JAMDA 14 (2013) 53e57 JAMDA. journal homepage:

Transcription

1 JAMDA 14 (2013) 53e57 JAMDA journal homepage: Original Study Body Mass Index as a Predictor of All-Cause Mortality in Nursing Home Residents During a 5-Year Follow-up Nicola Veronese MD a, *, Marina De Rui MD a, Elena Debora Toffanello MD a, Irene De Ronch MD a, Egle Perissinotto MD b, Francesco Bolzetta MD a, Barbara D Avanzo MD a, Fabrizio Cardin MD c, Alessandra Coin PhD a, Enzo Manzato PhD a, Giuseppe Sergi PhD a a Department of MedicineeDIMED, Geriatrics Division, University of Padova, Padova, Italy b Department of Cardiac, Thoracic and Vascular Sciences, University of Padova, Padova, Italy c Geriatric Department, Division of Geriatric Surgery, University of Padova, Italy abstract Keywords: Body mass index nursing home mortality elderly nutrition Background: Body mass index (BMI) is considered a short-term mortality predictor, but a consensus has not been reached on its role and that of other nutritional parameters in predicting long-term mortality in nursing home residents. Objectives: To correlate BMI, Mini Nutritional Assessment scores, and serum albumin levels with the 5-year mortality rate in institutionalized elderly subjects. Methods: A total of 181 nursing home residents aged 70 years were included in a 5-year longitudinal study. Data were collected on all participants nutritional, health, cognitive, and functional status by means of a comprehensive geriatric assessment. Data on the participants vital status were obtained 5 years after beginning the study, and a survival analysis was conducted using KaplaneMeier curves and multivariate Cox proportional hazards models. Results: The 5-year mortality rate was 63%. The deceased subjects (n ¼ 115) had a lower BMI ( vs kg/m 2 ; P ¼.03) and Mini Nutritional Assessment score ( vs ; P ¼.02) than those still alive. Serum albumin levels did not differ between the two groups. Among the three indicators of nutritional status considered in this study, only BMI 30 kg/m 2 was significantly associated with a lower mortality risk at 5 years (hazard ratio ¼ 0.432; 95% CI ; P ¼.04), the risk for death being greater the lower the BMI class (log-rank test: P <.001). Conclusions: Our findings suggest that BMI is the best of the three parameters considered as a nutritional predictor of nursing home residents mortality in the longer term, and indicate that a lower mortality risk coincides with a higher BMI. Published by Elsevier Inc. on behalf of the American Medical Directors Association, Inc. Malnutrition, which is related to frailty, physical disability, and a higher mortality risk, 1 is a common condition among institutionalized elderly adults, affecting 60% to 80% of nursing home residents. 2 Adequately assessing these people s nutritional status might enable appropriate measures to be adopted, possibly having a significant impact on their morbidity and mortality, particularly for the most severely disabled. 3 Several tools have been developed for the clinical assessment of nutritional status, including anthropometric parameters [ie, body mass index (BMI)], biochemical measures (serum albumin), and the Mini Nutritional Assessment (MNA), which is the multidimensional method most validated for use in geriatrics. The authors declare no conflicts of interest. * Address correspondence to Nicola Veronese, MD, Clinica GeriatricaeOspedale Giustinianeo (2 piano), Via Giustiniani 2, 35128, Padova, Italy. address: ilmannato@gmail.com (N. Veronese). BMI should be considered one of the most important parameters for identifying malnutrition in elderly adults. 4 Its relationship with mortality has been described as a U-shaped curve, where a greater risk coincides with the extremes of the BMI spectrum, or as an inverse J-shaped curve, with a greater risk only in association with the lowest BMI values. 5e7 Among the biochemical parameters considered, hypoalbuminemia has a strong predictive value for mortality in a number of diseases, 8 as well as in healthy elderly adults. 9 Finally, scores obtained with the MNA have been found to correlate with mortality in elderly individuals living at home and in elderly hospital patients. 10 Several studies on the relationship between nutritional status and mortality among nursing home residents indicate the BMI as the main short-term mortality predictor. Kaiser et al 11 found that the 1-year mortality rate among residents with a BMI > 30 kg/m 2 was lower (12.8%) than for residents whose BMI was <20 kg/m 2 (58.8%) /$ - see front matter Published by Elsevier Inc. on behalf of the American Medical Directors Association, Inc.

