits or vegetable (1)No 0 165(5.3)Yes 1 2953(94.7)

N. Mode of eating (2)Cannot eat with help (including tube-feeding) 0 5(0.2)Can self-feed but with difficulty 1 29(0.9)Can self-feed without difficulty 2 3084(98.9)

O. Self-rated nutritional status (2)Malnourished 0 181(5.8)Uncertain of status 1 238(7.6)No nutritional problems 2 2699(86.6)

P. Self-rated health status relative to peers (2)Not as good as others 0 1173(37.6)Not sure or about the same 1 1785(57.2)Better than others 2 160(5.1)

Q. MAC (cm) (2)<22.5/21 0 45(1.4)22.5e23.5/21e22 1 57(1.8)>23.5/22 2 3016(96.7)

R. CC (cm) (3)<28/25 0 30(1.0)28e29/25e26 1 34(1.1)29e30/26e27 2 47(1.5)�30/27 3 3007(96.4)

MAC ¼ Mid-arm circumference; CC ¼ calf circumference; MNA ¼ Mini NutritionalAssessment; T2 ¼ Taiwan version-2; SF ¼ short-form.ItemM (fluid intake) was not available. The final score was proportionately adjusted

A.C. Tsai, M.-Y. Lai / Clinical Nutrition xxx (2013) 1e64

Please cite this article in press as: Tsai AC, Lai M-Y, Mini Nutritional Assefuture risk of falling in older adults e Results of a national cohort study,

has added further evidence that the MNAs are able predicting thefuture risk of falling in older persons at the population level.

4.2. The impact of malnutrition on the risk of falling

Malnutrition in the elderly usually develops inconspicuouslyand slowly over a period of years. In a 10-year longitudinal study,Vellas et al.30 observed that the decrease in nutritional intake oc-curs before hospitalization or clinical disease in elderly patients,and long before the decrease of body weight or serum albumin.Chronic PEM can lead to sarcopenia and osteoporosis, conditionsthat increase the risk of falling, bone fracture, immobility, and evenpremature death.9 Vivanti et al.8 examined the association betweennutritional status and falls in hospitalized patients and found thatthose who were rated as malnourished had reduced mobility andpoorer nutritional status. Zoltick et al.31 found that older men andwomenwho had lower dietary protein intake were associated withhigher risk of subsequent falls. Thus, poor nutrition is a risk factorfor falling. Early detection of changes in nutritional status isessential for instituting a feeding program.

We have found that older adults who are rated as at risk ofmalnutrition with the MNA or the MNA-SF have increased futurerisk of falling and the present study is the first to observe this MNAcapability in a large population-based sample. It is particularlyimportant to note that the MNA-SF is quite effective in predictingthe risk of falling even three years later. It is probable that the short-term predictive ability would be much stronger. Unfortunately,because the TLSA survey collects data only every four years, we arenot able to analyze the shorter-term predictive ability.

We also found that both MNAs are slightly better in predictingthe risk of falling in older (�65-year old) persons than in younger(53e64-year old) persons. This slight difference could conceivablybe due to two reasons: (a) fewer falls are due to frailty, functionalimpairment, or malnutrition in younger persons than in olderpersons, and (b) the MNAmay be less accurate in pinpointing thosewho are at risk of malnutrition in the middle-aged persons since itis a tool designed for rating the nutritional status of elderly persons.

It is clinically significant to observe that the MNA-SF performsrelatively well in both age groups compared to the full MNA. TheMNA-SF consists only 1/3 of the items (6 vs.18) and takes about 1/3of the time (5 vs. 15e20 min) to complete the screening processcompared to the full MNA. The time-saving feature makes theMNA-SF a more acceptable/desired screening tool for routineapplication than the full MNA in most of today’s high-demandingclinical settings. Those who are identified as at risk of malnutri-tion should be further evaluated for the risk of falling and be givenproper intervention.

