Clinical applications of body composition and functional status tools for nutrition assessment of hospitalized adults: A systematic review

Abstract

Background

No global consensus exists on diagnostic criteria for malnutrition. Muscular deficits and functional impairments are major components of available malnutrition diagnostic frameworks because these facets of nutrition status significantly impact outcomes. The purpose of this review is to explore which body composition assessment (BCA) and functional status assessment (FSA) tools are being used for nutrition assessment (NA) and monitoring the response to nutrition interventions (RNIs) in adult inpatients.

Methods

A literature search of Embase, Medline (Ovid), Web of Science, and Cochrane Central was performed to identify studies that used BCA and/or FSA tools for NA (along with an accepted NA diagnostic framework) and/or for monitoring RNI in adult inpatients.

Results

The search yielded 3667 articles; 94 were included in the review. The number of studies using BCA and/or FSA tools for NA was 47 and also 47 for monitoring RNI. Seventy-nine percent of studies used bioimpedance for BCA, and 97% that included FSA utilized handgrip strength. When compared against sets of diagnostic criteria, many of the BCA and FSA tools showed promising associations with nutrition status.

Conclusion

Bioimpedance methods are the most widely used bedside BCA tools, and handgrip strength is the most widely used FSA tool; however, these methods are being used with a variety of protocols, algorithms, and interpretation practices in heterogeneous populations. To create a standardized nutrition status assessment process there is a need for validation studies on bedside methods and the development of globally standardized assessment protocols in clinical inpatient settings.

Abbreviations

AMA

arm muscle area

AND

Academy of Nutrition and Dietetics

ASPEN

American Society for Parenteral and Enteral Nutrition

BCA

body composition assessment

BIS

bioimpedance spectroscopy

BMI

body mass index

CC

calf circumference

CT

computed tomography

CTL3

skeletal muscle cross-sectional area at the level of L3, assessed by CT

DXA

dual-energy x-ray absorptiometry

EDC

ESPEN diagnostic criteria

ESPEN

European Society for Parenteral and Enteral Nutrition

FELANPE

Federacion Latinoamericana de Terapia Nutricional, Nutricion Clinica y Metabolismo

FM

fat mass

FFM

fat-free mass

FFMI

fat-free mass index

FSA

Functional Status Assessment

GLIM

Global Leadership Initiative on Malnutrition

HGS

handgrip strength

HU

Hounsfield units (attenuation)

MCC

Malnutrition Consensus Criteria

MeSH

medical subject headings of the National Library of Medicine

MF-BIA

multiple-frequency bioelectrical impedance analysis

MRI

magnetic resonance imaging

MUAC

mid-upper arm circumference

NA

nutrition assessment

NSA

nutrition status assessment

PENSA

Parenteral and Enteral Nutrition Society of Asia

PhA

phase angle

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analysis

RNI

response to nutrition intervention

SF-BIA

single-frequency bioelectrical impedance analysis

SGA

Subjective Global Assessment

SPhA

standardized phase angle

US

ultrasound

CLINICAL RELEVANCY STATEMENT

It is widely accepted that muscle is an essential component of nutrition status, and it is recognized as a core phenotypic criterion in all four of the major nutrition assessment diagnostic frameworks. Clinicians are using bedside assessment tools with a variety of protocols, algorithms, and interpretation practices in highly heterogenous populations. Our focus in conducting this review was to survey the literature to gain perspective on the ways body composition and functional status assessment tools are being used to evaluate muscle as a component of nutrition status in clinical settings.

INTRODUCTION

Nutrition status has generally been defined as an individual's health condition as influenced by the intake and utilization of nutrients,¹ whereas malnutrition has been defined as “an acute, subacute, or chronic state of nutrition in which a combination of varying degrees of undernutrition or overnutrition, with or without inflammatory activity, have led to a change in body composition and diminished physiological function.”^{2, 3} Nutrition assessment (NA) is the quantitative evaluation of nutrition status and involves the interpretation of nutrient balance, body composition, and laboratory and clinical parameters.⁴ Adequate nutrition status is achieved in the general population when an individual is able to ingest, digest, absorb, and use adequate proportions of macronutrients and micronutrients to support all of the body's physiological processes, without excessive gain of adipose tissue or loss of skeletal muscle.^{2, 3, 5} Because of the increased physiological demands that are experienced by patients recovering from disease, surgery, or injury, hospitalized patients may experience greater rates of malnutrition than the general population.⁶ When a patient is well nourished, the body is able to support physiological processes related to immunity and recovery, promoting optimal patient outcomes. On the other hand, a malnourished patient may experience decreased immunity, physical, and physiological function, along with increased readmission rates, length of stay, mortality, and cost of care.^{2, 7-11} Therefore, monitoring the nutrition status of hospitalized patients is a vital part of the patient care process.