2 54 N. Veronese et al. / JAMDA 14 (2013) 53e57 In contrast, Volpato et al 12 found no association between BMI levels and mortality risk, with normal or overweight subjects having much the same 4-year mortality risk as underweight individuals. Cereda et al 13 also found that neither BMI nor MNA score correlated with nursing home residents mortality risk after the 6-year follow-up. Thus, despite the numerous studies conducted on this topic, there is still no consensus in the literature on the role of BMI and other nutritional parameters in predicting short- or long-term mortality in nursing home residents. Knowing which nutritional parameter might best predict longterm mortality in nursing home residents might help clinicians to arrange appropriate dietary measures and nutritional strategies for this particular group of frail elderly individuals. We hypothesized that, among all the conventional nutritional parameters and assessment tools generally applied in geriatrics, BMI alone might suffice as a predictor of the long-term mortality risk for frail elderly nursing home residents. The goal of this study, therefore, was to evaluate the association between the most common nutritional parameters/tools used in geriatrics (ie, BMI, serum albumin levels, and MNA scores) and the 5- year mortality rate in a population of elderly nursing home residents, with a view to identifying which parameter can best predict longterm mortality in this particular group of frail older people. Methods Subjects From September 1, 2006, to October 31, 2006, all individuals admitted to the Istituto di Riposo per Anziani, a state-funded nursing home in Padova, Italy, providing care for elderly people who are no longer self-sufficient, were screened for eligibility for a 5-year longitudinal observation study. Residents older than 70 years who were moderately to severely impaired in the activities of daily living (ADL) and had resided at the nursing home for at least 2 months before the enrollment process began were considered eligible for the study. The exclusion criteria were a diagnosis of malignancy, parenteral or enteral artificial nutrition, past or present severe psychiatric diseases, and evidence of acute illness at the time of observation or during the previous 30 days. Among the 396 elderly adults institutionalized at the nursing home, 181 residents met the inclusion/exclusion criteria and were involved in this study. Vital status on January 1, 2012, was assessed for all participants by contacting them directly or by consulting the nursing home s registers, and the mortality rate was recorded. This study was designed in accordance with the Helsinki Declaration and approved by local ethical committee (institutional review board). All participants were fully informed about the study s nature, purpose, procedures, and risks; written informed consent was obtained from all participants, or from relatives or legal guardians for those unable to give their own consent. Measures Participants were assessed by a trained medical researcher who conducted a comprehensive geriatric assessment. The data detailed below were collected for each participant. Anthropometric Measurements Body weight and height were measured to the nearest 0.1 kg and 0.1 cm, respectively, using a standard balance and stadiometer (Seca, Hamburg, Germany) with subjects barefoot and wearing light clothes. In bedridden patients, weight was obtained using a medical patient lift scale, whereas height was estimated from their knee-to-heel length according to Chumlea s method. The body mass index (BMI) was calculated (kg/m 2 ). Comorbidities and Indicators of Disease Severity Comorbidity was assessed using the Cumulative Illness Rating Scale (CIRS), a validated physician-rated index calculated by collecting the subject s medical history as well as conducting a physical examination and performing laboratory tests. The CIRS classifies comorbidities among 13 organ systems and grades each condition from 0 (no problem) to 4 (severely incapacitating or life-threatening conditions). The comorbidity index is given by the number of conditions graded as 3. The severity index is the mean of the severity scores for each of the 13 organ systems. 14 Cognitive and functional status was assessed using the Short Portable Mental Status Questionnaire (SPMSQ). 