567902

research-article2015

PENXXX10.1177/0148607114567902Journal of Parenteral and Enteral NutritionAllard et al

Original Communication

Malnutrition at Hospital Admission—Contributors and Effect on Length of Stay: A Prospective Cohort Study From the Canadian Malnutrition Task Force

Journal of Parenteral and Enteral Nutrition Volume XX Number X Month 201X 1­–11 © 2015 American Society for Parenteral and Enteral Nutrition DOI: 10.1177/0148607114567902 jpen.sagepub.com hosted at online.sagepub.com

Johane P. Allard, MD1; Heather Keller, RD, PhD2; Khursheed N. Jeejeebhoy, PhD, MD3; Manon Laporte, RD, MSc4; Don R. Duerksen, MD5; Leah Gramlich, MD6; Helene Payette, PhD7; Paule Bernier, RD8; Elisabeth Vesnaver, PhD9; Bridget Davidson, RD, MHSc10; Anastasia Teterina, PhD11; and Wendy Lou, PhD12

Abstract Background: In hospitals, length of stay (LOS) is a priority but it may be prolonged by malnutrition. This study seeks to determine the contributors to malnutrition at admission and evaluate its effect on LOS. Materials and Methods: This is a prospective cohort study conducted in 18 Canadian hospitals from July 2010 to February 2013 in patients ≥ 18 years admitted for ≥ 2 days. Excluded were those admitted directly to the intensive care unit; obstetric, psychiatry, or palliative wards; or medical day units. At admission, the main nutrition evaluation was subjective global assessment (SGA). Body mass index (BMI) and handgrip strength (HGS) were also performed to assess other aspects of nutrition. Additional information was collected from patients and charts review during hospitalization. Results: One thousand fifteen patients were enrolled: based on SGA, 45% (95% confidence interval [CI], 42%–48%) were malnourished, and based on BMI, 32% (95% CI, 29%–35%) were obese. Independent contributors to malnutrition at admission were Charlson comorbidity index > 2, having 3 diagnostic categories, relying on adult children for grocery shopping, and living alone. The median (range) LOS was 6 (1–117) days. After controlling for demographic, socioeconomic, and disease-related factors and treatment, malnutrition at admission was independently associated with prolonged LOS (hazard ratio, 0.73; 95% CI, 0.62–0.86). Other nutrition-related factors associated with prolonged LOS were lower HGS at admission, receiving nutrition support, and food intake < 50%. Obesity was not a predictor. Conclusion: Malnutrition at admission is prevalent and associated with prolonged LOS. Complex disease and age-related social factors are contributors. (JPEN J Parenter Enteral Nutr. XXXX;xx:xx-xx)

Keywords hospital malnutrition; malnutrition contributors; obesity; nutrition status; length of stay

Clinical Relevancy Statement The results from this study underline the importance of early screening to identify patients at nutrition risk and to provide prompt nutrition care. Considering the preadmission contributors, attention should also be paid to extend nutrition care to the community after patients’ discharge.

Introduction Malnutrition and obesity are prevalent in hospital settings. The prevalence of hospital malnutrition is between 15% and 70%,1-8 and the prevalence of obesity is about 30%,9,10 depending on populations, types of institutions, and methods of assessment. For hospital malnutrition, contributors may relate to underlying illnesses, aging, socioeconomic situations,11,12 surgical procedures, lack of nutrition monitoring, and intervention.8,13 For malnutrition detected at admission, preadmission contributors are not well studied.

Both malnutrition and obesity are associated with detrimental outcomes. For obesity, prolonged length of stay (LOS), complications, readmissions, and mortality are often due to obesity-associated comorbidities,9,14-16 resulting in increased costs.17-20 A high body mass index (BMI) may also be protective,21 but it is important to consider that obese patients may be malnourished and have sarcopenia that is difficult to detect if not formally assessed. The presence of sarcopenia regardless of body weight may contribute to detrimental outcomes.22,23 This is supported by a recent study24 that showed that malnutrition risk is a frequent finding in newly hospitalized overweight/obese adults, increasing LOS and risk of inhospital mortality. Therefore, a more comprehensive nutrition assessment such as subjective global assessment (SGA) and handgrip strength (HGS) may be required to detect malnutrition, including muscle wasting and weakness. Hospital malnutrition is also associated with detrimental outcomes25-33 and increased healthcare costs,28,32 independent of underlying disease, presence of comorbidities, patient’s age,

Downloaded from pen.sagepub.com at UCSF LIBRARY & CKM on April 4, 2015

2

Journal of Parenteral and Enteral Nutrition XX(X)

and/or socioeconomic factors. However, how the presence of malnutrition at the time of admission and the presence of various components of hospital nutrition care, such as dietitian visits, food intake, and nutrition support, affect LOS remains unclear. The first aim of this study was to assess the nutrition status of patients at hospital admission using SGA, BMI, and HGS and determine the prevalence of malnutrition and obesity as well as preadmission contributors to malnutrition. The second aim was to determine if the nutrition status of the patients assessed at admission by these methods was a predictor of prolonged LOS.

Materials and Methods Study Design This is a prospective multicenter cohort study conducted from July 2010 to February 2013, in adult patients admitted directly to the surgical and medical wards of 18 participating acute care hospitals from 8 provinces across Canada. Hospitals from across Canada were made aware of the study by various modes of communication: presentations at national conferences and hospitals, direct contact with hospital dietitians and administrators, and creation of the website www.nutritioncareincanada.ca. Those expressing an interest and agreeing to participate submitted the protocol and consent form to their respective research ethics boards (REBs). These centers were classified as academic, community, and small (200 beds or fewer) or large (more than 200 beds). Small centers enrolled a total of 40 patients per facility, and large centers enrolled 60 patients each. Patients were enrolled according to a strict protocol to avoid selection bias. Days of

enrollment rotated from Monday to Friday, with Monday capturing the week-end admissions from Friday at 5 pm to Monday at 5 pm. Consecutive admissions were approached for consent and a maximum of 7 patients were followed at the same time.

Eligibility Criteria Patients were included if they were > 18 years old and consent was provided. Patients were excluded if admitted directly to the intensive care unit; obstetric, psychiatry, or palliative wards; or a medical day unit. In addition, age and sex of the total number of hospital admissions over the study period were collected where possible to evaluate the representativeness of the study sample. The study was approved by all institutions’ administration and REBs.

Data Collection Demography; contact information; living arrangements; foodrelated activities of daily living; level of education; primary, secondary, and new diagnoses; presence/absence of cancer; Charlson comorbidity index (CCI)34; and number of medications were obtained. Due to the variety of diagnoses, these were classified under 11 broad standard categories (Table 1). If there was more than 1 category for the same patient, a second and third diagnostic category was coded. Because of the high number of categories, the number of diagnostic categories (1, 2, or 3) was used for analysis, in addition to the presence/ absence of cancer, CCI, admission wards (medical vs surgical), and number of medications to account for type and severity of illnesses.