4.3. The MNA and MNA-SF are multifunctional scales

It appears that the MNA and the MNA-SF have multifunctionalproperties, presumably because both scales include items thatreflect functional, mental, neuropsychological, and general healthconditions. The MNA and especially the MNA-SF (because of itssimplicity effectiveness and efficiency) appear to be well suited toserve as a general screening tool for identifying at risk individualsfor nutritional support.14 The tool may also be useful for identifyingthose who are at risk of falling. It is possible that with further im-provements, the MNA-SF can be developed into a simple, effective

to a 30-point basis. MNA-T2-SF included items A, B, C, D, E and R.MNA-T2 rated 203 (6.5%) subjects as at risk of malnutrition (MNA score �23.5) and2915 (93.5%) as normal (MNA score �24); MNA-T2-SF rated 425 (13.6%) as at risk ofmalnutrition (MNA-T2-SF <12) and 2693 (86.4%) as normal (MNA-T2-SF�12).

ssment and short-form Mini Nutritional Assessment can predict theClinical Nutrition (2013), http://dx.doi.org/10.1016/j.clnu.2013.10.010

Table 3Logistic regression analyses of the ability of the baseline (1999) nutritional statusrated with the MNA-T2 andMNA-T2-SF in predicting incident falling four years later(2003) in older Taiwanese (N ¼ 3118).

Nutritionalstatus

MNA-T2a MNA-T2-SFa

% Oftotal

% Fell OR(95% CI)

p % Oftotal

% Fell OR(95% CI)

p

Normal 93.5 17.0 1 86.4 16.8 1At riskb 6.5 32.5 1.87

(1.33e2.63)<0.001 13.6 25.6 1.39

(1.07e1.80)0.014

MAC ¼ Mid-arm circumference; CC ¼ calf circumference; MNA ¼ Mini NutritionalAssessment; T2 ¼ Taiwan version-2; SF ¼ short-form.OR (95% CI) ¼ odds ratio (95% confidence interval); Nagelkerke R2 was 0.127 forMNA-T2 and 0.124 for MNA-T2-SF.

a All models were adjusted for age, sex, education, marital status, living status,alcohol-drinking, smoking, betel-nut chewing, physical activity, BMI, hypertension,diabetes, heart disease, stroke, osteoarthritis, hip fracture, gout, lower-back pain,visual acuity, hearing ability, incontinence and nutritional status at baseline.

b MNA scores � 23.5 for the full MNA, or �11 for the MNA-SF were considered asat risk of malnutrition.

A.C. Tsai, M.-Y. Lai / Clinical Nutrition xxx (2013) 1e6 5

and efficient multipurpose screening tool for identifying elderlywho are at risk of malnutrition, falling, frailty or even the onset ofcare-need. Early detection of nutritional inadequacy can help pre-vent the development of malnutrition and reduce the risk of fallingin older adults. With the rapid increase in most elderly populationsaround the world, such a tool would seem to be particularlywelcome in both clinical and community settings.

4.4. Limitations of the study

The study has some limitations. (a) The survey was carried outevery four years and the incidence of falling was for the past 12months. Thus, no data were available for the years immediatelyfollowing the baseline survey. It is most probable that the predic-tive ability would be even better for those immediate follow-upyears. (b) Data from surveys are generally “self-report”. Self-reports have acceptable accuracy in general, but inaccurate recallsare unavoidable. (c) The present study did not attempt to analyze

Fig. 2. Receiver operating characteristic (ROC) curves for the full- and short-form-Mini Nutr3118 male and female �53 years old Taiwanese. For subjects �65 years old (Chart A), the arp < 0.001, sensitivity ¼ 0.659 and specificity ¼ 0.548 for MNA-T2; and AUC 95%CI ¼ 0.597 (0subjects <65 years old (Chart B), the respective values were 0.637(0.588e0.685), p < 0.001,T2-SF.

Please cite this article in press as: Tsai AC, Lai M-Y, Mini Nutritional Assefuture risk of falling in older adults e Results of a national cohort study,

those who had multiple falls vs. those who had only one during thepast 12 months. (d) The study excluded those subjects who had nodata on BMI, falling, or other variables. It is probable that theseindividuals have poorer health status andmight have greater risk offalling. (e) The study analyzed the prediction of any fall by the MNAdisregard of personal or environmental conditions, or the severityof bodily injuries. The complexity of these variables makes itdifficult to include these factors in the analysis.

5. Conclusion

Results show that older adults who are rated as at risk ofmalnutrition by the full-MNA or MNA-SF have increased risk offalling three years later, suggesting that theMNA and its short-formare capable of predicting future risk of falling in the elderly. Theseresults underscore the multifunctional nature of the MNA andMNA-SF. The MNA-SF is well suited to serve as a reasonable alter-native to the full MNA in today’s clinical settings. The simplicity,effectiveness, efficiency, non-invasiveness and acceptable predic-tive ability of the MNA-SF make it a highly desired tool in clinicaland community settings.

