- Email: [email protected]
Body Mass Index and Mortality in Institutionalized Elderly
BRIEF REPORT
Body Mass Index and Mortality in Institutionalized Elderly Emanuele Cereda, MD, PhD, Carlo Pedrolli, MD, Annunciata Zagami, MD, Alfredo Vanotti, MD, Silvano Piffer, MD, Annalisa Opizzi, RD, Mariangela Rondanelli, MD, PhD, and Riccardo Caccialanza, MD
Objective: Malnutrition and sarcopenia in institutions are very common and significantly affect the prognosis. Aging is characterized by weight and lean body mass losses. Accordingly, in elderly patients, body mass index (BMI) is considered a marker of protein stores rather than of adiposity. Current guidelines suggest a BMI 21 kg/m2 or lower as major trigger for nutritional support. We evaluated the association between BMI and mortality in institutionalized elderly. Methods: This was a multicentric prospective cohort study involving 519 long-term care resident elderly individuals. Risk for mortality across BMI tertiles was estimated by the Cox hazards regression model adjusted for potential confounders recorded at inclusion and collected during the follow-up. Results: During a median follow-up of 5.7 years (25th to 75th percentile, 5.2–8.2], 409 (78.8%) elderly patients died. In primary analyses, based on variables collected at inclusion, patients in the first tertile of
Malnutrition in long-term care facilities is a very common comorbidity that negatively affects the patient’s prognosis.1–4 Several causes appear to contribute to this Nutrition and Dietetics Service, Fondazione IRCCS Policlinico San Matteo, Operativa di Dietetica e Nutrizione Clinica, Pavia, Italy (E.C., R.C.); Unita Ospedale ‘‘S. Chiara,’’ Azienda Provinciale per i Servizi Sanitari, Trento, Italy (C.P.); Fondazione Belluria Onlus, Appiano Gentile, Como, Italy (A.Z.); Servizio di Dietetica e Nutrizione Clinica, ASL Como, Como, Italy (A.V.); Servizio Osservatorio Epidemiologico, Direzione per la Promozione e l’Educazione alla Salute, Azienda Provinciale per i Servizi Sanitari, Trento, Italy (S.P.); Servizio Endocrino-nutrizionale, Dipartimento di Scienze Sanitarie Applicate e Psicocomportamentali, Sezione di Nutrizione, Azienda di Servizi alla Persona degli Studi di Pavia, Pavia, Italy (A.O., M.R.). di Pavia, Universita The study was supported by the Fondazione IRCCS Policlinico San Matteo, Pavia, Italy, and by an Investigator Grant from Nutricia, Milano, Italy (to E.C.). Address correspondence to Emanuele Cereda, MD, PhD, Nutrition and Dietetics Service, Fondazione IRCCS Policlinico San Matteo, Viale Golgi 19, 27100 Pavia, Italy. E-mail: [email protected]
Published by Elsevier Inc. on behalf of the American Medical Directors Association, Inc. DOI:10.1016/j.jamda.2010.11.013 174 Cereda et al
BMI (#21 kg/m2) were at higher risk for all-cause (hazard ratio [HR] 1.38; 95% confidence interval [CI] 1.04– 1.84; P 5 .025) and cardiovascular mortality (HR 5 1.49; 95% CI, 1.00–2.08; P 5 .045). Increased risk was confirmed even after adjusting for nutritional support during the follow-up (all-cause HR 5 1.53; 95% CI, 1.13–2.06; P 5 .006; cardiovascular HR 5 1.62; 95% CI, 1.09–2.40; P 5 .018), which in turn was associated with a reduced risk (all-cause HR 5 0.74; 95% CI, 0.55–0.97; P 5 .035; cardiovascular HR 5 0.62; 95% CI, 0.42–0.91; P 5 .016). Conclusion: BMI is significantly associated with all-cause and cardiovascular mortality in institutionalized elderly patients. A value of 21 kg/m2 or lower can be considered a useful trigger for nutritional support. These results support intending BMI as nutritional reserve in institutionalized elderly patients. (J Am Med Dir Assoc 2011; 12: 174–178) Keywords: Body mass index; long-term care; elderly; mortality; malnutrition
picture, particularly age-related factors, multiple chronic diseases, and insufficient attention for nutritional care.1–3,5,6 Aging is frequently characterized by a reduction of food intake, changes in hormone profile (eg, insulin resistance), and proneness to acute diseases that could result in weight loss and sarcopenia.3,5–8 Current guidelines agree that nutritional screening and nutritional support should be part of routine care.1–3 Several screening tools have been introduced, and the use of those providing a multidimensional assessment of the patient is now recommended.9 However, a patient’s evaluation by these tools may require variable amounts of time and adequate training of health care professionals. Body mass index (BMI), being based only on weight and height measurements, is an unsophisticated, inexpensive, and easy-to-interpret method in nutritional assessment. However, mainly in view of age-related changes in body composition, World Health Organization (WHO) criteria for classifying nutritional status on the basis of BMI appear overly restrictive as they apply to the elderly.10–12 Accordingly, guidelines for JAMDA – March 2011
Outcome
nutritional management in long-term care suggest a BMI of 21 kg/m2 or lower, rather than lower than 18.5 kg/m2, as a majortrigger for nutritional support. However, few large prospective studies have examined the relationship between BMI and mortality in long-term care elderly residents.13–15 We designed the present study to further investigate the potential value of this anthropometric indicator in guiding clinical practice.
Vital status and cause of death were ascertained by active follow-up (Como and Pavia) or by means of record linkage with centralized provincial registry (Trento). Data on death were coded according to the International Classification of Diseases, 10th Revision (ICD-10)17 and classified as cardiovascular (CV), respiratory, neoplasms, or others/not reported.