15 Pressure sore risk was assessed with the Exton-Smith scale (ESS). 16 Nutritional status was evaluated with the MNA tool, an internationally validated method consisting of 18 items covering anthropometric measures, health status, dietary patterns, and subjective assessments of the participant s nutritional and health status. A total score 23.5 distinguishes people with a good nutritional status from those at nutritional risk (MNA score ¼ 17e23.5) and those with proteinecalorie malnutrition (MNA score < 17). 17 Functional status was assessed using ADL 18 and instrumental ADL (IADL). 19 Finally, aggregating the total scores obtained in eight different domains (ADL and IADL indexes SPMSQ score, comorbidity index obtained with the CIRS, total MNA score, ESS score, number of drugs taken, and living arrangements) enabled an estimation of the new multidimensional prognostic index (MPI) for each resident. The MPI is a risk score that varies from 0 to 1, where an MPI 0.33 (class I) indicates a low risk for death, MPI (class II) indicates an intermediate risk of death and an MPI 0.67 (class III) a high risk for death at 1 year. 20 Serum albumin levels (g/l) were tested at the central laboratory of the University of Padova using nephelometry. Blood samples were obtained in the morning after a 12-hour overnight fast, kept at 41 C for 1 hour, and then centrifuged at 3000 rpm for 10 minutes at 41 C to obtain serum or plasma. Statistical Analysis Our sample s baseline characteristics were reported as means SD for continuous variables and as frequencies or percentages for categorical variables. Baseline comparisons between the residents who died and those who were still alive were drawn using generalized linear models accounting for age differences. Rank analyses were performed in cases of skewness in the distribution of continuous variables. For the survival analyses, KaplaneMeier curves were calculated for all the categorical variables (sex, diabetes, hypertension, dementia, MNA categories, BMI classes, and MPI classes). To explore whether variables should be included as predictors in the final survival model, we performed the log-rank test of equality across strata for all the categorical variables and Cox univariate proportional hazards regression for all the continuous variables (age, BMI, number of ADL preserved, number of IADL preserved, SPMSQ score, CIRS comorbidity index, number of drugs taken, ESS score, and serum albumin levels). The predictors included in the final model were all the variables with P <.2 in the univariate analyses: sex, age, comorbidity index, serum albumin levels, number of drugs taken, ADL and IADL indexes, MNA risk classes, BMI classes, and MPI classes. BMI was classified in 4 categories: underweight (<20 kg/m 2 ), normal weight (20e24.9 kg/m 2 ), overweight (25e29.9 kg/m 2 ), and obese (30 kg/m 2 ). A BMI threshold of 20 kg/m 2 has been widely accepted as defining underweight in elderly adults. 9 For the purposes of this

3 N. Veronese et al. / JAMDA 14 (2013) 53e57 55 Table 1 Baseline of Nursing Home Residents by Survival Status All (N ¼ 181) Dead (n ¼ 115) Alive (n ¼ 66) Age-Adjusted P value* Mean age, y SD (unadjusted) Female sex, n (%) 146 (81.7) Mean BMI SD, (kg/m 2 ) Mean serum albumin level SD, (g/l) Mean ADL total SD Mean IADL total SD Mean SPMSQ score SD, (0e10) Mean MNA score SD, (0e30) Mean ESS score SD, (5e20) Mean CIRS-CI SD Mean number of drugs taken SD Mean MPI score SD Diabetes mellitus, n (%) 32 (17.9) Hypertension, n (%) 101 (57.3) Dementia, n (%) 125 (70.2) ADL total, activities of daily living preserved; BMI, body mass index; CIRS-CI, Cumulative Illness Rating ScaleeComorbidity index; ESS, Exton-Smith scale; IADL total, instrumental activities of daily living preserved; MNA, Mini Nutritional Assessment; MPI, multidimensional prognostic index; SPMSQ, Short Portable Mental Status Questionnaire. *P values obtained by analysis of variance fitting generalized linear models, accounting for age differences. The c 2 test was used to compare categorical data. study, a BMI in the range of 20 to 24.9 kg/m 2 was taken as a reference category. Hazard ratios (HRs) and 95% CIs were estimated for all the variables entered in the final model. Finally, a graph with the survival