Presently, no global consensus on diagnostic criteria for the identification of malnutrition exists. The European Society for Clinical Nutrition and Metabolism (ESPEN) have published the ESPEN diagnostic criteria (EDC)¹²; the American Society for Parenteral and Enteral Nutrition (ASPEN) and the Academy of Nutrition and Dietetics (AND) have published the Malnutrition Consensus Criteria (MCC); and, more recently, the Global Leadership Initiative on Malnutrition (GLIM), which has involved leaders from ASPEN, ESPEN, the Federacion Latinoamericana de Terapia Nutricional, Nutricion Clinica y Metabolismo, and the Parenteral and Enteral Nutrition Society of Asia, has proposed the GLIM criteria with the intent that a global consensus may be reached once the criteria are validated.^{9, 13} Additionally, the Subjective Global Assessment (SGA), an NA diagnostic framework that has been validated for use in clinical populations since 1987,¹⁴ is also widely used today by clinicians. The diagnostic criteria implemented by the four most widely accepted NA diagnostic frameworks are presented in Table 1. Although each of these NA diagnostic frameworks vary in the exact criteria used to diagnose malnutrition, each includes at least one indicator of muscle mass and/or function. The objective measurement of these parameters using available bedside tools provides clinicians with a quantitative interpretation of nutrition status, whereas the other components of the assessment may be subjective or prone to bias. Muscular deficits and functional impairments are major components of the SGA, EDC, MCC, and GLIM criteria because these facets of nutrition status significantly impact disease prognosis and patient outcomes.^{9, 15-20}

TABLE 1. Overview of diagnostic frameworks for nutrition assessment

Criteria	SGA	EDC	ASPEN/AND MCC	GLIM criteria
1. Body weight	Change over previous: 6 months; 2 weeks	>5% loss over the last 3 months; >10% loss over any time course	2%, 5%, or 7%–10% loss over 1 week, 1 month, or 3–6 months, respectively	Current BMI <18.5, <20, or <22; varies by age and severity
2. Dietary intake	Assesses change in diet: change duration; change type	N/A	Percentage of estimated needs received	Percentage of estimated needs received
3. Gastrointestinal symptoms	Includes assessment of GI symptoms	N/A	N/A	Includes assessment of GI symptoms
4. Functional capacity	Subjective assessment of functional capacity	N/A	Handgrip strength assessment	N/A
5. Disease and its relation to nutrition requirements	Included as potential risk factor	N/A	Included as etiology	Included as etiology
6. Physical	Loss of subcutaneous fat, muscle wasting, edema, and/or ascites; assessed on a scale of 0–3 (0 = none; 3 = severe)	Cutoff values for low BMI and low FFMI	Loss of subcutaneous fat, loss of muscle, and fluid accumulation; assessed as mild, moderate, severe	Reduced muscle mass, assessed by a validated assessment tool

• Abbreviations: ASPEN/AND MCC, American Society for Parenteral and Enteral Nutrition/Academy of Nutrition and Dietetics Malnutrition Consensus Criteria; BMI, body mass index; EDC, European Society for Parenteral and Enteral Nutrition diagnostic criteria; FFMI, fat-free mass index; GLIM, Global Leadership Initiative on Malnutrition; N/A, not applicable; SGA, Subjective Global Assessment.

Identifying how clinicians are currently assessing and monitoring muscle and function in clinical settings is a necessary first step toward developing a global consensus on preferred bedside tools, optimal assessment protocols, and appropriate interpretation of measures, which would allow for the implementation of muscle assessments as part of a standardized NA process. In recent years, there have been a limited number of publications on the validation of various bedside tools for determining body composition in clinical populations. However, there is no comprehensive report on which bedside body composition assessment (BCA) and/or functional status assessment (FSA) tools are being used for NA in inpatient settings.