MATERIALS AND METHODS
Statistical Analyses
Study Design We designed a multicentric prospective cohort study in 4 long-term care facilities in the provinces of Como (n 5 1), Pavia (n 5 1), and Trento (n 5 2). Baseline data collection began in May 2002 and ended in May 2007. Recruitment of patients was performed as follows: every year, for 2 weeks, all the subjects newly admitted were assessed for eligibility. Inclusion criteria were age older than 65 years and agreement to participate by written informed consent (direct enquiry or legal guardians). Exclusion criteria were presence of acute infections or illnesses, established terminal diseases, and known neoplastic disorder. The study was performed in adherence with the principles of the Declaration of Helsinki and protocol was approved by local Institutional Ethics Committee. Nutritional Assessment Baseline nutritional assessment by anthropometry was based on body weight, height, knee-height, mid-arm circumference, and triceps skinfold thickness. Particularly, weight (to the nearest 0.1 kg) was measured using a calibrated flat scale (ambulatory patients) or by a chair scale or a hoistprovided weighting device (nonambulatory or bedridden patients). Height (to the nearest 0.1 cm) was directly assessed in those able to stand or estimated from knee height (nonambulatory patients or in case of abnormal spinal curvature).10,16 BMI was calculated as the ratio between weight (kg) and height (m) squared (kg/m2).10 Patients were categorized into tertiles of this index, setting a cutoff of 21 kg/m2 or lower for tertile I (tertile II, 21–25; tertile III, $25) in agreement with current clinical guidelines for nutritional management in long-term care.1 Eight- to 12-hour fasting venous blood samples were also drawn for the evaluation of hemoglobin, total lymphocyte count, serum albumin, prealbumin, transferrin, total cholesterol, and creatinine. Covariates At baseline, information was collected on age, gender, admission diagnosis, and major comorbidities (diabetes and hypertension). Data were retrieved from medical records or ascertained through patient interview and physical examination. The selection of main admission diagnosis groups used in the analyses was based on literature review by reason of an established association with nutritional status and/or survival. Finally, data on prescription of nutritional support (other than by gastrostomy) during the stay were also retrieved through direct consultation of clinical records. BRIEF REPORT
Statistical analyses were performed using MEDCALC for Windows Version 11.3.0.0 (MedCalc Software, Mariakerke, Belgium) with the test’s significance set to a 2-tailed P value of less than .05. Data were presented as mean and standard deviation, median and interquartile range (25th–75th percentile), or counts and percentage. Comparisons of groups were performed using 1-way analysis of variance, Kruskall-Wallis test analysis (continuous variables), or chi-square test (categorical variables). Associations with all-cause and causespecific mortality were assessed by the Cox proportional hazards regression model. Results were presented as hazard ratio (HR) and 95% confidence interval (CI). Previous to inclusion in the analyses, collinearity between variables was checked with the Pearson statistics. Therefore, only variables with a P less than .10 by univariate analysis were retained in the final model. Primary analysis was based only on baseline features of the population. Then, a further adjustment for nutritional support during the follow-up was also considered as secondary analysis. RESULTS In total, 533 patients were assessed during the recruitment phases. Written consent was obtained for every patient. Fourteen patients were lost to follow-up. The distribution and the features of the analyzed cohort (n 5 519; 97.4% of the original sample) according to tertiles of BMI are presented in Table 1. During a median follow-up of 5.7 years (25th–75th, 5.2– 8.2), 409 deaths (78.8%) occurred. Frequency of events according to BMI were as follows: tertile I, 84.9% (n 5 157); tertile II, 78.0% (n 5 138); tertile III, 72.6% (n 5 114) (P 5 .021). CV causes of death were the most frequent (56.5%; coronary artery disease 24.2%, heart failure 25.1%, stroke 32.0%, others or unspecified 18.7%). Other causes were as follows: 17.4%, respiratory diseases (90% by pneumonia); 6.8%, neoplasm; 19.3%, other diseases. Median survival time was significantly reduced in lower BMI tertiles: tertile I 5 2.0 years (25th–75th, 1.7–2.7); tertile II 5 2.3 years (25th–75th, 1.8–3.4); tertile III 5 3.1 years (25th–75th, 2.3–4.9) (P for trend 5 .003). Then, adjusted HRs for allcause and cause-specific mortality were computed using Cox regression models. The HRs for tertile of BMI were calculated taking the highest one as reference standard. After adjusting for confounders recorded at baseline, participants in the lowest tertile of BMI had a significant elevation in risk for all-cause mortality (HR 5 1.38, 95% CI, 1.04–1.84, P 5 .025; Table 2). Risks for cause-specific mortality were also computed. BMI was significantly associated only with death risk for CV causes (tertile I HR 5 1.49, 95% CI Cereda et al 175
Table 1.
Baseline Clinical and Demographic Characteristics of the Population by Body Mass Index Tertiles
Characteristic*
Overall Population (n 5 519)
1st Tertile BMI #21 (n 5 185)
2nd Tertile BMI 21–25 (n 5 177)
3rd Tertile BMI $25 (n 5 157)
P Value†
Male, n (%) Age, mean (SD), y Body mass index, mean (SD), kg/m2 Arm circumference, mean (SD), cm Triceps skinfold, mean (SD), mm Albumin, mean (SD), g/L Prealbumin, mean (SD), mg/dL Transferrin, mean (SD), mg/dL Hemoglobin, mean (SD), g/dL Total lymphocytes count, mean (SD), /mm3 Total cholesterol, mean (SD), mg/dL Creatinine, mean (SD), mg/dL All-type dementia, n (%) COPD, n (%) Heart disease, n (%) Hip fracture, n (%) Stroke, n (%) Other main admission diagnoses, n (%) Diabetes, n (%) Hypertension, n (%) Gastrostomy, n (%)
150 (28.9) 84 (8.4) 23.1 (2.3) 24.6 (3.6) 12.9 (5.8) 35.4 (4.5) 19.7 (7.4) 201 (41) 12.2 (2.4) 1868 (786) 184 (43) 1.10 (0.53) 229 (44.1) 37 (7.1) 55 (10.6) 41 (7.9) 86 (16.6) 71 (13.7) 74 (14.3) 225 (43.3) 28 (5.4)
55 (29.7) 84.4 (8.5) 18.4 (2.0)‡ 21.7 (3.3)‡ 9.1 (5.4)‡ 34.3 (4.3)‡ 18.2 (6.2) 198 (47) 11.9 (2.4) 1816 (850) 174 (42) 0.99 (0.37) 91 (49.2) 24 (13.0) 21 (11.3) 13 (7.0) 20 (10.8) 16 (8.7) 11 (5.9) 65 (35.3) 16 (8.6)
52 (29.4) 84.2 (8.0) 23.0 (1.1)‡ 24.1 (3.6)‡ 11.6 (5.2)‡ 35.6 (4.6) 20.2 (5.6) 203 (35) 12.1 (2.3) 1833 (624) 187 (42) 1.09 (0.46) 83 (46.9) 10 (5.6) 13 (7.4) 13 (7.4) 36 (20.3) 22 (12.4) 29 (16.7) 85 (47.9) 8 (4.5)
53 (33.7) 83.2 (8.16) 28.6 (3.3)‡ 27.7 (3.9)‡ 17.6 (6.7)‡ 36.3 (4.5) 20.6 (9.5) 201 (39) 12.4 (2.6) 1958 (891) 191 (44)‡ 1.22 (0.69)‡ 55 (35.0) 3 (1.9) 21 (13.4) 15 (9.6) 30 (19.1) 33 (21.0) 34 (21.4) 75 (47.6) 4 (2.5)
.879 .359 \.001 \.001 \.001 \.001 .059 .628 .458 .555 .025 .029 .021 \.001 .186 .651 .068 .002 \.001 .019 .037
BMI, body mass index (kg/m2); COPD, chronic obstructive pulmonary disease; SD, standard deviation. * Data are reported as mean (SD) , median (25th–75th) or as counts (%). Percentages are calculated within single groups. † Continuous and categorical variables were compared between groups with 1-way analysis of variance, the Kruskall Wallis test, or the chisquare test, respectively. ‡ Significantly different versus all the other groups by post-hoc test.