Statement of authorship

ACT conceived the idea, directed the study, and drafted themanuscript; MYL performed the statistical analysis and reviewedthe manuscript.

Conflict of interest statement

No potential conflicts of interest were disclosed.

Sources of funding

The study received no financial support from any source.

itional Assessment Taiwan Version-2 (MNA-T2 and MNA-T2-SF) in a national cohort ofea under the curve (AUC) and 95% confidence interval (95%CI) ¼ 0.635 (0.609e0.660),.570e0.624), p < 0.001, sensitivity ¼ 0.504 and specificity ¼ 0.680 for MNA-T2-SF. For

0.720 and 0.498 for MNA-T2; 0.566 (0.514e0.618), p ¼ 0.010, 0.329 and 0.806 for MNA-

ssment and short-form Mini Nutritional Assessment can predict theClinical Nutrition (2013), http://dx.doi.org/10.1016/j.clnu.2013.10.010

A.C. Tsai, M.-Y. Lai / Clinical Nutrition xxx (2013) 1e66

Acknowledgments

This study is based on data from the “Taiwan Longitudinal Studyon Aging” (TLSA), conducted by the Bureau of Health Promotion ofTaiwan. Descriptions or conclusions herein do not represent theviewpoint of the Bureau.

References

1. Salva A, Roque M, Rojana X, Inzitari M, Andrieu S, Schiffrin EJ, et al. Falls andrisk factors for falls in community e dwelling adults with dementia (NutriAlzTrial). Alzheimer Dis Assoc Disord 2012;26:74e80.

2. Ferrer A, Formiga F, Plana-Ripoll O, Tobella MA, Gil A, Pujol R, et al. Risk of fallsin 85-year-olds is associated with functional and cognitive status: the Octabaixstudy. Arch Gerontol Geriatr 2012;54:352e6.

3. Tinetti ME, Speechley M, Ginter SF. Risk factors for falls among elderly personsliving in the community. N Engl J Med 1988;319:1701e7.

4. De Vries OJ, Peeters GM, Elders PJ, Muller M, Knol DL, Danner SA, et al.Multifactorial intervention to reduce falls in older people at high risk ofrecurrent falls: a randomized controlled trial. Arch Intern Med 2010;170:1110e7.

5. Cameron ID, Murray GR, Gillespie LD, Robertson MC, Hill KD, Cumming RG,et al. Interventions for preventing falls in older people in nursing care facilitiesand hospitals. Cochrane Database Syst Rev 2010;(1). http://dx.doi.org/10.1002/14651858.CD005465.pub2. Art. No.: CD005465.

6. Hartikainen S, Lonnroos E, Louhivuori K. Medication as a risk factor for falls:critical systematic review. J Gerotol Med Sci 2007;62A:1172e81.

7. Vivanti A, McDonald CK, Palmer MA, Sinnott M. Malnutrition associated withincreased risk of frail mechanical falls among older people presenting to anemergency department. Emerg Med Australas 2009;21:386e94.

8. Vivanti A, Ward N, Haines T. Nutritional status and associations with falls,balance, mobility and functionality during hospital admission. J Nutr HealthAging 2011;15:388e91.

9. Vanltallie TB. Frailty in the elderly: contributions of sarcopenia and visceralprotein depletion. Metabolism 2003;52:22e6.

10. Guigoz Y, Vellas B, Garry P. Mini nutritional assessment: a practical assessmenttool for grading the nutritional state of elderly patients. Facts, Res Gerontol1994;4(Suppl. 2):15e59.

11. Guigoz Y. The mini nutritional assessment (MNA) review of the literature-whatdose it tell us? J Nutr Health Aging 2006;10:466e85.

12. Rubenstein LZ, Harker JO, Salva A, Guigoz Y, Vellas B. Screening for undernu-trition in geriatric practice: developing the short-form Mini-NutritionalAssessment (MNA-SF). J Gerontol Series A: Med Sci 2001;56A:M366e72.

13. Kaiser MJ, Bauer JM, Uter W, Donini LM, Stange I, Volkert D, et al. Prospectivevalidation of the modified Mini Nutritional Assessment short-forms in thecommunity, nursing home, and rehabilitation settings. JAG 2011;59:2124e8.

14. Cereda E. Mini nutritional assessment. Curr Opin Clin Nutr Metab Care 2012;15:29e41.

15. Cereda E, Valzolgher L, Pedrolli C. Mini nutritional assessment is a good pre-dictor of functional status in institutionalised elderly at risk of malnutrition.Clin Nutr 2008;27:700e5.