functions for the different BMI categories was generated. All analyses were repeated after excluding residents who died in the first year of follow-up. The level of statistical significance was set at All statistical analyses were performed with SAS software, release 9.2 (SAS Institute, Inc, Cary, NC). Results From the beginning of the study, the mortality rate was 3% at 1 year and 63% at 5 years. Residents baseline characteristics in the sample as a whole and according to their survival status are listed in Table 1. Almost all of the participants were female (81.7%), the mean age of the sample as a whole was years (range, 70e102 years), and the mean BMI was kg/m 2 (14.10e42.54). The residents enrolled in this study were almost completely dependent in ADL ( ) and had moderate to severe cognitive impairments (SPMSQ score, ). A total of 167 subjects (92.3%) were incapable of any IADL. The MNA score in the sample as a whole was , indicating a high risk for malnutrition. None of the elderly subjects in our sample was in MPI class 1 (with a low mortality risk); their mean MPI score was , indicating a severe risk for death at 1 year. Compared with the residents still alive at 5 years, those who died had a significantly lower baseline BMI ( vs kg/m 2 ; P ¼.03), a lower number of independent ADL ( vs ; P ¼.01), and a lower MNA score ( vs ; P ¼.02), whereas the baseline MPI score was significantly higher in the deceased than in the living subjects ( vs ; P ¼.006). Serum albumin levels and SPMSQ differed only marginally between the two groups. Table 2 lists our participants distribution by BMI, MNA, and MPI classes vis-à-vis their vital status. The proportions of the residents still alive gradually increased from the underweight to the overweight and obese BMI categories (c 2 test, P ¼.01). The same trend was apparent for the distribution between the MNA classes: the residents who died amounted to 71.1%, 65.3%, and 35.3%, respectively, of the undernourished, at risk for malnutrition, and normal MNA classes (c 2 test, P ¼.03). Finally, the prevalence of the residents who died was significantly greater in the group in MPI class 3 (c 2 test, P ¼.02). Table 3 lists the HR and the 95% CI for all the quantitative and categorical variables identified as significant mortality predictors at univariate analysis (details not shown). Among all the variables entered in the final model, in addition to age at baseline, the predictors of 5-year mortality were sex (HR ¼ for female subjects; P ¼.01), baseline BMI class (HR ¼ for BMI class 30 kg/m 2 ; P ¼.04), and number of drugs being taken at baseline (HR ¼ 1.075; P ¼.04). Neither serum albumin levels (HR ¼ 0.661; P ¼.16) nor MNA class (undernourished subjects: HR ¼ 1.190, P ¼.72; subjects at risk for malnutrition: HR ¼ 1.310, P ¼.63) predicted 5-year mortality. Our results were confirmed even after excluding residents who died in the first year of follow-up from the analyses. Figure 1 shows the survival function, exploring the association between BMI class and time to death. The mortality risk was significantly lower for individuals with a BMI 30 kg/m 2 than for other BMI classes (log-rank test: P <.001). The mortality risk was greater the lower the BMI class. Discussion The association between nutritional parameters and the 5-year mortality rate was examined in a group of frail elderly nursing home residents. After adjusting for several potential confounders (ie, sex, age, comorbidity, and number of drugs taken), residents with a greater baseline BMI were at significantly smaller risk for death than those with a lower BMI. Among the nutritional parameters investigated in this study (ie, BMI, serum albumin levels, and MNA Table 2 Residents Distribution in Body Mass Index, Mini Nutritional Assessment, Multidimensional Prognostic Index Classes at the Baseline, in the Sample as a Whole, and by Survival Status All (N ¼ 181) Dead (n ¼ 115) Survived (n ¼ 66) P value* BMI classes, %.01 <20 kg/m e24.9 kg/m e29.9 kg/m kg/m MNA classes, %.03 <17 (malnutrition) e23.5 (risk for malnutrition) >24 (normal nutrition) MPI classes, % (class 2) (class 3) BMI, body mass index; MNA, Mini Nutritional Assessment; MPI, Multidimensional Prognostic Index. *The c 2 test was used to test the significance of the differences in the distributions.