1.00–2.08, P 5 .045; linear increase over tertiles HR 5 0.83, 95% CI 0.68–1.00, P 5 .048). During the follow-up 128 (24.7%) patients were prescribed a nutritional support other than by gastrostomy (93% sip feeding). Frequency of nutritional support was significantly higher in the lowest tertile of BMI: tertile I, 34.6% (n 5 64); tertile II, 19.2% (n 5 34); tertile III, 19.1% (n 5 30). Analyses adjusted for this further confounder confirmed the associations found in primary analyses between BMI and both all-cause (HR 5 1.53, 95% CI 1.13–2.06, P 5 .006) and CV (HR 5 1.62, 95% CI 1.09–2.40, P 5 .018; for linear increase over tertile of BMI, HR 5 0.78, 95% CI 0.64–0.96, P 5 .018) mortality. Nutritional support was also associated with reduced risk (all-cause HR 5 0.74, 95% CI 0.55–0.97, P 5 .035; CV HR 5 0.62, 95% CI 0.42–0.91, P 5 .016). Interestingly, analyses restricted to patients in the lowest tertile of BMI showed that nutritional support was associated with reduced risk of both all-cause (HR 5 0.59, 95% CI 0.39– 0.91, P 5 .018) and cardiovascular mortality (HR 5 0.45, 95% CI 0.26–0.81, P 5 .008). The associations were confirmed even if nutritional status groups were set to exact tertiles of the population (I, #20.7, n 5 175; II, 20.8–24.6, n 5 175; III, $24.7, n 5 169). DISCUSSION The present study demonstrated that BMI is associated with increased mortality from all-cause and cardiovascular disease in long-term care elderly residents. In this respect, we also strengthened the recommendation of current guidelines1,2 that a value of 21 kg/m2 or lower should be considered 176 Cereda et al
a major trigger for nutritional support. Moreover, this was the first prospective study reporting a reduction in death risk in patients receiving nutritional support. In the general adult population the association between BMI and mortality has been described as a J-shaped curve with higher increase in risk for smaller variation of BMI at low body weight rather than in overweight and obesity.17–19 These studies provided compelling evidence that World Health Organization’s BMI thresholds (mainly based on morbidity risk) are significantly lower than those associated with reduced death risk. A value lower than 20 kg/m2 is now accepted as underweight in developed countries.20 Ad hoc studies or subgroup analyses have reported a U-shaped relationship in elderly persons, with a less steep curve on the right; a wide, flat bottom; and a minimum risk shifted to higher BMI values.11,12,17–19,21,22 Different threshold values for the elderly should account for a greater variability in functional and metabolically active components of the body in advanced age, which have been shown to explain most of the outcome prediction by BMI.13,14,22 Moreover, aging is associated with height reduction owing to accentuation of spinal curvature and narrowing of intervertebral disk spaces.23 BMI shows limitations in reflecting body composition and it is likely to reflect lean body mass stores in elderly persons. Aging is characterized by a reduction in skeletal muscle mass because of multiple contributing and interacting factors, such as reduced physical activity, presence of chronic diseases, endocrine changes, increased inflammatory background, and primary and/or secondary anorexia.5–8 In the presence of an impaired energy balance, older people are prone to higher JAMDA – March 2011
Table 2.
Predictors*of All-Cause Mortality at Multivariable Cox Proportional Hazard Regression Models
Characteristic†
Dead (n 5 409)
Survivors (n 5 110)
Hazard Ratio‡ (95% CI)
P Value
Hazard Ratio§ (95% CI)
P Value
Male, n (%) Age, mean (SD), y BMI (kg/m2) categories, n (%) \21 21–25 $25 All-type dementia, n (%) COPD, n (%) Heart disease, n (%) Hip fracture, n (%) Diabetes, n (%) Gastrostomy, n (%) Nutritional support, n (%)
132 (32.3) 85.2 (8.0)
18 (16.4) 79.5 (8.4)
157 (38.4) 138 (33.7) 114 (27.9) 170 (41.6) 32 (7.8) 51 (12.5) 39 (9.5) 71 (17.4) 28 (6.8) 111 (27.1)
28 (25.5) 39 (35.4) 43 (39.1) 59 (53.5) 5 (4.5) 4 (3.6) 2 (1.8) 3 (2.7) 0 (0) 17 (15.4)
1.78 (1.40–2.25) 1.02 (1.00–1.03) 0.84 (0.73–0.97)k 1.38 (1.04–1.84) 1.01 (0.75–1.36) 1 (reference) 0.82 (0.63–1.07) 2.30 (1.17–4.13) 1.25 (0.84–1.85) 1.73 (1.13–2.64) 1.99 (1.51–2.62) 4.66 (2.82–7.70) —
\.001 .031 .021k .025 .923
1.81 (1.42–2.29) 1.02 (1.00–1.03) 0.80 (0.69–0.93)k 1.53 (1.13–2.06) 1.04 (0.77–1.40) 1 (reference) 0.76 (0.58–1.00) 1.96 (1.03–3.71) 1.22 (0.82–1.80) 1.55 (1.00–2.38) 1.94 (1.47–2.56) 4.33 (2.60–7.22) 0.74 (0.55–0.97)
\.001 .018 .005k .006 .793
.150 .009 .274 .011 .013 \.001 —
.058 .025 .333 .049 .017 \.001 .035
BMI, body mass index; CI, confidence interval; COPD, chronic obstructive pulmonary disease. * Variables included in the final model were those with a univariate P value \.10. † Data are reported as mean (SD) , median (25th–75th), or as counts (%). Percentages are calculated within single groups. ‡ Model based on baseline variables. § Model further adjusted for oral nutritional support during the follow-up. k Linear increase in risk over categories assumed (checked with likelihood ratio test).