The purpose of this review is to provide a general overview of the BCA and FSA tools that are being used in clinical populations to assess nutrition status and/or monitor response to nutrition interventions (RNIs) in hospitalized adult patients.¹⁹

METHODS

Given our aim to document the BCA and FSA bedside tools being used in adult inpatient settings to assess nutrition status and RNIs, our search strategy was necessarily broad and comprehensive.

Search strategy

This systematic review is registered at FigShare.

The search strategy followed the Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines,¹ and used a combination of keywords and medical subject headings of the National Library of Medicine (MeSH) terms related to body composition, anthropometry, nutrition status assessment, and functional status. In December 2019, the following electronic databases were searched for eligible studies: Embase, Medline (Ovid), Web of Science, and Cochrane Central. A second database search, following the same procedure, was performed in December 2021. The specific search strategies were created by a health science librarian with expertise in systematic review searching. File S9 outlines the exact search strategy.

Screening strategy

Inclusion criteria for this review include (1) a population of adults (age ≥18 years); (2) at least one clinical assessment of body composition and/or functional status; (3) the use of BCA and/or FSA to describe nutrition status and/or RNI; and (4) that the participants are hospitalized inpatients during the initial assessment. Clinical assessments that are related to nutrition status but do not provide measures of body composition, such as indirect calorimetry, were not included as part of this review. Additionally, assessment methods that would not be feasible in clinical settings, such as tritiated water dilution and in vivo neutron activation analysis, were not included in this review. Finally, the search was limited to English-language studies.

Four independent reviewers (K.A.B., J.F.O., C.P.E., and L.O.S.) were split into two teams of two. During the first round of abstract screening, articles were divided in half alphabetically by the first author's last name. Team 1 (K.A.B. and C.P.E.) screened articles in which the first author's last name began with letters A–J, and team 2 (J.FO. and L.O.S.) screened abstracts in which the first author's last name began with letters K–Z. Papers that seemingly met the inclusion criteria were pushed into a second round of abstract screening. This second round was performed having team 1 and team 2 swap alphabetical halves (team 1 screened K–Z and team 2 screened A–J), and abstracts of included articles were rescreened in a complementary fashion. Articles with discrepancies between reviewers were discussed by all reviewers to determine which papers would move on to full-text screening.

Two reviewers (J.F.O. and L.O.S.) conducted full-text screening. Both reviewers independently screened all of the articles that had made it to this phase. All discrepancies were discussed between reviewers, and disagreements were settled by a third reviewer (C.P.E.) before the final inclusion of articles into the review.

Articles published between the two search dates were screened following the same screening protocol as the first round. Because this search yielded fewer articles, only one round of abstract screening was performed. After this final round, full-text articles that seemingly met the inclusion criteria were independently screened for inclusion by all four reviewers, and eligible articles were included in the review.

Data analysis

One reviewer (L.O.S.) extracted all relevant body composition, functional status, clinical, demographic, and nutrition data from the included articles (Figure 1).

FIGURE 1

Flow diagram of literature search and results. BCA, body composition assessment; FSA, functional status assessment; NA, nutrition assessment.

RESULTS

Ultimately, 94 articles were included in this review. Figure 2 displays the number of studies in which the various BCA and FSA tools were utilized.

FIGURE 2

Categorization of studies by BCA and FSA tools. AMA, arm muscle area; BCA, body composition assessment; BIS, bioimpedance spectroscopy; CC, calf circumference; CT, computed tomography; DXA, dual-energy x-ray absorptiometry; FSA, functional status assessment; MAMC, midarm muscle circumference; MF-BIA, multifrequency bioelectrical impedance analysis; MUAC, mid-upper arm circumference; SF-BIA, single-frequency bioelectrical impedance analysis; US, ultrasound.

Tables 2 and 3 provide overviews of the included articles regarding population, study design, BCA and FSA techniques, and NA tools used for comparison. Table 2 relates to articles in which the primary focus was NA, and Table 3 relates to articles in which the primary focus was RNIs. Instrument specifics, protocols, timing of assessments, parameters assessed, and significant findings are provided in the Supporting Information files. Tables S1–S4 correspond to the NA articles included in Table 2, and Tables S5–S8 correspond to articles included in Table 3.