4 56 N. Veronese et al. / JAMDA 14 (2013) 53e57 Table 3 Risk for 5-Year All-Cause Mortality According to Participants Baseline Hazard ratio 95% CI P value* Female sex e Age, y e Number of drugs taken e Serum albumin (g/l) e ADL total e IADL total e CIRS-CI e BMI (kg/m 2 ) < e e (reference) d d 25e e e MNA score < e e e >24 1 (reference) d d MPI class 2 1 (reference) d d e ADL total, number of preserved activities of daily living; BMI, body mass index; CIRS-CI, Cumulative Illness Rating ScaleeComorbidity index; IADL total, number of preserved instrumental activities of daily living; MNA, Mini Nutritional Assessment; MPI, multidimensional prognostic index. *P values obtained by fitting proportional hazard regression models, accounting for all the variables found to be significant predictors at univariate analyses. score), only BMI appeared to predict 5-year mortality in this group of elderly nursing home residents. The main finding emerging from our study is that the mortality risk was lower in individuals with a BMI 30 kg/m 2 and increased linearly toward lower BMI classes. Our study found that obesity (BMI 30 kg/m 2 ) in advanced age halves the 5- year mortality risk (HR ¼ 0.432; 95% CI, 0.20e0.70), indicating that BMI is not just the expression of fat mass in frail elderly nursing home residents. Our results contrast with a previous study indicating that the relationship between BMI and mortality was no longer significant after adjusting for body cell mass 12 ; this could be because of the different BMI cutoffs used to stratify the sample population, and to the younger age and lower rates of obesity, disability, and cognitive impairment in the sample considered in the previous study than in our present report. The relationship between moderate obesity and mortality in the elderly remains controversial. Janssen et al 21 found moderate obesity to be associated with a slightly greater mortality risk (odds ratio, 1.10; 95% CI, 1.06e1.13), whereas other studies on elderly adults (age 75 years) seem to demonstrate that obesity does not increase the mortality risk. The different role of obesity at different times of life could be explained by a selective survival mechanism; that is, obese individuals surviving into old age seem to be resistant to the negative effect of a high BMI. It is reasonable to hypothesize that people with a greater BMI have higher proportions not only of fat mass but also of fat-free mass, which may help to protect against death. 22 Another possibility is that people with larger fat reserves are better able to withstand periods of low calorie intake associated with acute illness, minimizing the wasting effects associated with chronic illness. 7 The other two nutritional parameters considered in our study, serum albumin levels and MNA, were not significant mortality predictors in our population. Hypoalbuminemia is generally considered an indicator of malnutrition (and protein malnutrition, in particular), 23 but it was no longer significant in predicting death for our nursing home residents after adjusting for potential confounders such as disability and comorbidity index. Similar findings were reported in Volpato et al s 12 study, which found that low serum albumin levels were not a mortality predictor in nursing home residents after 4 years of follow-up. As for MNA, several studies have indicated that undernourished individuals (with MNA scores < 17.5) and those at risk for malnutrition (MNA scores between 17.5 and 24) had a mortality risk 7 and 2.5 times greater, respectively, than those with an adequate nutritional status (MNA scores > 24). 24 In contrast, Beck et al 25 found that MNA had some weaknesses in predicting outcome over periods longer than 12 months, and concluded that this tool might be more sensitive in predicting short-term than long-term mortality rates. Among the non-nutritional variables considered in this study, the number of drugs taken by participants was a significant 5-year mortality predictor. Drug prescription for the elderly is a challenge because the prevalence of chronic diseases increases with age and frequently leads to several drugs being administered. Polypharmacy may carry a greater risk for adverse drug reactions, however, which can frequently lead to hospitalization or rehospitalization and death. 26 In this study, higher amounts of medication used by our nursing home residents was associated with a slightly greater mortality risk, which increased by approximately 7.5% for each drug added, as reported in other studies. 27,28 We consequently analyzed the relationship between mortality and MPI, an aggregated index that has been shown to strongly predict Fig. 1. Survival function, adjusted for sex, age, and number of drugs taken, exploring the association between body mass index (kg/m 2 ) categories and all-cause mortality after a 5-year follow-up. The P values were obtained using the log-rank test of equality across strata.