lean body mass and functional losses compared with young persons.24 Moreover, long-term care is frequently the destination of patients surviving to hospital stay, which in turn could be associated with nutritional deficits often persisting after discharge.2,4,25 Current guidelines on nutritional management in longterm care recommend the use of trigger BMI of 21 kg/m2. A more sensitive cutoff would result in earlier intervention.1,2 In our study, nutritional support appeared to result in lower death risk. These results are consistent with the recent meta-analysis showing the positive effect of oral supplementation on the survival of malnourished patients.26 Some important aspects should be taken into account for appropriate results interpretation. The limitations are those of observational studies. The effect of potential events occurring during the follow-up has not been considered. Several confounders have been investigated but we did not assess history of weight loss, body composition, and functional status, factors that mostly explain the association between increased mortality and common nutrition-related geriatric syndromes such as frailty, sarcopenia, and cachexia.2,5,6,13 Given also the planned way of recruitment, we cannot exclude that part of the subjects included were those surviving complications leading to institutionalization. Perhaps a systematic inclusion of those newly admitted would allow more accurate estimates of the risk. BMI may be also the result of chronic diseases.2,3,5,6 To avoid a possible overestimation of risk in underweight conditions, analyses were adjusted for the most important disease associated with adverse outcome and nutritional deterioration. With regard to BMI, limitations could also be related to the choice of the reference group. However, it has been recently reported in the European Prospective Investigation into Cancer and Nutrition cohort that the lowest risk of death occurred at a BMI of 25.3 and 24.3 for adult men and women, respectively.19 We recognize that compliance with or administration of BRIEF REPORT
nutritional support has not been monitored and the lack of ability to detect whether or not nutritional support was actually accepted may be a marker for physician selection bias in prescribing habits. Finally, we cannot exclude different practices of global patient management in different institutions. However, our findings are comforted by previous intervention trials in malnourished patients.26 CONCLUSION BMI is significantly associated with all-cause and cardiovascular mortality in institutionalized elderly patients. The present study supports the importance of nutritional assessment and the protective effect of BMI. BMI can be considered a useful screening tool to be performed on a routine basis and a value of 21 kg/m2 or lower appears a valid trigger for nutritional support. REFERENCES 1. Thomas DR, Ashmen W, Morley JE, Evans WJ. Nutritional management in long-term care: Development of a clinical guideline. Council for Nutritional Strategies in Long-Term Care. J Gerontol A Biol Sci Med Sci 2000;55:M725–M734. 2. Arvanitakis M, Beck A, Coppens P, et al. Nutrition in care homes and home care: How to implement adequate strategies (report of the Brussels Forum (22–23 November 2007)). Clin Nutr 2008;27:481–488. 3. Salva A, Coll-Planas L, Bruce S, et al. Nutritional assessment of residents in long-term care facilities (LTCFs): Recommendations of the task force on nutrition and ageing of the IAGG European region and the IANA. J Nutr Health Aging 2009;13:475–483. 4. Cereda E, Pedrolli C. The Geriatric Nutritional Risk Index. Curr Opin Clin Nutr Metab Care 2009;12:1–7. Erratum in: Curr Opin Clin Nutr Metab Care 2009;12:683. 5. Muscaritoli M, Anker SD, Argiles J, et al. Consensus definition of sarcopenia, cachexia and pre-cachexia: Joint document elaborated by Special Interest Groups (SIG) ‘‘cachexia-anorexia in chronic wasting diseases’’ and ‘‘nutrition in geriatrics.’’. Clin Nutr 2010;29:154–159. 6. Cruz-Jentoft AJ, Baeyens JP, Bauer JM, et al. Sarcopenia: European consensus on definition and diagnosis: Report of the European Working Group on Sarcopenia in Older People. Age Ageing 2010;39:412–423. Cereda et al 177
7. Morley JE. Anorexia, sarcopenia, and aging. Nutrition 2001;17:660–663. 8. Nieuwenhuizen WF, Weenen H, Rigby P, Hetherington MM. Older adults and patients in need of nutritional support: Review of current treatment options and factors influencing nutritional intake. Clin Nutr 2010;29:160–169. 9. Kondrup J, Allison SP, Elia M, et al. Educational and Clinical Practice Committee, European Society of Parenteral and Enteral Nutrition (ESPEN). ESPEN guidelines for nutrition screening 2002. Clin Nutr 2003; 22:415–421. 10. World Health Organization. Physical status: The use and interpretation of anthropometry. Report of a WHO Expert Committee. World Health Organ Tech Rep Ser 1995;854:1–452. 11. Heiat A, Vaccarino V, Krumholz HM. An evidence-based assessment of federal guidelines for overweight and obesity as they apply to elderly persons. Arch Intern Med 2001;161:1194–1203. 12. Janssen I. Morbidity and mortality risk associated with an overweight BMI in older men and women. Obesity (Silver Spring) 2007;15: 1827–1840. 13. 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;52:886–891. 14. 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;51:227–230. 15. Miller MD, Thomas JM, Cameron ID, et al. BMI: A simple, rapid and clinically meaningful index of under-nutrition in the oldest old? Br J Nutr 2009;101:1300–1305.
178 Cereda et al
16. Cereda E, Bertoli S, Battezzati A. Height prediction formula for middleaged (30–55 y). Caucasians. Nutrition 2010;26:1075–1081. 17. Stevens J, Cai J, Pamuk ER, et al. The effect of age on the association between body-mass index and mortality. N Engl J Med 1998;338:1–7. 18. Flegal KM, Graubard BI, Williamson DF, Gail MH. Cause-specific excess deaths associated with underweight, overweight, and obesity. JAMA 2007;298:2028–2037. 19. Pischon T, Boeing H, Hoffmann K, et al. General and abdominal adiposity and risk of death in Europe. N Engl J Med 2008;359:2105–2120. 20. James WP, Francois PJ. The choice of cut-off point for distinguishing normal body weights from underweight or ‘chronic energy deficiency’ in adults. Eur J Clin Nutr 1994;48:S179–S184. 21. Tayback M, Kumanyika S, Chee E. Body weight as a risk factor in the elderly. Arch Intern Med 1990;150:1065–1072. 22. Janssen I, Mark AE. Elevated body mass index and mortality risk in the elderly. Obes Rev 2007;8:41–59. 23. Rossman I. Anatomic and body composition changes with aging. In: Finch CE, Hayflick L, editors. Handbook of the Biology of Aging. New York: Reinhold VN; 1977. 24. Schneider SM, Al-Jaouni R, Pivot X, et al. Lack of adaptation to severe malnutrition in elderly patients. Clin Nutr 2002;21:499–504. 25. Sullivan DH, Walls RC, Lipschitz DA. Protein-energy undernutrition and the risk of mortality within 1 y of hospital discharge in a select population of geriatric rehabilitation patients. Am J Clin Nutr 1991;53: 599–605. 26. Milne AC, Potter J, Vivanti A, Avenell A. Protein and energy supplementation in elderly people at risk from malnutrition. Cochrane Database Syst Rev 2009;(2). CD003288.