From 1Department of Medicine, University Health Network, University of Toronto, Ontario, Canada; 2Schlegel-UW Research Institute for Aging, Applied Health Sciences, University of Waterloo, Waterloo, Ontario, Canada; 3Department of Medicine, St Michael’s Hospital, University of Toronto, Toronto, Ontario, Canada; 4Clinical Nutrition Department, Réseau de Santé Vitalité Health Network, Campbellton Regional Hospital, New Brunswick, Canada; 5Department of Medicine, St Boniface Hospital, University of Manitoba, Winnipeg, Manitoba, Canada; 6Department of Medicine, University of Alberta, Alberta Health Services, Edmonton, Alberta, Canada; 7Facultée de la Médecine et des Sciences de la Santé, Université de Sherbrooke, Sherbrooke, Québec, Canada; 8Jewish General Hospital, McGill University, Montréal, Québec, Canada; 9Department of Family Relations and Applied Nutrition, University of Guelph, Guelph, Ontario, Canada; 10Canadian Nutrition Society, Toronto, Ontario, Canada; 11University Health Network, Toronto, Ontario, Canada; and 12Dalla Lana School of Public Health, University of Toronto, Toronto, Ontario, Canada. Supplementary material for this article is available on the Journal of Parenteral and Enteral Nutrition website at http://jpen.sagepub.com/supplemental. Financial disclosure: Funding by the Canadian Nutrition Society after receiving unrestricted grants from Abbott Nutrition, Baxter, Fresenius-Kabi Canada, Nestle Healthcare Nutrition, and Pfizer. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The researchers were totally independent. Conflict of interest disclosure: All authors have completed the ICMJE uniform disclosure form at www.icmje.org/coi_disclosure.pdf and declare financial support from the Canadian Nutrition Society, which funded the study with unrestricted grants from Abbott Nutrition, Baxter, Fresenius-Kabi Canada, Nestle Healthcare Nutrition, and Pfizer. J.P.A., H.K., K.N.J., M.L., D.R.D., L.G., H.P., and P.B. are also members of the Abbott Nutrition speaker bureau. There exist no other relationships or activities that could appear to have influenced the submitted work. Disclaimer: The authors had full access to and take full responsibility for the integrity of the data. The views expressed in this article are those of the authors and do not necessarily represent the views of the Canadian Nutrition Society. Received for publication October 18, 2014; accepted for publication November 14, 2014. Corresponding Author: Dr Johane P. Allard, MD, Department of Medicine, Division of Gastroenterology, University Health Network, Toronto General Hospital, 585 University Avenue, 9N-973, Toronto, ON M5G 2C4, Canada. Email: [email protected]

Downloaded from pen.sagepub.com at UCSF LIBRARY & CKM on April 4, 2015

Allard et al

3

Table 1.  Frequencies of Diagnostic Categories in Study Sample and Canadian Patient Population Based on Canadian Institute for Health Information 2009–2011. Primary Diagnostic Category

CMTF Patientsa (n = 1015), No. (%)

CIHIb (n = 78,497), %

159 (15.7) 309 (30.4) 127 (12.5) 189 (18.6) 105 (10.3) 56 (5.5) 7 (0.7) 79 (7.8) 10 (1.0) 21 (2.1) 84 (8.3) 188 (18.5) 143 (14.1) 200 (19.8)

22.19 16.4 7.74 10.9 13.91 2.5 NA 3.63 0.35 NA 1.21 3.25 17.91 11.4

Cardiovascular Gastrointestinal Genitourinary Respiratory Musculoskeletal Neurologic Autoimmune Metabolic Sensory organ Trauma Hematopoietic Infection Other Presence of cancer

CIHI, Canadian Institute for Health Information; CMTF, Canadian Malnutrition Task Force; NA, not available. a Percentage of those having this diagnostic category out of total number of patients in the cohort; some patients may have more than 1 diagnostic category. b Parts of this material are based on the CIHI Discharge Abstract Database Research Analytic Files (sampled from fiscal years 2009–2010 and 2010– 2011). However, the analysis, conclusions, opinions, and statements expressed herein are those of the authors and not those of the CIHI.

At admission, SGA was performed by trained coordinators and was the primary measure of nutrition status,35 where malnutrition was defined as SGA B or C (SGA B+C). In addition, BMI (kg/m2) was calculated from weight (measured with Seca 952 chair scale; Weigh and Measure, LLC, Olney, MD, USA) and height (standing height [cm], or if > 65 years or unable to stand, estimated from knee height [SHORR knee height caliper; Weigh and Measure, LLC]) and used in further analysis as categories according to recognized cutoff points36,37 (< 20, ≥ 20 and < 30, and ≥ 30 kg/m2) to indicate undernourished, well-nourished, and obese groups. The HGS was measured using a hydraulic hand dynamometer (Jamar hydraulic hand dynamometer; Weigh and Measure, LLC) according to standardized procedure.38 During hospital stay, dietary intake was estimated by a nutritionDay form38,39 at lunch time for 3 days in the first week of stay and recorded as 0%, 25%, 50%, or 100% consumption of the main meal. In addition, numbers of daily medications for the first 10 days of admission, diet orders, dietitian visits, and type and timing of nutrition support were recorded.

Clinical Outcomes and Covariates The main outcome was LOS defined as the difference (days) between the date of discharge, death, or transfer to another hospital and the date of admission to the hospital ward. Covariates selected to assess the preadmission contributors to malnutrition (SGA B+C) at admission included demographic (age and sex), socioeconomic (living arrangements, education level, primary person responsible for grocery shopping and cooking), and disease-related (number of diagnoses, presence of cancer, CCI) characteristics.

For multivariate analysis of the outcome LOS, SGA was included in the model as a binary covariate: SGA A (well-nourished) vs SGA B+C (malnourished). In addition, other nutritionrelated parameters were recorded at admission and during hospitalization. The preadmission weight loss ≥ 5% variable was calculated using the ratio of usual weight, self-reported by the patients at admission, to the weight measured at admission to the ward. It was categorized using the 5% cutoff point to distinguish patients with a substantial amount of weight loss, indicating impaired nutrition status. The admission BMI variable was categorized as previously described (BMI < 20 kg/m2; 20 to < 30; ≥ 30 kg/m2). The admission HGS was included in the models as a linear predictor. During hospitalization, the variables used were the following: (1) average food intake estimates from nutritionDay over the first week, dichotomized at 50% (< 50% vs ≥ 50%); (2) registered dietitian visit; (3) nil per os (NPO; nothing by mouth) for ≥ 3 days (without any additional oral nutrition, enteral nutrition [EN], or parenteral nutrition [PN]), as this was reported in the literature as associated with prolonged hospital stay25; and (4) presence of nutrition support, which included oral supplementation, EN, or PN. The latter 3 variables were treated as time-dependent indicators, taking value 1 starting from the day of the first occurrence of the event. Other covariates to control for in the model were selected based on literature and clinical experience and included age, sex, education level, living arrangements (living at home vs living in supportive housing, retirement home/assisted living facilities, nursing home, and 4 patients with no permanent home), number of diagnostic categories (defined as 1, 2, or 3 diagnostic categories based on admission and new diagnoses made during hospitalization, and treated as categorical predictor in the analysis), presence of cancer