Table 2. Overview of Studies Using Body Composition and/or Functional Status Assessment Tools for Nutrition Assessment

First author, year	Study design/duration	Population	Sample characteristics N; age (mean/median (SD/range)); %F Female	Body composition/Functional assessment Techniques	Assessment operator	Nutrition assessment comparator
Al-Kalaldeh, 2020^21	CS	Intensive Care Unit	N = 411	SF-BIA	Trained investigator	NUTRIC
			age = 60.7 (18.71)
			%F = 40.1%
Al-Kalaldeh, 2018^22	CS	Intensive Care Unit	N = 321	MF-BIA	Trained investigator	MUST
			age = 60.03
			%F = 34.3%
Almasaudi, 2019^23	CS	Oncology/Surgical	N = 363	CT	Nurse	MUST
			age = 66 (12)
			%F = 45.18%
Bakshi, 2016^24	CS	Liver Transplant Recipients	N = 54	MF-BIA	Operator not described	SGA
			age = 48.3 (10.2)	Anthropometric
			%F = 27.8%	Functional
Cao, 2021^25	Longitudinal/38D	Head and Neck Cancer	N = 287	MF-BIA	Trained investigators	ESPEN
			age = 54 (44-63)
			%F = 34.8%
Chapple, 2017^26	Prospective observational/LOS	Traumatic Brain Injury Intensive Care Unit	N = 37	US	Trained investigator	SGA
			age = 45.3 (15.8)	DXA
			%F = 13%
Del Giorno, 2020^27	Retrospective observational/CS	COVID-19	N = 90	SF-BIA	Dietitian	NRS-2002
			age = 64.5 (13.7)
			%F = 32.2%
Ding, 2018^28	Longitudinal/6W	Oncology	N = 48	MF-BIA	Operator not described	PG-SGA SPEN
			age = 47 (32-66)
			%F = 25%
do Amaral Paes, 2018^29	CS	Oncology/Sepsis	N = 31	SF-BIA	Operator not described	NUTRIC
			age = 25.1 (21.7-30.1)
			%F = 51.6%
Faisy, 2000^30	CS	Chronic Obstructive Pulmonary Disease	N = 51	MF-BIA	Physician performed anthropometry. MF-BIA operator not described	HSNAT
			age = 69 (11)	Anthropometric
			%F = 17.6%
Fetterplace, 2019^31	Longitudinal/LOS	Mechanically Ventilated Intensive Care Unit	N = 60	BIS	BIS and Anthropometric assessed by Dietitian.	SGA
			age = 58 (16)	Anthropometric		NUTRIC
					FSAs assessed by Physiotherapists
			%F = 45%	Functional
Gabrielson, 2021^32	Longitudinal/8W-12W	Gastric/Colorectal Cancer	N = 41	Anthropometric	Dietitian	PG-SGA
			age = 58.5 (11.3)	SF-BIA
			%F = 43.9%	Functional
Galata, 2019^33	CS	Malnourished Oncology	N = 36	MF-BIA	Operator not described	NRS-2002
			age = 62.6 (30-87)	Functional
			%F = 55.6%
Hengsterman, 2007^34	CS	Elderly	N = 484	SF-BIA	Trained investigator	MNA
			age = 79.6 (7.6)
			%F = 67.4%
Huang D.D., 2021^35	CS	Elderly Gastrectomy	N = 597	Functional	Operator not described	GLIM
			age = 72 (8)	CT		NRS-2002
			%F = 22.5%
Irisawa, 2020^36	Longitudinal/4W	Stroke Rehabilitation	N = 179	MF-BIA	Operator not described	GNRI
			age = 79.7
			%F = 50.3%
Jansen, 2019^37	CS	Malnourished Intensive Care Unit	N = 169	SF-BIA	Operator not described	SGA
			age = 60.27 (16.74)	Anthropometric
			%F = 43.2%
Knappe-Drzikova, 2019^38	Longitudinal ≤ 108M (mean 54.8M)	Gastroenterological/Hepatological	N = 644	MF-BIA	Nutritionists	NRS
			age = 58	Anthropometric		SGA
			%F = 45.2%
Koroušić Seljak, 2020^39	CS	General Hospital	N = 207	MF-BIA	Dietitians	NRS-2002
			age = not stated	Functional		MNA-SF
			%F = 52.2%
Kyle, 2002^40	CS	General Hospital	N = 995	SF-BIA	Hospital Nutrition Unit	SGA
			age = not stated
			%F = 47.2%
Kyle, 2012^41	CS	General Hospital	N = 649	SF-BIA	Hospital Nutrition Unit	SGA
			age = not clearly stated			NRS-2002
			%F = 41.1%
Laimer, 2021^42	Longitudinal/10D	Medication-Related Osteonecrosis of the Jaw	N = 58	MF-BIA	Operator not described	NRS-2002
			age = 70.5 mean (46-86)			NRI
			%F = 53.0%
Lambell, 2021^43	Prospective observational/CS	Intensive Care Unit	N = 50	CT	BIS operator not described. CT analyzed by trained investigators	SGA
			age = 52 (20)	BIS
			%F = 24%	Anthropometric
Lee, 2015^44	Retrospective/CS	Intensive Care Unit	N = 66	MF-BIA	Operator not described	HSNAT