JAMDA – March 2011
Body Mass Index and Mortality in Institutionalized Elderly Emanuele Cereda, MD, PhD, Carlo Pedrolli, MD, Annunciata Zagami, MD, Alfredo Vanotti, MD, Silvano Piffer, MD, Annalisa Opizzi, RD, Mariangela Rondanelli, MD, PhD, and Riccardo Caccialanza, MD
Objective: Malnutrition and sarcopenia in institutions are very common and significantly affect the prognosis. Aging is characterized by weight and lean body mass losses. Accordingly, in elderly patients, body mass index (BMI) is considered a marker of protein stores rather than of adiposity. Current guidelines suggest a BMI 21 kg/m2 or lower as major trigger for nutritional support. We evaluated the association between BMI and mortality in institutionalized elderly. Methods: This was a multicentric prospective cohort study involving 519 long-term care resident elderly individuals. Risk for mortality across BMI tertiles was estimated by the Cox hazards regression model adjusted for potential confounders recorded at inclusion and collected during the follow-up. Results: During a median follow-up of 5.7 years (25th to 75th percentile, 5.2–8.2], 409 (78.8%) elderly patients died. In primary analyses, based on variables collected at inclusion, patients in the first tertile of
Malnutrition in long-term care facilities is a very common comorbidity that negatively affects the patient’s prognosis.1–4 Several causes appear to contribute to this Nutrition and Dietetics Service, Fondazione IRCCS Policlinico San Matteo, Operativa di Dietetica e Nutrizione Clinica, Pavia, Italy (E.C., R.C.); Unita Ospedale ‘‘S. Chiara,’’ Azienda Provinciale per i Servizi Sanitari, Trento, Italy (C.P.); Fondazione Belluria Onlus, Appiano Gentile, Como, Italy (A.Z.); Servizio di Dietetica e Nutrizione Clinica, ASL Como, Como, Italy (A.V.); Servizio Osservatorio Epidemiologico, Direzione per la Promozione e l’Educazione alla Salute, Azienda Provinciale per i Servizi Sanitari, Trento, Italy (S.P.); Servizio Endocrino-nutrizionale, Dipartimento di Scienze Sanitarie Applicate e Psicocomportamentali, Sezione di Nutrizione, Azienda di Servizi alla Persona degli Studi di Pavia, Pavia, Italy (A.O., M.R.). di Pavia, Universita The study was supported by the Fondazione IRCCS Policlinico San Matteo, Pavia, Italy, and by an Investigator Grant from Nutricia, Milano, Italy (to E.C.). Address correspondence to Emanuele Cereda, MD, PhD, Nutrition and Dietetics Service, Fondazione IRCCS Policlinico San Matteo, Viale Golgi 19, 27100 Pavia, Italy. E-mail: [email protected]
Published by Elsevier Inc. on behalf of the American Medical Directors Association, Inc. DOI:10.1016/j.jamda.2010.11.013 174 Cereda et al
BMI (#21 kg/m2) were at higher risk for all-cause (hazard ratio [HR] 1.38; 95% confidence interval [CI] 1.04– 1.84; P 5 .025) and cardiovascular mortality (HR 5 1.49; 95% CI, 1.00–2.08; P 5 .045). Increased risk was confirmed even after adjusting for nutritional support during the follow-up (all-cause HR 5 1.53; 95% CI, 1.13–2.06; P 5 .006; cardiovascular HR 5 1.62; 95% CI, 1.09–2.40; P 5 .018), which in turn was associated with a reduced risk (all-cause HR 5 0.74; 95% CI, 0.55–0.97; P 5 .035; cardiovascular HR 5 0.62; 95% CI, 0.42–0.91; P 5 .016). Conclusion: BMI is significantly associated with all-cause and cardiovascular mortality in institutionalized elderly patients. A value of 21 kg/m2 or lower can be considered a useful trigger for nutritional support. These results support intending BMI as nutritional reserve in institutionalized elderly patients. (J Am Med Dir Assoc 2011; 12: 174–178) Keywords: Body mass index; long-term care; elderly; mortality; malnutrition
picture, particularly age-related factors, multiple chronic diseases, and insufficient attention for nutritional care.1–3,5,6 Aging is frequently characterized by a reduction of food intake, changes in hormone profile (eg, insulin resistance), and proneness to acute diseases that could result in weight loss and sarcopenia.3,5–8 Current guidelines agree that nutritional screening and nutritional support should be part of routine care.1–3 Several screening tools have been introduced, and the use of those providing a multidimensional assessment of the patient is now recommended.9 However, a patient’s evaluation by these tools may require variable amounts of time and adequate training of health care professionals. Body mass index (BMI), being based only on weight and height measurements, is an unsophisticated, inexpensive, and easy-to-interpret method in nutritional assessment. However, mainly in view of age-related changes in body composition, World Health Organization (WHO) criteria for classifying nutritional status on the basis of BMI appear overly restrictive as they apply to the elderly.10–12 Accordingly, guidelines for JAMDA – March 2011
Outcome
nutritional management in long-term care suggest a BMI of 21 kg/m2 or lower, rather than lower than 18.5 kg/m2, as a majortrigger for nutritional support. However, few large prospective studies have examined the relationship between BMI and mortality in long-term care elderly residents.13–15 We designed the present study to further investigate the potential value of this anthropometric indicator in guiding clinical practice.
Vital status and cause of death were ascertained by active follow-up (Como and Pavia) or by means of record linkage with centralized provincial registry (Trento). Data on death were coded according to the International Classification of Diseases, 10th Revision (ICD-10)17 and classified as cardiovascular (CV), respiratory, neoplasms, or others/not reported.