Downloaded from pen.sagepub.com at UCSF LIBRARY & CKM on April 4, 2015

4

Journal of Parenteral and Enteral Nutrition XX(X)

(recorded at admission or diagnosed in hospital), and CCI at admission. In addition, treatment-related factors were included: admission ward (surgical vs medical) and number of medications per day defined based on longitudinal follow-up of the number of medications per day during the first 10 days of hospitalization (or until discharge if LOS was fewer than 10 days) as median number of daily medications during that period.

Sample Size The sample size was first calculated to estimate the prevalence of malnutrition. Using a simple random sample survey model, with a confidence level of 95%, a margin of error of 5%, and a prevalence of malnutrition of 31.6%,33 it was estimated at 333 participants. Since our study was based on a cluster sample, this estimate was multiplied by the effect of the plan of sampling (2 for nutrition surveys). The total sample size reached 666 participants. We then enrolled more than 1000 patients to assess contributors of malnutrition and predictors of prolonged LOS, based on previous similar studies,25,33 considering the potential for missing measurements.

Statistical Analysis Patient characteristics and clinical outcomes are presented by counts with percentages for categorical variables, and with medians including first and third quartiles (q1, q3) for continuous variables. The associations between these variables and the nutrition status SGA (A, B, and C) were examined using chisquare or Fisher exact tests for categorical variables and Kruskal-Wallis test for continuous variables. For associations found to be significant, pairwise comparisons were performed for all SGA categories, and the Bonferroni correction was employed for adjusting multiple comparisons. Based on the results of bivariable analyses and clinical input, SGA categories B and C were combined for multivariable analyses. To examine the associations between malnutrition (SGA B+C) and preadmission and disease-related factors while accounting for other covariates (including age, education, etc), multivariable logistic regression was used. To take into account the correlations between patients admitted to the same hospital, the generalized estimating equations approach was employed. The final multivariate model was developed based on bivariable analyses, literature reviews, model selection procedures, and clinical input. Odds ratios (ORs) and 95% confidence intervals (CIs) were derived for each of the parameters. Model assumptions and model fits were examined using graphical and model assessment techniques.40 The statistical analysis named “frailty Cox PH (proportional hazards) model” was used for predicting LOS; this specific type of statistical approach accounted for differences between hospitals. Univariate and final multivariate models were built with malnutrition as the primary variable of interest. Patients who died during hospitalization or were transferred to

other hospitals were treated as censored observations. Variable selection for the final PH frailty model was established through literature reviews, Kaplan-Meier plots, log-rank tests, clinical considerations, and variable selection techniques; for model assessments, Schoenfeld residuals and cumulative-hazard and deviance-residual plots41 were examined. For all parameters, hazard ratios (HRs) and 95% CIs were obtained. Two sensitivity analyses were carried out to examine the effects of inclusion/exclusion of decreased patients and patients who stayed for 2 days or fewer post-entry. All statistical tests were 2-sided, and the significance level was set at 5%. All analyses were performed using SAS v9.3.

Role of the Funding Source This study was funded by the Canadian Nutrition Society. The funding source did not influence the design, conduct, or analysis of the study or the decision to submit the manuscript for publication.

Results Patient Characteristics and Nutrition Parameters A total of 1015 patients were enrolled. The distribution of patients by primary diagnostic category at the time of admission was compared with the distribution of diagnostic category in the Canadian patient population based on the Canadian Institute for Health Information (CIHI) database 2009-2011 (Table 1). In our study, the most frequent primary diagnostic category was gastrointestinal disorder, followed by respiratory disease, infection, and cardiovascular disease. Cancer was present in 19.8% of patients at admission. The CIHI database reported cardiovascular, “other,” gastrointestinal, and musculoskeletal diseases as the most frequent diagnoses with presence of cancer at admission at 11.4%. Some of the differences may be attributed to the study entry criteria. In 9 hospitals where the data were available, enrollment for age and sex was overall representative when compared with all admissions. There was no significant difference in age except in 1 hospital, where third party consent was not approved by REB, and where patients age 65 and older were underrepresented (P = .002). For sex, 8 hospitals had no differences and 1 had a higher proportion of males (P = .005). Characteristics by type of hospital and province are in Supplementary Table 1. There was no statistically significant difference between academic and community hospitals, except for age (corresponding medians 64 and 68 years), LOS (7 days vs 6 days), and CCI (2 vs 1.5). Patient characteristics are presented for all the patients and by SGA category in Table 2. The median age (first and third quartiles [q1, q2]) was 66 years (54 years, 77 years) and 527

Downloaded from pen.sagepub.com at UCSF LIBRARY & CKM on April 4, 2015

Allard et al

5

Table 2.  Patient Characteristics According to Nutrition Status (Subjective Global Assessment Categories) at Baseline.a Parameter

No.b All Patients (N = 1015)c SGA A (n = 558)c SGA B (n = 341)c SGA C (n = 116)c P Valued

Demography and socioeconomic parameters 1014  Age 1014  Sex  Male    Ethnicity  Canadian    European    Othere   Living arrangements prior to 1015 admission   Alone at home     With others at home     Other (not at home)f   Having postsecondary education 1007   Primary person responsible for 996 grocery shopping   Patient or spouse     Adult child    Otherg   Primary person responsible for 993 cooking   Patient or spouse     Adult child    Otherg   Preadmission parameters   Use of oral nutrition supplements 1011 prior to hospitalization   Use of vitamin/mineral 726 supplements prior to hospitalization   No. of surgeries in the past 5 1000 years  None     1 or 2     3 or more  

66 (54, 77)

64ab (53, 75)

68a (54, 79)

68b (58, 79)

< .001 .49     .083             .056

527 (52.0)

292 (52.4)

170 (49.9)

65 (56.0)

835 (81.8) 117 (11.5) 69 (6.8)

463 (55.7) 52 (45.2) 42 (61.8)

279 (33.6) 46 (40.0) 16 (23.5)

89 (10.7) 17 (14.8) 10 (14.7)

272 (26.8) 651 (64.1) 92 (9.1) 405 (40.2)

136 (24.4) 378 (67.7) 44 (7.9) 242 (43.6)

96 (28.2) 207 (60.7) 38 (11.1) 125 (37.2)

40 (34.5) 66 (56.9) 10 (8.6) 38 (32.8)