Please cite this article in press as: Tsai AC, Lai M-Y, Mini Nutritional Assefuture risk of falling in older adults e Results of a national cohort study,

16. Lee L, Tsai AC. Mini-nutritional assessment predicts functional decline ofelderly Taiwanese: result of a population representative sample. Br J Nutr2012;107:1707e13.

17. Andre MB, Dumavibhat N, Ngatu NR, Eitoku M, Hirota R, Suganuma N. MiniNutritional Assessment and functional capacity in community-dwelling elderlyin Rural Luozi, Democratic Republic of Congo. Geriatr Gerontol Int 2012. http://dx.doi.org/10.1111/j.1447-0594.2012.00852.x.

18. Neyens J, Halfens R, Spreeuwenberg M, Meijers J, Luiking Y, Verlaan G, et al.Malnutrition is associated with an increased risk of falls and impaired activityin elderly patients in Dutch residential long-term care (LTC): a cross-sectionalstudy. Arch Geront Geriatr 2013;56:265e9.

19. Bureau of Health Promotion, Department of Health, Taiwan. Survey of theelderly in Taiwan [accessed 10.06.12], <http://www.bhp.doh.gov.tw/BHPnet/Portal/Them.aspx?No ¼200712270002>; 1989.

20. Tsai AC, Yang SF, Wang JY. Validation of population-specific Mini-NutritionalAssessment with its long-term mortality-predicting ability: results of apopulation-based longitudinal 4-year study in Taiwan. Br J Nutr 2010;104:93e9.

21. Tsai AC, Ku PY, Tsai JD. Population-specific anthropometric cutoff standardsimprove the functionality of the mini nutritional assessment without BMI ininstitutionalized elderly in Taiwan. J Nutr Health Aging 2008;12:696e700.

22. Tsai AC, Chang TL, Wang YC, Liao CY. Population-specific short-form mininutritional assessment with body mass index or calf circumference can predictrisk of malnutrition in community-living or institutionalized elderly people inTaiwan. J Am Diet Assoc 2010;110:1328e34.

23. Tsai AC, Ho CS, Chang MC. Population-specific anthropometric cut-pointsimprove the functionality of the mini nutritional assessment (MNA) inelderly Taiwanese. Asia Pac J Clin Nutr 2007;16:656e62.

24. Tsai AC, Chou YT, Chang TL, Chang-Lee SN, Tsay SF. A modified mini nutritionalassessment without BMI can effectively assess the nutritional status ofneuropsychiatric patients. J Clin Nurs 2009;18:1916e22.

25. Bauer JD, Isenring E, Torma J, Horsley P, Martineau J. Nutritional status of pa-tients who have experienced a fall in an acute care setting. J Hum Nutr Diet2007;20:558e64.

26. Salvi F, Giorgi R, Grilli A, Morichi V, Espinosa E, Spazzafumo L, et al. MiniNutritional Assessment (short form) and functional decline in older patientadmitted to an acute medical ward. Aging Clin Exp Res 2008;20:322e8.

27. Bloch F, Thibaud M, Dugué B, Brèque C, Rigaud AS, Kemoun G. Episodes offalling among elderly people: a systematic review and meta-analysis of socialand demographic pre-disposing characteristics. Clinics (Sao Paulo) 2010;65:895e903.

28. Kwan MMS, Close JCT, Wong AKW, Lord SR. Falls incidence, risk factors, andconsequences in Chinese people: a systematic review. J Am Geriatr Soc2011;59:536e43.

29. Isenring E, Baker J, Kerr G. Malnutrition and falls risk in community-dwellingolder adults. J Nutr Health Aging 2013;17:277e9.

30. Vellas B, Hunt WC, Romero LJ, Koehler KM, Baumqartner RN, Garry PJ. Changesin nutritional status and patterns of morbidity among free-living elderly per-sons: a 10-year longitudinal study. Nutrition 1997;13:515e9.

31. Zoltick ES, Sahni S, McLean RR, Quach L, Casey VA, Hannan MT. Dietary proteinintake and subsequent falls in older men and women: the Framingham Study.J Nutr Health Aging 2011;15:147e52.

ssment and short-form Mini Nutritional Assessment can predict theClinical Nutrition (2013), http://dx.doi.org/10.1016/j.clnu.2013.10.010