MATERIALS AND METHODS
Statistical Analyses
Study Design We designed a multicentric prospective cohort study in 4 long-term care facilities in the provinces of Como (n 5 1), Pavia (n 5 1), and Trento (n 5 2). Baseline data collection began in May 2002 and ended in May 2007. Recruitment of patients was performed as follows: every year, for 2 weeks, all the subjects newly admitted were assessed for eligibility. Inclusion criteria were age older than 65 years and agreement to participate by written informed consent (direct enquiry or legal guardians). Exclusion criteria were presence of acute infections or illnesses, established terminal diseases, and known neoplastic disorder. The study was performed in adherence with the principles of the Declaration of Helsinki and protocol was approved by local Institutional Ethics Committee. Nutritional Assessment Baseline nutritional assessment by anthropometry was based on body weight, height, knee-height, mid-arm circumference, and triceps skinfold thickness. Particularly, weight (to the nearest 0.1 kg) was measured using a calibrated flat scale (ambulatory patients) or by a chair scale or a hoistprovided weighting device (nonambulatory or bedridden patients). Height (to the nearest 0.1 cm) was directly assessed in those able to stand or estimated from knee height (nonambulatory patients or in case of abnormal spinal curvature).10,16 BMI was calculated as the ratio between weight (kg) and height (m) squared (kg/m2).10 Patients were categorized into tertiles of this index, setting a cutoff of 21 kg/m2 or lower for tertile I (tertile II, 21–25; tertile III, $25) in agreement with current clinical guidelines for nutritional management in long-term care.1 Eight- to 12-hour fasting venous blood samples were also drawn for the evaluation of hemoglobin, total lymphocyte count, serum albumin, prealbumin, transferrin, total cholesterol, and creatinine. Covariates At baseline, information was collected on age, gender, admission diagnosis, and major comorbidities (diabetes and hypertension). Data were retrieved from medical records or ascertained through patient interview and physical examination. The selection of main admission diagnosis groups used in the analyses was based on literature review by reason of an established association with nutritional status and/or survival. Finally, data on prescription of nutritional support (other than by gastrostomy) during the stay were also retrieved through direct consultation of clinical records. BRIEF REPORT
Statistical analyses were performed using MEDCALC for Windows Version 11.3.0.0 (MedCalc Software, Mariakerke, Belgium) with the test’s significance set to a 2-tailed P value of less than .05. Data were presented as mean and standard deviation, median and interquartile range (25th–75th percentile), or counts and percentage. Comparisons of groups were performed using 1-way analysis of variance, Kruskall-Wallis test analysis (continuous variables), or chi-square test (categorical variables). Associations with all-cause and causespecific mortality were assessed by the Cox proportional hazards regression model. Results were presented as hazard ratio (HR) and 95% confidence interval (CI). Previous to inclusion in the analyses, collinearity between variables was checked with the Pearson statistics. Therefore, only variables with a P less than .10 by univariate analysis were retained in the final model. Primary analysis was based only on baseline features of the population. Then, a further adjustment for nutritional support during the follow-up was also considered as secondary analysis. RESULTS In total, 533 patients were assessed during the recruitment phases. Written consent was obtained for every patient. Fourteen patients were lost to follow-up. The distribution and the features of the analyzed cohort (n 5 519; 97.4% of the original sample) according to tertiles of BMI are presented in Table 1. During a median follow-up of 5.7 years (25th–75th, 5.2– 8.2), 409 deaths (78.8%) occurred. Frequency of events according to BMI were as follows: tertile I, 84.9% (n 5 157); tertile II, 78.0% (n 5 138); tertile III, 72.6% (n 5 114) (P 5 .021). CV causes of death were the most frequent (56.5%; coronary artery disease 24.2%, heart failure 25.1%, stroke 32.0%, others or unspecified 18.7%). Other causes were as follows: 17.4%, respiratory diseases (90% by pneumonia); 6.8%, neoplasm; 19.3%, other diseases. Median survival time was significantly reduced in lower BMI tertiles: tertile I 5 2.0 years (25th–75th, 1.7–2.7); tertile II 5 2.3 years (25th–75th, 1.8–3.4); tertile III 5 3.1 years (25th–75th, 2.3–4.9) (P for trend 5 .003). Then, adjusted HRs for allcause and cause-specific mortality were computed using Cox regression models. The HRs for tertile of BMI were calculated taking the highest one as reference standard. After adjusting for confounders recorded at baseline, participants in the lowest tertile of BMI had a significant elevation in risk for all-cause mortality (HR 5 1.38, 95% CI, 1.04–1.84, P 5 .025; Table 2). Risks for cause-specific mortality were also computed. BMI was significantly associated only with death risk for CV causes (tertile I HR 5 1.49, 95% CI Cereda et al 175
Table 1.
Baseline Clinical and Demographic Characteristics of the Population by Body Mass Index Tertiles
Characteristic*
Overall Population (n 5 519)
1st Tertile BMI #21 (n 5 185)
2nd Tertile BMI 21–25 (n 5 177)
3rd Tertile BMI $25 (n 5 157)
P Value†
Male, n (%) Age, mean (SD), y Body mass index, mean (SD), kg/m2 Arm circumference, mean (SD), cm Triceps skinfold, mean (SD), mm Albumin, mean (SD), g/L Prealbumin, mean (SD), mg/dL Transferrin, mean (SD), mg/dL Hemoglobin, mean (SD), g/dL Total lymphocytes count, mean (SD), /mm3 Total cholesterol, mean (SD), mg/dL Creatinine, mean (SD), mg/dL All-type dementia, n (%) COPD, n (%) Heart disease, n (%) Hip fracture, n (%) Stroke, n (%) Other main admission diagnoses, n (%) Diabetes, n (%) Hypertension, n (%) Gastrostomy, n (%)
150 (28.9) 84 (8.4) 23.1 (2.3) 24.6 (3.6) 12.9 (5.8) 35.4 (4.5) 19.7 (7.4) 201 (41) 12.2 (2.4) 1868 (786) 184 (43) 1.10 (0.53) 229 (44.1) 37 (7.1) 55 (10.6) 41 (7.9) 86 (16.6) 71 (13.7) 74 (14.3) 225 (43.3) 28 (5.4)
55 (29.7) 84.4 (8.5) 18.4 (2.0)‡ 21.7 (3.3)‡ 9.1 (5.4)‡ 34.3 (4.3)‡ 18.2 (6.2) 198 (47) 11.9 (2.4) 1816 (850) 174 (42) 0.99 (0.37) 91 (49.2) 24 (13.0) 21 (11.3) 13 (7.0) 20 (10.8) 16 (8.7) 11 (5.9) 65 (35.3) 16 (8.6)
52 (29.4) 84.2 (8.0) 23.0 (1.1)‡ 24.1 (3.6)‡ 11.6 (5.2)‡ 35.6 (4.6) 20.2 (5.6) 203 (35) 12.1 (2.3) 1833 (624) 187 (42) 1.09 (0.46) 83 (46.9) 10 (5.6) 13 (7.4) 13 (7.4) 36 (20.3) 22 (12.4) 29 (16.7) 85 (47.9) 8 (4.5)
53 (33.7) 83.2 (8.16) 28.6 (3.3)‡ 27.7 (3.9)‡ 17.6 (6.7)‡ 36.3 (4.5) 20.6 (9.5) 201 (39) 12.4 (2.6) 1958 (891) 191 (44)‡ 1.22 (0.69)‡ 55 (35.0) 3 (1.9) 21 (13.4) 15 (9.6) 30 (19.1) 33 (21.0) 34 (21.4) 75 (47.6) 4 (2.5)
.879 .359 \.001 \.001 \.001 \.001 .059 .628 .458 .555 .025 .029 .021 \.001 .186 .651 .068 .002 \.001 .019 .037
BMI, body mass index (kg/m2); COPD, chronic obstructive pulmonary disease; SD, standard deviation. * Data are reported as mean (SD) , median (25th–75th) or as counts (%). Percentages are calculated within single groups. † Continuous and categorical variables were compared between groups with 1-way analysis of variance, the Kruskall Wallis test, or the chisquare test, respectively. ‡ Significantly different versus all the other groups by post-hoc test.