            .037   .006

838 (84.1) 63 (6.3) 95 (9.5)

480 (87.4) 22 (4.0) 47 (8.6)

266 (79.9) 29 (8.7) 38 (11.4)

92 (80.7) 12 (10.5) 10 (8.8)

            .098

867 (87.3) 33 (3.3) 93 (9.4)

491 (89.4) 12 (2.2) 46 (8.4)

282 (84.9) 14 (4.2) 36 (10.8)

94 (83.9) 7 (6.3) 11 (9.8)

           

215 (21.3) 415 (56.2)

65 (11.7) 225 (58.4)

95 (27.9) 141 (56.6)

55 (47.4) 49 (53.3)

< .001   .65   .99

498 (49.8) 415 (41.5) 87 (8.7)

275 (50.0) 227 (41.3) 48 (8.7)

164 (49.0) 141 (42.1) 30 (9.0)

59 (51.3) 47 (40.9) 9 (7.8)

            (continued)

Downloaded from pen.sagepub.com at UCSF LIBRARY & CKM on April 4, 2015

6

Journal of Parenteral and Enteral Nutrition XX(X)

Table 2.  (continued) Parameter

No.b All Patients (N = 1015)c SGA A (n = 558)c SGA B (n = 341)c SGA C (n = 116)c P Valued

  2 or more acute care admissions 984 545 268 in the past 5 years (55.4) (49.5) Admission ward and disease-related characteristics   Admission ward: surgical 955 295 177   (30.9) (33.7)   Number of diagnostic categories 1015   1 category 653 385   (64.3) (69.0)   2 categories 262 133   (25.8) (23.8)   3 categories 100 40   (9.9) (7.2)   CCI at admission > 2 1011 418 197   (41.8) (36.0)   Cancer present 1011 200 93   (19.8) (16.8)   Infection is present in 1 of 1015 188 94 diagnosis categories (18.5) (16.9) 1012 10 10   Number of medications prescribed at admission (6, 15) (6, 15) Additional nutrition-related parameters on admission and length of stay 966 26.9 29.0ab   BMI, m/kg2   (23.2, 31.8) (24.9, 33.9)   Handgrip strength, kg 993 20.5 22.7ab   (14, 30) (15.6, 32)   Length of stay, d 922 6 (4, 11) 6.0ab   (1, 81)

199 (59.9)

78 (70.3)

< .001  

95 (29.4)

23 (21.5)

193 (56.6) 99 (29.0) 49 (14.4) 160 (46.9) 75 (22.1) 77 (22.6) 10 (6, 14)

75 (64.7) 30 (25.9) 11 (9.5) 61 (54.5) 32 (27.6) 17 (14.7) 10.5 (7,15)

.035   .001             < .001   .013   .052   .81  

25.3ac (22.7, 29.8) 18.2ac (13, 26) 7.0ac (2, 117.0)

21.0bc (17.5, 23.6) 15.88bc (9, 22.3) 9.0bc (2, 46)

< .001   < .001   < .001  

BMI, body mass index; CCI, Charlson comorbidity index; SGA A, well-nourished; SGA B, moderately malnourished; SGA C, severely malnourished. a Continuous parameters presented as median (q1, q3); categorical parameters represented as No. (%). b Total number of nonmissing observations for the variable. c Number of patients for some variables may be lower due to missing values. d Kruskal-Wallis test, Pearson chi-square test, or Fisher exact test as appropriate; the same letter subscripts indicate significant differences between respective categories. e “Other” includes West, South, East, and Southeast Asian, African, Pacific Islander, Central and South American, Caribbean, Aboriginal/Native, Arab, and other (maximum number of patients in any of these categories was 20 [2%]). f “Other” includes supportive housing, retirement home/assisted living, nursing home, and 8 patients with no permanent home. g “Other” includes other family (not children or spouse), neighbors, friends, and community support service.

(51.9%) were male. The median LOS was (6 days, 11 days) with a range of 1–117 days; 77 patients had LOS ≤ 2 days. For the 1015 patients, 558 (55%) were well nourished (SGA A), 341 (33.6%) were moderately malnourished (SGA B), and 116 (11.4%) were severely malnourished (SGA C). Thus, the overall proportion of malnourished patients (SGA B+C) at admission was 45% (95% CI, 42%–48%). Admission BMI was measured in 966 patients: the median (q1, q3) was 26.9 (23.3, 31.8). Only 87 (9%) had a BMI below 20; 294 (30.3%) were between 20 and ≤ 25; 272 (28.2%) were above 25 and below 30; and 313 (32.4%) were ≥ 30 kg/m2 (obese). Admission HGS was measured in 993 patients. When compared with reference values,42 661 patients (66.6%) had lower than normal HGS, defined as smaller than the lower limit of 95% CI of HGS for healthy subjects of the same age and sex, and 332 (33.43%)

were within normal range. BMI and overall HGS were significantly associated with SGA categories. During the first week of hospitalization, 931 patients had food intake assessed: 31.4% ate < 50% of their meals. During hospital stay, 106 patients (10.5%) were kept NPO ≥ 3 days, 71 patients (7.01%) received some form of nutrition support (oral, EN, or PN), and 275 patients (26.9%) had at least 1 visit from a dietitian.

Factors Associated With Malnutrition at Admission Nine hundred seventy patients who had nonmissing values on 7 potential predictors identified based on clinical experience and preliminary analyses were included in the multivariate

Downloaded from pen.sagepub.com at UCSF LIBRARY & CKM on April 4, 2015

Allard et al

7

Table 3.  Contributors to Malnutrition (SGA B+C) at Admission: Logistic Regression Models. Univariate Models Predictor Age   < 65 y   ≥ 65 y Living arrangements   Lives with others at home   Lives alone at home  Othera Education   High school or less   University or college Grocery shopping   Patient or spouse   Adult child  Otherb Cooking   Patient or spouse   Adult child  Otherb No. of diagnostic categories  1  2  3 Cancer   No cancer   Cancer present CCI at admission   ≤2   >2

Odds Ratio (95% CI)

Final Multivariate Model P Value

Odds Ratio (95% CI)

P Value

Reference 1.40 (1.11, 1.77)

.005

Reference 1.07 (0.84, 1.40)

  .53

Reference 1.37 (0.997, 1.89) 1.55 (0.89, 2.70)

.058 .120

Reference 1.38 (0.99, 1.90) 1.17 (0.63, 2.17)

  .055 .62

Reference 0.74 (0.58, 0.95)

.018

Reference 2.64 (1.40, 4.96) 1.49 (1.01, 2.19)

.003 .046

Reference 2.44 (1.07, 2.44) 1.48 (1.06, 2.05)

.034 .021

Reference 1.39 (0.97, 2.00) 2.19 (1.42, 3.36)

.072 < .001

Reference 1.73 (1.14, 2.62)

.010

Reference 1.75 (1.35, 2.27)

< .001

    Reference 2.41 (1.27, 4.57) 1.31 (0.83, 2.07)

  .007 .25      

Reference 1.36 (0.95, 1.96) 1.97 (1.29, 3.01)

  .096 .002    

Reference 1.62 (1.22, 2.13)

  < .001

CCI, Charlson comorbidity index; CI, confidence interval; SGA, subjective global assessment; SGA B, moderately malnourished; SGA C, severely malnourished. a “Other” includes supportive housing, retirement home/assisted living, nursing home, and 8 patients with no permanent home. b “Other” includes other family (not children or spouse), neighbors, friends, and community support service.

analysis (Supplementary Figure 1). Among them, 437 were malnourished (SGA B + SGA C). Independent factors associated with malnutrition at admission were CCI > 2 and having 3 diagnostic categories on admission, relying on adult children for grocery shopping (compared with patient or spouse), and living alone (Table 3).