5 N. Veronese et al. / JAMDA 14 (2013) 53e57 57 short-term mortality in hospitalized patients with several acute diseases. 29 In our sample, residents MPI scores did not predict their long-term mortality (although there were significantly more deaths in the severe risk class, ie, MPI class 3); this is probably because MPI scores tend to level out in institutionalized individuals, as demonstrated also by the absence of any cases in MPI class 1 in our sample, because their living arrangements are the same and their ADL and IADL indexes are fairly homogeneous. MPI might be useful in predicting the mortality risk in nursing home residents when an acute event strongly modifies their score, for example, after a hospital stay, when their disabilities or comorbidities often become more severe. One of the shortcomings of our study lies in that our sample of nursing home residents consisted mainly of women (81.7%), so it cannot be regarded as representative of the general population. Given the age of our study population, the predominance of women is probably due to their longer life expectancy and to the fact that women are institutionalized more frequently than men because of the two sexes traditionally different roles. 30 In addition, we obtained no direct information on muscle mass, physical activity, or body composition, and collecting such data might have shed some useful light to enable a better interpretation of the relationship between BMI and mortality. The main strengths of our study lie in the multidimensional assessment conducted and the long follow-up (considering the short life expectancy of nursing home residents, with their various degrees of frailty and disability). Conclusion Our results appear to suggest that, among the parameters considered, BMI was the best nutritional predictor of long-term mortality in nursing home residents. The mortality risk was greater the lower the BMI, providing important prognostic information on these institutionalized elderly adults. Accurate nutritional and dietary assessments should be performed in all individuals on admission to a nursing home to identify their degree of frailty and prevent any further weight loss. Appropriate nutritional measures for nursing home residents should also take people s dietary habits and preferences into account, their appetite, and potential negative influences on their appreciation of food, even in normal-weight elderly residents. References 1. Sutcliffe C, Burns A, Challis D, et al. Depressed mood, cognitive impairment, and survival in older people admitted to care homes in England. Am J Geriatr Psychiatry 2007;15:708e Salva A, Coll-Planas L, Bruce S, et al. Nutritional assessment of residents in longterm 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;13:475e Buffa R, Floris GU, Putzu PF, Marini E. Body composition variations in ageing. Coll Antropol 2011;35:259e Kimyagarov S, Klid R, Levenkrohn S, et al. Body mass index (BMI), body composition and mortality of nursing home elderly residents. Arch Gerontol Geriatr 2010;51:227e Sergi G, Perissinotto E, Pisent C, et al. An adequate threshold for body mass index to detect underweight condition in elderly persons: The Italian Longitudinal Study on Aging (ILSA). J Gerontol A Biol Sci Med Sci 2005;60: 866e Mazza A, Zamboni S, Tikhonoff V, et al. Body mass index and mortality in elderly men and women from general population. The experience of Cardiovascular Study in the Elderly (CASTEL). Gerontology 2007;53:36e Grawbowski DC, Campbell CM, Ellis JE. Obesity and mortality in elderly nursing home residents. J Gerontol A Biol Sci Med Sci 2005;60:1184e Menon V, Greene T, Wang X, et al. C-reactive protein