			age = 63.1 (15.7)
			%F = 36.4%
Maeda, 2017^45	CS	Geriatric	N = 1,164	MF-BIA	Dietitians	ESPEN
			age = 83.5 (8.2)	Anthropometric
			%F = 56.2%
Marin, 2011^46	Longitudinal/LOS	Amyotrophic Lateral Sclerosis	N = 92	MF-BIA	Operator not described	HSNAT
			age = 65.6	Anthropometric
			%F = 50%
Mulasi, 2018^47	Longitudinal/6M	Head and Neck Cancer	N = 19	MF-BIA	Operator not described	PG-SGA
			age = 59 (7)	Functional		ASPEN MCC
			%F = 5%
Nematy, 2012^48	Longitudinal/LOS	Neurological Intensive Care Unit	N = 29	MF-BIA	Dietitians and medical team	NRS-2002
			age = ~44	Anthropometric
			%F = not stated
Ni Bhuachalla, 2018^49	CS	Oncology	N = 725	CT	Operator not described	MUST
			age = 64.3 (55.9-71.0)			MST
			%F = 40.3%			NRI
Nishiyama, 2018^50	Retrospective/CS	Colorectal Disease/Surgical	N = 40	SF-BIA	Trained nutritionist	SGA
			age = 59.4 (12.3)
			%F = 52.5%
Norman, 2005^51	CS	General Hospital	N = 287	SF-BIA	Trained investigator	SGA
			age = 64.8 (18.9-96.6)	Anthropometric
			%F = 56.4%	Functional
Norman, 2008^52	CS	Gastrointestinal Disease	N = 242	SF-BIA	Operator not described	SGA
			age = 60.3 (42.1-68.3)	Anthropometric
			%F = 50%	Functional
Oey, 2020^53	Longitudinal/LOS	Liver Transplant Candidates	N = 102	CT	CT analyzed by research physicians. Other assessments performed by dietitians.	HSNAT Royal Free Hospital-nutritional prioritizing tool
			age = 56.3 (43.9-64.0)	Functional
			%F = 33.3%	MF-BIA
Pena, 2019^54	Longitudinal/LOS	Oncology/Surgical	N = 121	SF-BIA	Trained investigators	SGA
			age = 58.8 (12.5)
				Anthropometric
			%F = 47.1%
Pichard, 2004^55	Longitudinal/LOS	General Hospital	N = 952	SF-BIA	Hospital Nutrition Unit	SGA
			age = not clearly stated
			%F = 50.2%
Ramos da Silva, 2021^56	Post-hoc analysis/~7M	Breast Cancer	N = 61	BIS	Operator not described	NRI
			age = 46.4 (26-64)	Functional
			%F = 100%
Razzera, 2020^57	Prospective cohort/CS	Intensive Care Unit	N = 89	SF-BIA/BIVA	Operator not described	NUTRIC
			age = 62.5 (14.1)
			%F = 50.6%
Ribeiro, 2020^58	Longitudinal/24M	Liver Transplant Recipients	N = 29	SF-BIA	Trained investigators	HSNAT
			age = 54.1 (11.5)	Anthropometric
			%F = 20.7%	Functional
Rietveld, 2018^59	Longitudinal/LOS	Esophageal Cancer	N = 101	MF-BIA	Dietitians	ESPEN
			age = 65.3 (9.5)	Functional
			%F = 26.7%
Ringaitiene, 2016^60	Longitudinal/1M	Cardiac Surgery	N = 342	MF-BIA	Operator not described	HSNAT
			age = 65 (58-72)
			%F = 34.2%
Sanchez, 2019^61	Longitudinal/3M	Cervical Cancer	N = 55	Anthropometric	Anthropometric operator not described. CT conducted by radiologist.	PG-SGA
			age = 50.45 (11.4)	CT
			%F = 100%
Sanson, 2018^62	Longitudinal/12M	Malnourished Elderly	N = 81	SF-BIA	Operator not described	Instant Nutrition Assessment
			age = 80.7 (11.5)	Anthropometric
						NRS-2002
			%F = 54.3%
Spychalska-Zwolińska, 2020^63	CS	Lower Limb Ischemia	“N = 70	Anthropometric	Operator not described	NRS-2002 MNA
			age = 65.5 (11.2)	SF-BIA
			%F = 42.9%	Functional
Teixeira, 2021^64	CS	Chronic Obstructive Pulmonary Disease	N = 176	Anthropometric	Trained investigators	SGA
			age = 68.2 (10.4)
			%F = 56.2%
Tobberup, 2019^65	Longitudinal/3 treatment cycles	Non-Small Cell Lung Cancer	N = 52	CT	CT analyzed by radiologist and oncologist	PG-SGA-SF
			age = 67.5 (49-80)
			%F = 41.9%
Visser, 2013^66	Longitudinal/duration not described	Cardiac Surgery	N = 325	BIS	Operator not described	Sarcopenic obesity
			age = mostly ≥65	Functional
			%F = 27.7%
Zalizko, 2020^67	Prospective observational/Duration not stated	Inflammatory Bowel Disease	“N = 50	Anthropometric	Operator not described	NRS-2002 MUST
			age = 36.5 (28.5-51.5)
				MF-BIA
			%F = 44%