Factors Associated With Prolonged LOS Seven hundred fifty-five patients with complete data on LOS and candidate predictors were included in the analysis (Table 4, Supplementary Figure 2). Among them, 16 patients died in hospital and 32 were transferred to another hospital. HR < 1 indicates reduced chances for discharge on any particular day, that is, association with prolonged LOS. The final multivariate model, which was controlled for age, sex, living conditions, number of medications, and number of diagnostic categories, showed that malnutrition (SGA B+C) at admission was independently associated with prolonged LOS (HR, 0.73; 95% CI,

0.62–0.86). Other nutrition-related covariates found to be significant were HGS at admission, nutrition support during hospitalization, and reduced food intake during the first week. Two sensitivity analyses were performed, the first excluding in-hospital deceased patients (n = 16), and the second excluding patients with LOS ≤ 2 days (n = 31) from the sample but including deceased patients. Both models were very similar to the original one (Supplementary Table 2).

Discussion This is a comprehensive prospective multicenter cohort study demonstrating that the prevalence of malnutrition (45%) and obesity (32%) is high at admission, that complex disease and social factors are preadmission contributors to malnutrition, and that malnutrition at admission (SGA and HGS), as well as other hospital nutrition factors (food intake < 50% during the first week of stay and the use of nutrition support), are significantly and independently associated with prolonged LOS.

Downloaded from pen.sagepub.com at UCSF LIBRARY & CKM on April 4, 2015

8

Journal of Parenteral and Enteral Nutrition XX(X)

Table 4.  Association Between Nutrition-Related Parameters and Prolonged Length of Hospital Stay.a Univariate Models Parameter SGA B/C at admission BMI at admission   < 20 kg/m2   20 to < 30 kg/m2   ≥ 30 kg/m2 Handgrip strength at admission, kg, per 10 units change Preadmission weight loss ≥ 5% NPO for 3 or more daysb Nutrition supportb Food intake < 50% Had a registered dietitian visitb Age per 10-year difference Sex, male Education   Grade school or less   High school   University or college Living arrangements   Lives alone or with others at home  Otherc Admission ward: surgical Number of medications, per unit increase Number of diagnostic categories  1  2  3 CCI at admission > 2 Cancer

Final Multivariate Model

Hazard Ratio (95% CI)

P Value

Hazard Ratio (95% CI)

P Value

0.64 (0.55, 0.75)

< .001

0.73 (0.62, 0.86)

< .001

0.88 (0.67, 1.17) Reference 1.00 (0.85, 1.18) 1.19 (1.11, 1.27) 0.73 (0.62, 0.87) 0.94 (0.73, 1.21) 0.65 (0.46, 0.93) 0.72 (0.61, 0.85) 0.79 (0.66, 0.96) 0.88 (0.84, 0.92) 0.97 (0.84, 1.13)

.39 .99 < .001 < .001 .63 .018 < .001 .015 < .001 .70

Reference 1.24 (1.01, 1.52) 1.34 (1.09, 1.65)

.041 .006

Reference 0.57 (0.43, 0.75) 1.09 (0.91, 1.29) 0.95 (0.94, 0.97)

< .001 .37 < .001

Reference 0.68 (0.57, 0.81) 0.50 (0.38, 0.65) 0.73 (0.62, 0.85) 0.81 (0.66, 0.99)

< .001 < .001 < .001 .035

1.12 (1.01, 1.23)

0.61 (0.42, 0.88) 0.73 (0.62, 0.87) 0.95 (0.90, 1.01) 0.77 (0.63, 0.93)

      .026     .008 < .001   .074 .008      

Reference 0.72 (0.53, 0.96) 0.96 (0.95, 0.98) Reference 0.70 (0.59, 0.84) 0.58 (0.44, 0.76)

  .027   < .001   < .001 < .001    

BMI, body mass index; CCI, Charlson comorbidity index; CI, confidence interval; NPO, nil per os; SGA, subjective global assessment; SGA B, moderately malnourished; SGA C, severely malnourished. a Hazard ratio < 1 indicates fewer chances for discharge on any particular day and, thus, association with prolonged hospital stay. n = 755; 706 events, censored = 49 (deceased = 16, discharged to another hospital = 32, imputed from in-hospital follow-up variables, exact discharge date unknown = 1). b Time-dependent indicator variable. c Includes supportive housing, retirement home/assisted living, nursing home, and 4 patients with no permanent home.

Obesity or other BMI categories are not associated with LOS, as BMI is not a good indicator of overall nutrition status and may be influenced by edema. Others also did not find associations between BMI and LOS.2 In addition, prevalence alone would suggest that BMI on its own is not sensitive enough, as only 9% of admitted patients were in the underweight BMI category, well below reported malnutrition prevalence rates.1-8 These findings suggest that an anthropometric snapshot in time like BMI is not adequate to assess the nutrition status of patients at hospital admission. In addition, although some21 reported that higher BMI may be protective, there is a false perception that these patients are in a better nutrition state and have better outcomes; in fact, they may be disadvantaged as they may develop sarcopenia that could be difficult to detect unless a more comprehensive assessment such as SGA is completed.