1.00–2.08, P 5 .045; linear increase over tertiles HR 5 0.83, 95% CI 0.68–1.00, P 5 .048). During the follow-up 128 (24.7%) patients were prescribed a nutritional support other than by gastrostomy (93% sip feeding). Frequency of nutritional support was significantly higher in the lowest tertile of BMI: tertile I, 34.6% (n 5 64); tertile II, 19.2% (n 5 34); tertile III, 19.1% (n 5 30). Analyses adjusted for this further confounder confirmed the associations found in primary analyses between BMI and both all-cause (HR 5 1.53, 95% CI 1.13–2.06, P 5 .006) and CV (HR 5 1.62, 95% CI 1.09–2.40, P 5 .018; for linear increase over tertile of BMI, HR 5 0.78, 95% CI 0.64–0.96, P 5 .018) mortality. Nutritional support was also associated with reduced risk (all-cause HR 5 0.74, 95% CI 0.55–0.97, P 5 .035; CV HR 5 0.62, 95% CI 0.42–0.91, P 5 .016). Interestingly, analyses restricted to patients in the lowest tertile of BMI showed that nutritional support was associated with reduced risk of both all-cause (HR 5 0.59, 95% CI 0.39– 0.91, P 5 .018) and cardiovascular mortality (HR 5 0.45, 95% CI 0.26–0.81, P 5 .008). The associations were confirmed even if nutritional status groups were set to exact tertiles of the population (I, #20.7, n 5 175; II, 20.8–24.6, n 5 175; III, $24.7, n 5 169). DISCUSSION The present study demonstrated that BMI is associated with increased mortality from all-cause and cardiovascular disease in long-term care elderly residents. In this respect, we also strengthened the recommendation of current guidelines1,2 that a value of 21 kg/m2 or lower should be considered 176 Cereda et al
a major trigger for nutritional support. Moreover, this was the first prospective study reporting a reduction in death risk in patients receiving nutritional support. In the general adult population the association between BMI and mortality has been described as a J-shaped curve with higher increase in risk for smaller variation of BMI at low body weight rather than in overweight and obesity.17–19 These studies provided compelling evidence that World Health Organization’s BMI thresholds (mainly based on morbidity risk) are significantly lower than those associated with reduced death risk. A value lower than 20 kg/m2 is now accepted as underweight in developed countries.20 Ad hoc studies or subgroup analyses have reported a U-shaped relationship in elderly persons, with a less steep curve on the right; a wide, flat bottom; and a minimum risk shifted to higher BMI values.11,12,17–19,21,22 Different threshold values for the elderly should account for a greater variability in functional and metabolically active components of the body in advanced age, which have been shown to explain most of the outcome prediction by BMI.13,14,22 Moreover, aging is associated with height reduction owing to accentuation of spinal curvature and narrowing of intervertebral disk spaces.23 BMI shows limitations in reflecting body composition and it is likely to reflect lean body mass stores in elderly persons. Aging is characterized by a reduction in skeletal muscle mass because of multiple contributing and interacting factors, such as reduced physical activity, presence of chronic diseases, endocrine changes, increased inflammatory background, and primary and/or secondary anorexia.5–8 In the presence of an impaired energy balance, older people are prone to higher JAMDA – March 2011
Table 2.
Predictors*of All-Cause Mortality at Multivariable Cox Proportional Hazard Regression Models
Characteristic†
Dead (n 5 409)
Survivors (n 5 110)
Hazard Ratio‡ (95% CI)
P Value
Hazard Ratio§ (95% CI)
P Value
Male, n (%) Age, mean (SD), y BMI (kg/m2) categories, n (%) \21 21–25 $25 All-type dementia, n (%) COPD, n (%) Heart disease, n (%) Hip fracture, n (%) Diabetes, n (%) Gastrostomy, n (%) Nutritional support, n (%)
132 (32.3) 85.2 (8.0)
18 (16.4) 79.5 (8.4)
157 (38.4) 138 (33.7) 114 (27.9) 170 (41.6) 32 (7.8) 51 (12.5) 39 (9.5) 71 (17.4) 28 (6.8) 111 (27.1)
28 (25.5) 39 (35.4) 43 (39.1) 59 (53.5) 5 (4.5) 4 (3.6) 2 (1.8) 3 (2.7) 0 (0) 17 (15.4)
1.78 (1.40–2.25) 1.02 (1.00–1.03) 0.84 (0.73–0.97)k 1.38 (1.04–1.84) 1.01 (0.75–1.36) 1 (reference) 0.82 (0.63–1.07) 2.30 (1.17–4.13) 1.25 (0.84–1.85) 1.73 (1.13–2.64) 1.99 (1.51–2.62) 4.66 (2.82–7.70) —
\.001 .031 .021k .025 .923
1.81 (1.42–2.29) 1.02 (1.00–1.03) 0.80 (0.69–0.93)k 1.53 (1.13–2.06) 1.04 (0.77–1.40) 1 (reference) 0.76 (0.58–1.00) 1.96 (1.03–3.71) 1.22 (0.82–1.80) 1.55 (1.00–2.38) 1.94 (1.47–2.56) 4.33 (2.60–7.22) 0.74 (0.55–0.97)
\.001 .018 .005k .006 .793
.150 .009 .274 .011 .013 \.001 —
.058 .025 .333 .049 .017 \.001 .035
BMI, body mass index; CI, confidence interval; COPD, chronic obstructive pulmonary disease. * Variables included in the final model were those with a univariate P value \.10. † Data are reported as mean (SD) , median (25th–75th), or as counts (%). Percentages are calculated within single groups. ‡ Model based on baseline variables. § Model further adjusted for oral nutritional support during the follow-up. k Linear increase in risk over categories assumed (checked with likelihood ratio test).