and albumin as predictors of all-cause and cardiovascular mortality in chronic kidney disease. Kidney Int 2005;68:766e Corti MC, Guralnik JM, Salive ME, Sorkin JD. Serum albumin level and physical disability as predictors of mortality in older persons. JAMA 1994;272: 1036e Donini LM, Savina C, Rosano A, et al. MNA predictive value in the follow-up of geriatric patients. J Nutr Health Aging 2003;7:282e Kaiser R, Winning K, Uter W, et al. Functionality and mortality in obese nursing home residents: An example of risk factor paradox? J Am Med Dir Assoc 2010; 11:428e Volpato S, Romagnoni F, Soattin L, et al. Body mass index, body cell mass, and 4-year all-cause mortality risk in older nursing home residents. J Am Geriatr Soc 2004;52:886e Cereda E, Pedrolli C, Zagami A, et al. Nutritional screening and mortality in newly institutionalised elderly: A comparison between the geriatric nutritional risk index and the mini nutritional assessment. Clin Nutr 2011;30: 793e Parmelee PA, Thuras PD, Katz IR, Lawton MP. Validation of the Cumulative Illness Rating Scale in a geriatric residential population. J Am Geriatr Soc 1995; 43:130e Pfeiffer E. A short portable mental status questionnaire for the assessment of organic brain deficit in elderly patients. J Am Geriatr Soc 1975;23:433e Bliss MR, McLaren R, Exton-Smith AN. Mattresses for preventing pressure sores in geriatric patients. Mon Bull Minist Health Public Health Lab Serv 1966;25: 238e Vellas B, Guigoz Y, Garry PJ, et al. The Mini Nutritional Assessment (MNA) and its use in grading the nutritional state of elderly patients. Nutrition 1999;15: 116e Katz S, Downs TD, Cash HR, Grotz RC. Progress in development of the index of ADL. Gerontologist 1970;10:20e Lawton MP, Brody EM. Assessment of older people: Self-maintaining and instrumental activities of daily living. Gerontologist 1969;9:179e Pilotto A, Ferrucci L, Franceschi M, et al. Development and validation of a multidimensional prognostic index for one-year mortality from comprehensive geriatric assessment in hospitalized older patients. Rejuvenation Res 2008;11:151e Janssen I, Mark AE. Elevated body mass index and mortality risk in the elderly. Obes Rev 2007;8:41e Bigaard J, Frederiksen K, Tjonneland A, et al. Body fat and fat-free mass and allcause mortality. Obes Res 2004;12:1042e Sergi G, Coin A, Enzi G, et al. Role of visceral proteins in detecting malnutrition in the elderly. Eur J Clin Nutr 2006;60:203e Tsai AC, Yang SF, Wang JY. Validation of population-specific Mini-Nutritional Assessment with its long-term mortality-predicting ability: Results of a population-based longitudinal 4-year study in Taiwan. Br J Nutr 2010;10: 93e Beck AM, Holst M, Rasmussen HH. Efficacy of the Mini Nutritional Assessment to predict the risk of developing malnutrition or adverse health outcomes for old people. e-spen 2008;3:102e Sergi G, De Rui M, Sarti S, Manzato E. Polypharmacy in the elderly: Can comprehensive geriatric assessment reduce inappropriate medication use? Drugs Aging 2011;28:509e Franceschi M, Scarcelli C, Niro V, et al. Prevalence, clinical features and avoidability of adverse drug reactions as cause of admission to a geriatric unit: A prospective study of 1756 patients. Drug Saf 2008;31:545e Dale MC, Burns A, Panter L, Morris J. Factors affecting survival of elderly nursing home residents. Int J Geriatr Psychiatry 2001;16:70e Pilotto A, Addante F, Franceschi M, et al. Multidimensional Prognostic Index based on a comprehensive geriatric assessment predicts short-term mortality in older patients with heart failure. Circ Heart Fail 2010;3:14e Wingard DL. Health among older women in the United States. Public Health Rep 1987;102:62e67.