SGA has been used in various patient populations worldwide5,11,32,43-48 and is associated with clinical outcomes.26,32,43,49-54 Our findings are in line with these studies in terms of prevalence and association with prolonged LOS. SGA takes into account declining body mass associated with poor food intake, poor function, and disease stress. Therefore, this makes it an excellent tool to assess the global effect of malnutrition, which occurs when intake of protein-energy falls below requirements while disease increases metabolic rate, further increasing energy-protein deficit. Other parameters reflecting poor nutrition intake were independent predictors of prolonged LOS. We found that 31% of patients ate < 50% of their food, similar to a previous study.55 Monitoring food intake during hospitalization is important to prevent nutrition deterioration, but the effect of nutrition support is unclear. As previously reported,25 nutrition support was

Downloaded from pen.sagepub.com at UCSF LIBRARY & CKM on April 4, 2015

Allard et al

9

associated with prolonged LOS, which was longer for those on PN (median 20 days) and EN (median 21 days) as opposed to oral supplementation (median 10 days) (data not shown). The prolonged LOS, particularly in those receiving EN or PN, may reflect a special group of patients who had prior medical or surgical complications, which lead to weight loss and poor recovery, triggering a nutrition support consultation. These results may also reflect the nature of observational studies, as nutrition intervention studies reported benefits.56,57 HGS was an independent factor associated with LOS. HGS can be influenced not only by malnutrition but also by muscle depletion from lack of mobility or sarcopenia secondary to inflammation or aging. Associations with malnutrition, decline, and outcomes have been reported previously,38,58-61 with HGS being particularly useful for assessing acute changes in nutrition status61-63 and predicting certain outcomes,59,61,64,65 including mortality.66 Issues to be considered when using this parameter as the sole measurement of nutrition status are level of participation, level of cognitive function, presence of musculoskeletal disease, drugs, and operator skills. Therefore, it is preferable not to use this alone when assessing nutrition status. Several preadmission factors were associated with malnutrition at admission and with LOS. Understanding who the malnourished are at admission, in addition to identifying those who are malnourished or at risk, can help to more readily provide nutrition care and plan better discharge once the patients return to the community. The independent contributors reflected complex disease (CCI > 2 on admission, having 3 diagnostic categories) and social factors (relying on adult children for grocery shopping [compared with patient or spouse], and living alone). Cancer alone and age ≥ 65 were not independent predictors of malnutrition, contrary to what was reported,2 but CCI, which includes cancer in addition to other comorbidities, was a better predictor of malnutrition in our study. Another study, using a nutrition screening tool (NRS-2002) rather than an assessment tool, found that the loss of functional autonomy, illiteracy, age, being male, living alone (single, divorced, widowed), and smoking were factors associated with risk of undernutrition at admission.12 We did not find that male sex was a contributor to malnutrition at admission, but being male was associated with prolonged LOS. Whether this is due to higher nutrition requirements that are not being met or other medical or physiological reasons is unclear. No other studies reported the social contributors to malnutrition assessed by SGA. The strength of this study was its large and representative sample from several hospitals across the country and the assessment of multiple nutrition aspects and contributors. We used standardized assessments of nutrition status, which were performed by trained coordinators and considered preadmission social and disease-related contributors of malnutrition. This is the only study using multivariate Cox PH analysis to examine the relationship between malnutrition and LOS while controlling for disease, socioeconomic, and treatment-related

factors. It is also the first study where relationship between LOS and complete in-hospital fasting, dietitian visits, and nutrition support was examined, accounting for the timing of these events in the course of hospitalization and to show independent relationship between reduced food intake and prolonged LOS. These unique strengths not only establish more firmly the relevance of nutrition to LOS but also provide direction for further research and practice. For example, we should consider more fully social factors before and after hospitalization that may affect the response to nutrition intervention. The limitation was the lack of nutrition intervention that would confirm causality between malnutrition and LOS. Other limitations include presence of missing data, which reflects the reality of observational studies in a complex and busy health system environment, and hospital voluntary enrollment, which may have introduced a selection bias. We also cannot fully exclude the possibility that there is heterogeneity in the patient population, clinical practice, and policies between institutions and provinces that may have affected outcomes.

Conclusion In this study, the prevalence of malnutrition and obesity was 45% and 32%, respectively, and preadmission contributors to malnutrition reflect complex disease and age-associated social situations. Malnutrition at admission and poor food intake early during hospitalization were associated with prolonged LOS, suggesting that prompt nutrition care or intervention and monitoring should be performed when patients are admitted to hospital. Although nutrition support was associated with prolonged LOS, this may reflect the use of EN and PN in patients who are already malnourished as a result of prior complications and may suggest delays in initiating nutrition support.

Author Contributions J.P.A. and H.K. with the participation of K.N.J., M.L., D.R.D., L.G., and H.P. contributed to the conception and design of the study; A.T., J.P.A., W.L., E.V. with the help of H.K., K.N.J., M.L., D.R.D., L.G., H.P., and P.B. contributed to the analysis and interpretation of the data; J.P.A. and A.T. drafted the article; J.P.A., H.K., M.L., K.N.J., D.R.D., L.G., A.T., and W.L. critically revised the article for important intellectual content; J.P.A., H.K., K.N.J., M.L., L.G., D.R.D., P.B., E.V., A.T., H.P., B.D., and W.L. had final approval of the article; J.P.A., D.R.D., L.G., P.B., K.N.J., M.L., and the centers mentioned in the acknowledgment provided study materials or patients; A.T. and W.L. provided statistical expertise; J.P.A. and H.K. obtained funding; B.D., E.V. and A.T. provided administrative, technical, or logistical support; and B.D., E.V., and A.T. collected and assembled the data.

Acknowledgments The following institutions participated in the study: South West Health, Regional Health Authority A, Réseau de Santé Vitalité Health Network, Jewish General Hospital, Centre de santé et de

Downloaded from pen.sagepub.com at UCSF LIBRARY & CKM on April 4, 2015

10

Journal of Parenteral and Enteral Nutrition XX(X)

services sociaux du Sud de Lanaudière, Centre hospitalier de l’Université de Montréal, University Health Network, University of Waterloo, University of Guelph, Toronto East General Hospital, St Michael’s Hospital, McMaster University, Niagara Health System, University of Manitoba, Regina Qu’Appelle Health Region, University of Alberta, University of Calgary, University of British Columbia, Vancouver Island Health Authority. Hospital name followed by site investigator and site coordinator: Yarmouth Regional Hospital (Patrice Simpson, Natalie Pothier); Réseau de Santé Vitalité Health Network: Campbellton Regional Hospital (Manon Laporte, Isabelle Caissie), George L. Dumont Hospital (Manon Laporte, Claire Johnson); Jewish General Hospital (Paule Bernier, Marissa Ranallo); Hopital Notre Dame du CHUM (Danuta Balicki, Josee Beaudoin); Hopital Pierre Le-Gardeur (Catherine Talbot, Genevieve St. Jean); University Health Network (Johane Allard, Bianca Arendt); St Michael’s Hospital (Khursheed Jeejeebhoy, Stanley Zhang); Toronto East General Hospital (Angela Chen, Cathy You); Hamilton Health Sciences (David Armstrong, Anda Andic); Niagara Health System (Adam Rahman, Andrea Digweed); St Boniface General Hospital (Donald Duerksen, Laura Toews); Regina Qu’Appelle Health Region (Roseann Nasser, Sharisse Kushniruk); Foothills Medical Centre (Maitreyi Raman, Sally Ho); University of Alberta (Leah Gramlich, Nicole Journault); Vancouver General Hospital (Theresa Cividin, Claudia Dow); Nanaimo Regional General Hospital (Heather Tant, Tracy Lister).