lean body mass and functional losses compared with young persons.24 Moreover, long-term care is frequently the destination of patients surviving to hospital stay, which in turn could be associated with nutritional deficits often persisting after discharge.2,4,25 Current guidelines on nutritional management in longterm care recommend the use of trigger BMI of 21 kg/m2. A more sensitive cutoff would result in earlier intervention.1,2 In our study, nutritional support appeared to result in lower death risk. These results are consistent with the recent meta-analysis showing the positive effect of oral supplementation on the survival of malnourished patients.26 Some important aspects should be taken into account for appropriate results interpretation. The limitations are those of observational studies. The effect of potential events occurring during the follow-up has not been considered. Several confounders have been investigated but we did not assess history of weight loss, body composition, and functional status, factors that mostly explain the association between increased mortality and common nutrition-related geriatric syndromes such as frailty, sarcopenia, and cachexia.2,5,6,13 Given also the planned way of recruitment, we cannot exclude that part of the subjects included were those surviving complications leading to institutionalization. Perhaps a systematic inclusion of those newly admitted would allow more accurate estimates of the risk. BMI may be also the result of chronic diseases.2,3,5,6 To avoid a possible overestimation of risk in underweight conditions, analyses were adjusted for the most important disease associated with adverse outcome and nutritional deterioration. With regard to BMI, limitations could also be related to the choice of the reference group. However, it has been recently reported in the European Prospective Investigation into Cancer and Nutrition cohort that the lowest risk of death occurred at a BMI of 25.3 and 24.3 for adult men and women, respectively.19 We recognize that compliance with or administration of BRIEF REPORT
nutritional support has not been monitored and the lack of ability to detect whether or not nutritional support was actually accepted may be a marker for physician selection bias in prescribing habits. Finally, we cannot exclude different practices of global patient management in different institutions. However, our findings are comforted by previous intervention trials in malnourished patients.26 CONCLUSION BMI is significantly associated with all-cause and cardiovascular mortality in institutionalized elderly patients. The present study supports the importance of nutritional assessment and the protective effect of BMI. BMI can be considered a useful screening tool to be performed on a routine basis and a value of 21 kg/m2 or lower appears a valid trigger for nutritional support. REFERENCES 1. Thomas DR, Ashmen W, Morley JE, Evans WJ. Nutritional management in long-term care: Development of a clinical guideline. Council for Nutritional Strategies in Long-Term Care. J Gerontol A Biol Sci Med Sci 2000;55:M725–M734. 2. Arvanitakis M, Beck A, Coppens P, et al. Nutrition in care homes and home care: How to implement adequate strategies (report of the Brussels Forum (22–23 November 2007)). Clin Nutr 2008;27:481–488. 3. Salva A, Coll-Planas L, Bruce S, et al. Nutritional assessment of residents in long-term care facilities (LTCFs): Recommendations of the task force on nutrition and ageing of the IAGG European region and the IANA. J Nutr Health Aging 2009;13:475–483. 4. Cereda E, Pedrolli C. The Geriatric Nutritional Risk Index. Curr Opin Clin Nutr Metab Care 2009;12:1–7. Erratum in: Curr Opin Clin Nutr Metab Care 2009;12:683. 5. Muscaritoli M, Anker SD, Argiles J, et al. Consensus definition of sarcopenia, cachexia and pre-cachexia: Joint document elaborated by Special Interest Groups (SIG) ‘‘cachexia-anorexia in chronic wasting diseases’’ and ‘‘nutrition in geriatrics.’’. Clin Nutr 2010;29:154–159. 6. Cruz-Jentoft AJ, Baeyens JP, Bauer JM, et al. Sarcopenia: European consensus on definition and diagnosis: Report of the European Working Group on Sarcopenia in Older People. Age Ageing 2010;39:412–423. Cereda et al 177
7. Morley JE. Anorexia, sarcopenia, and aging. Nutrition 2001;17:660–663. 8. Nieuwenhuizen WF, Weenen H, Rigby P, Hetherington MM. Older adults and patients in need of nutritional support: Review of current treatment options and factors influencing nutritional intake. Clin Nutr 2010;29:160–169. 9. Kondrup J, Allison SP, Elia M, et al. Educational and Clinical Practice Committee, European Society of Parenteral and Enteral Nutrition (ESPEN). ESPEN guidelines for nutrition screening 2002. Clin Nutr 2003; 22:415–421. 10. World Health Organization. Physical status: The use and interpretation of anthropometry. Report of a WHO Expert Committee. World Health Organ Tech Rep Ser 1995;854:1–452. 11. Heiat A, Vaccarino V, Krumholz HM. An evidence-based assessment of federal guidelines for overweight and obesity as they apply to elderly persons. Arch Intern Med 2001;161:1194–1203. 12. Janssen I. Morbidity and mortality risk associated with an overweight BMI in older men and women. Obesity (Silver Spring) 2007;15: 1827–1840. 13. 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;52:886–891. 14. 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;51:227–230. 15. Miller MD, Thomas JM, Cameron ID, et al. BMI: A simple, rapid and clinically meaningful index of under-nutrition in the oldest old? Br J Nutr 2009;101:1300–1305.
178 Cereda et al
16. Cereda E, Bertoli S, Battezzati A. Height prediction formula for middleaged (30–55 y). Caucasians. Nutrition 2010;26:1075–1081. 17. Stevens J, Cai J, Pamuk ER, et al. The effect of age on the association between body-mass index and mortality. N Engl J Med 1998;338:1–7. 18. Flegal KM, Graubard BI, Williamson DF, Gail MH. Cause-specific excess deaths associated with underweight, overweight, and obesity. JAMA 2007;298:2028–2037. 19. Pischon T, Boeing H, Hoffmann K, et al. General and abdominal adiposity and risk of death in Europe. N Engl J Med 2008;359:2105–2120. 20. James WP, Francois PJ. The choice of cut-off point for distinguishing normal body weights from underweight or ‘chronic energy deficiency’ in adults. Eur J Clin Nutr 1994;48:S179–S184. 21. Tayback M, Kumanyika S, Chee E. Body weight as a risk factor in the elderly. Arch Intern Med 1990;150:1065–1072. 22. Janssen I, Mark AE. Elevated body mass index and mortality risk in the elderly. Obes Rev 2007;8:41–59. 23. Rossman I. Anatomic and body composition changes with aging. In: Finch CE, Hayflick L, editors. Handbook of the Biology of Aging. New York: Reinhold VN; 1977. 24. Schneider SM, Al-Jaouni R, Pivot X, et al. Lack of adaptation to severe malnutrition in elderly patients. Clin Nutr 2002;21:499–504. 25. Sullivan DH, Walls RC, Lipschitz DA. Protein-energy undernutrition and the risk of mortality within 1 y of hospital discharge in a select population of geriatric rehabilitation patients. Am J Clin Nutr 1991;53: 599–605. 26. Milne AC, Potter J, Vivanti A, Avenell A. Protein and energy supplementation in elderly people at risk from malnutrition. Cochrane Database Syst Rev 2009;(2). CD003288.
JAMDA – March 2011