References 1. Guigoz Y. The Mini Nutritional Assessment (MNA) review of the literature—what does it tell us? J Nutr Health Aging. 2006;10(6):466-485; discussion 85-87. 2. Pirlich M, Schutz T, Norman K, et al. The German hospital malnutrition study. Clin Nutr. 2006;25(4):563-572. 3. Banks M, Ash S, Bauer J, Gaskill D. Prevalence of malnutrition in adults in Queensland public hospitals and residential aged care facilities. Nutrition & Dietetics. 2007;64(3):172-178. 4. Kondrup J, Sorensen JM. The magnitude of the problem of malnutrition in Europe. Nestle Nutrition Workshop Series Clinical & Performance Program. 2009;12:1-14. 5. Waitzberg DL, Caiaffa WT, Correia MI. Hospital malnutrition: the Brazilian national survey (IBRANUTRI): a study of 4000 patients. Nutrition. 2001;17(7-8):573-580. 6. Aghdassi E, McArthur M, Liu B, McGeer A, Simor A, Allard JP. Dietary intake of elderly living in Toronto long-term care facilities: comparison to the dietary reference intake. Rejuvenation Res. 2007;10(3):301-309. 7. Singh H, Watt K, Veitch R, Cantor M, Duerksen DR. Malnutrition is prevalent in hospitalized medical patients: are housestaff identifying the malnourished patient? Nutrition. 2006;22(4):350-354. 8. Villalon L, Laporte M, Carrier N. Nutrition screening for seniors in health care facilities: a survey of health professionals. Can J Diet Pract Res. 2011;72(4):162-169. 9. Choban PS, Flancbaum L. The impact of obesity on surgical outcomes: a review. J Am Coll Surg. 1997;185(6):593-603. 10. Wells JC. Obesity as malnutrition: the role of capitalism in the obesity global epidemic. Am J Hum Biol. 2012;24(3):261-276. 11. Pirlich M, Schutz T, Kemps M, et al. Social risk factors for hospital malnutrition. Nutrition. 2005;21(3):295-300. 12. Amaral TF, Matos LC, Teixeira MA, Tavares MM, Alvares L, Antunes A. Undernutrition and associated factors among hospitalized patients. Clin Nutr. 2010;29(5):580-585.

13. Kondrup J, Johansen N, Plum LM, et al. Incidence of nutritional risk and causes of inadequate nutritional care in hospitals. Clin Nutr. 2002;21(6):461-468. 14. Hernandez AV, Kaw R, Pasupuleti V, et al. Association between obesity and postoperative atrial fibrillation in patients undergoing cardiac operations: a systematic review and meta-analysis. Ann Thorac Surg. 2013;96(3): 1104-1116. 15. Markoff B, Amsterdam A. Impact of obesity on hospitalized patients. Mt Sinai J Med. 2008;75(5):454-459. 16. Stein PD, Goldman J. Obesity and thromboembolic disease. Clin Chest Med. 2009;30(3):489-493. 17. Finkelstein EA, Khavjou OA, Thompson H, et al. Obesity and severe obesity forecasts through 2030. Am J Prev Med. 2012;42(6):563-570. 18. Wolf AM, Finer N, Allshouse AA, et al. PROCEED: Prospective Obesity Cohort of Economic Evaluation and Determinants: baseline health and healthcare utilization of the US sample. Diabetes Obes Metab. 2008;10(12):1248-1260. 19. Anis AH, Zhang W, Bansback N, Guh DP, Amarsi Z, Birmingham CL. Obesity and overweight in Canada: an updated cost-of-illness study. Obes Rev. 2010;11(1):31-40. 20. Moffatt E, Shack LG, Petz GJ, Sauve JK, Hayward K, Colman R. The cost of obesity and overweight in 2005: a case study of Alberta, Canada. Can J Public Health. 2011;102(2):144-148. 21. Hogue CW Jr, Stearns JD, Colantuoni E, et al. The impact of obesity on outcomes after critical illness: a meta-analysis. Intensive Care Med. 2009;35(7):1152-1170. 22. Lieffers JR, Bathe OF, Fassbender K, Winget M, Baracos VE. Sarcopenia is associated with postoperative infection and delayed recovery from colorectal cancer resection surgery. Br J Cancer. 2012;107(6): 931-936. 23. Peng PD, van Vledder MG, Tsai S, et al. Sarcopenia negatively impacts short-term outcomes in patients undergoing hepatic resection for colorectal liver metastasis. HPB. 2011;13(7):439-446. 24. Leibovitz E, Giryes S, Makhline R, Zikri Ditch M, Berlovitz Y, Boaz M. Malnutrition risk in newly hospitalized overweight and obese individuals: Mr NOI. Eur J Clin Nutr. 2013;67(6):620-624. 25. Caccialanza R, Klersy C, Cereda E, et al. Nutritional parameters associated with prolonged hospital stay among ambulatory adult patients. CMAJ. 2010;182(17):1843-1849. 26. Lim SL, Ong KC, Chan YH, Loke WC, Ferguson M, Daniels L. Malnutrition and its impact on cost of hospitalization, length of stay, readmission and 3-year mortality. Clin Nutr. 2012;31(3):345-350. 27. Kirkland LL, Kashiwagi DT, Brantley S, Scheurer D, Varkey P. Nutrition in the hospitalized patient. J Hosp Med. 2013;8(1):52-58. 28. Norman K, Pichard C, Lochs H, Pirlich M. Prognostic impact of diseaserelated malnutrition. Clin Nutr. 2008;27(1):5-15. 29. Sorensen J, Kondrup J, Prokopowicz J, et al. EuroOOPS: an international, multicentre study to implement nutritional risk screening and evaluate clinical outcome. Clin Nutr. 2008;27(3):340-349. 30. Holmes S. The effects of undernutrition in hospitalised patients. Nurs Stand. 2007;22(12):35-38. 31. Leandro-Merhi VA, de Aquino JLB, Sales Chagas JF. Nutrition status and risk factors associated with length of hospital stay for surgical patients. JPEN J Parenter Enteral Nutr. 2011;35(2):241-248. 32. Correia MITD, Waitzberg DL. The impact of malnutrition on morbidity, mortality, length of hospital stay and costs evaluated through a multivariate model analysis. Clin Nutr. 2003;22(3):235-239. 33. Kyle UG, Schneider SM, Pirlich M, Lochs H, Hebuterne X, Pichard C. Does nutritional risk, as assessed by Nutritional Risk Index, increase during hospital stay? A multinational population-based study. Clin Nutr. 2005;24(4):516-524. 34. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40(5):373-383.

Downloaded from pen.sagepub.com at UCSF LIBRARY & CKM on April 4, 2015

Allard et al

11

35. Detsky AS, McLaughlin JR, Baker JP, et al. What is subjective global assessment of nutritional status? JPEN J Parenter Enteral Nutr. 1987;11(1):8-13. 36. World Health Organization. Obesity: preventing and managing the global epidemic. Report of a WHO consultation. World Health Organ Tech Rep Ser. 2000;894:i-xii, 1-253. 37. Haute Autorité de Santé. Clinical Practice Guidelines. Nutritional Support Strategy for Protein-Energy Malnutrition in the Elderly. Saint Denis, France: Haute Autorité de Santé; 2007. 38. Matos LC, Tavares MM, Amaral TF. Handgrip strength as a hospital admission nutritional risk screening method. Eur J Clin Nutr. 2007;61(9):1128-1135. 39. Schindler K, Pernicka E, Laviano A, et al. How nutritional risk is assessed and managed in European hospitals: a survey of 21,007 patients findings from the 2007-2008 cross-sectional nutritionDay survey. Clin Nutr. 2010;29(5):552-559. 40. Agresti A. Categorical Data Analysis. 3rd ed. Hoboken, NJ: Wiley; 2013. 41. Klein JP, Moeschberger ML. Survival Analysis: Techniques for Censored and Truncated Data. 2nd ed. New York: Springer; 2003. 42. Bohannon RW, Peolsson A, Massy-Westropp N, Desrosiers J, Bear-Lehman J. Reference values for adult grip strength measured with a Jamar dynamometer: a descriptive meta-analysis. Physiotherapy. 2006;92(1):11-15. 43. Sacks GS, Dearman K, Replogle WH, Cora VL, Meeks M, Canada T. Use of subjective global assessment to identify nutrition-associated complications and death in geriatric long-term care facility residents. J Am Coll Nutr. 2000;19(5):570-577. 44. Baccaro F, Moreno JB, Borlenghi C, et al. Subjective global assessment in the clinical setting. JPEN J Parenter Enteral Nutr. 2007;31(5):406-409. 45. Baker JP, Detsky AS, Wesson DE, et al. Nutritional assessment: a comparison of clinical judgement and objective measurements. N Engl J Med. 1982;306(16):969-972. 46. Kondrup J, Allison SP, Elia M, Vellas B, Plauth M. ESPEN guidelines for nutrition screening 2002. Clin Nutr. 2003;22(4):415-421. 47. Cooper BA, Bartlett LH, Aslani A, Allen BJ, Ibels LS, Pollock CA. Validity of subjective global assessment as a nutritional marker in endstage renal disease. Am J Kidney Dis. 2002;40(1):126-132. 48. Pham NV, Cox-Reijven PL, Greve JW, Soeters PB. Application of subjective global assessment as a screening tool for malnutrition in surgical patients in Vietnam. Clin Nutr. 2006;25(1):102-108. 49. Campbell KL, Ash S, Bauer JD, Davies PS. Evaluation of nutrition assessment tools compared with body cell mass for the assessment of malnutrition in chronic kidney disease. J Ren Nutr. 2007;17(3):189-195. 50. Raslan M, Gonzalez MC, Torrinhas RS, Ravacci GR, Pereira JC, Waitzberg DL. Complementarity of Subjective Global Assessment (SGA) and Nutritional Risk Screening 2002 (NRS 2002) for predicting poor clinical outcomes in hospitalized patients. Clin Nutr. 2011;30(1):49-53.

51. Schiesser M, Muller S, Kirchhoff P, Breitenstein S, Schafer M, Clavien PA. Assessment of a novel screening score for nutritional risk in predicting complications in gastro-intestinal surgery. Clin Nutr. 2008;27(4): 565-570. 52. Wakahara T, Shiraki M, Murase K, et al. Nutritional screening with Subjective Global Assessment predicts hospital stay in patients with digestive diseases. Nutrition. 2007;23(9):634-639. 53. Kyle UG, Genton L, Pichard C. Hospital length of stay and nutritional status. Curr Opin Clin Nutr Metab Care. 2005;8(4):397-402. 54. Raslan M, Gonzalez MC, Dias MC, et al. Comparison of nutritional risk screening tools for predicting clinical outcomes in hospitalized patients. Nutrition. 2010;26(7-8):721-726. 55. Sullivan DH, Sun S, Walls RC. Protein-energy undernutrition among elderly hospitalized patients: a prospective study. JAMA. 1999;281(21):2013-2019. 56. Kruizenga HM, Van Tulder MW, Seidell JC, Thijs A, Ader HJ, Van Bokhorst-de van der Schueren MA. Effectiveness and cost-effectiveness of early screening and treatment of malnourished patients. Am J Clin Nutr. 2005;82(5):1082-1089. 57. Milne AC, Avenell A, Potter J. Meta-analysis: protein and energy supplementation in older people. Ann Intern Med. 2006;144(1):37-48. 58. Humphreys J, de la Maza P, Hirsch S, Barrera G, Gattas V, Bunout D. Muscle strength as a predictor of loss of functional status in hospitalized patients. Nutrition. 2002;18(7-8):616-620. 59. Klidjian AM, Foster KJ, Kammerling RM, Cooper A, Karran SJ. Relation of anthropometric and dynamometric variables to serious postoperative complications. Br Med J. 1980;281(6245):899-901. 60. Vaz M, Thangam S, Prabhu A, Shetty PS. Maximal voluntary contraction as a functional indicator of adult chronic undernutrition. Br J Nutr. 1996;76(1):9-15. 61. Webb AR, Newman LA, Taylor M, Keogh JB. Hand grip dynamometry as a predictor of postoperative complications reappraisal using age standardized grip strengths. JPEN J Parenter Enteral Nutr. 1989;13(1):30-33. 62. Lopes J, Russell DM, Whitwell J, Jeejeebhoy KN. Skeletal muscle function in malnutrition. Am J Clin Nutr. 1982;36(4):602-610. 63. Windsor JA, Hill GL. Grip strength: a measure of the proportion of protein loss in surgical patients. Br J Surg. 1988;75(9):880-882. 64. Klidjian AM, Archer TJ, Foster KJ, Karran SJ. Detection of dangerous malnutrition. JPEN J Parenter Enteral Nutr. 1982;6(2):119-121. 65. Mahalakshmi VN, Ananthakrishnan N, Kate V, Sahai A, Trakroo M. Handgrip strength and endurance as a predictor of postoperative morbidity in surgical patients: can it serve as a simple bedside test? Int Surg. 2004;89(2):115-121. 66. Newman AB, Kupelian V, Visser M, et al. Strength, but not muscle mass, is associated with mortality in the health, aging and body composition study cohort. J Gerontol A Biol Sci Med Sci. 2006;61(1):72-77.

Downloaded from pen.sagepub.com at UCSF LIBRARY & CKM on April 4, 2015