• Note: Sample characteristics: “ indicates that sample characteristics are based on the group with disease described in the “Population” column. Age is given in years as mean (SD) or median (range). “F” indicates the percentage of females in the given studies.
• Abbreviations: ASPEN MCC, American Society for Parenteral and Enteral Nutrition Malnutrition Consensus Criteria; BCA, body composition assessment; BIS, bioimpedance spectroscopy; BIVA, bioelectrical impedance vector analysis; CS, cross-sectional; CT, computed tomography; D, days; ESPEN, European Society for Clinical Nutrition and Metabolism; FSA, functional status assessment; GLIM, Global Leadership Initiative on Malnutrition; GNRI, geriatric nutrition risk index; HSNAT, hospital-specific nutritional assessment tool; LOS, length of stay; M, months; MF-BIA, multifrequency bioelectrical impedance analysis; MNA, Mini Nutritional Assessment; MUST, Malnutrition Universal Screening Tool; NA, nutrition assessment; NRI, Nutrition Risk Index; NRS-2002, Nutrition Risk Screening 2002; NUTRIC, nutrition risk in critically ill; PG-SGA, Patient-Generated Subjective Global Assessment; PG-SGA-SF, Patient-Generated Subjective Global Assessment Short Form; SF-BIA, single-frequency bioelectrical impedance analysis; SGA, Subjective Global Assessment; W, weeks.

Table 3. Overview of studies using body composition and/or functional status assessment tools to monitor response to nutrition interventions

Author, year	Study duration/design/intervention type	Population	Sample characteristics: N; age (mean [SD] or median [range]), years; % female	Body composition/functional assessment techniques	Assessment operator	Nutrition assessment comparator
Akita, 2019^68	5W/RCT/ONS	Pancreatic cancer	“N = 31 Age = 67.8 (10.7) %F = 64.5	MF-BIA CT	Operator not described	MUST
Aredes, 2019^69	45D/triple blind, placebo-controlled RCT/ONS	Cervical cancer	“N = 20 Age = 45.14 (9.67) %F = 100	CT	Trained investigator	PG-SGA
Bos, 2001^70	10D/prospective intervention trial/ONS	Malnourished older adult	N = 23 Age = 79 %F = 56.5	DXA SF-BIA Anthropometrics Functional	Operator not described	XXX
Bouillanne, 2013^71	6W/RT/pulse diet	Malnourished older adult	“N = 29 Age = 84.1 (2.3) %F = 79.3	DXA MF-BIA Functional	Operator not described	MNA
Breitkreutz, 2005^72	8W/RT/high-fat diet + ONS	Malnourished GI cancer	“N = 12 Age = 57.8 (1.3) %F = 25	SF-BIA	Operator not described	HSNAT
Brown, 2020^73	16W/DB PC RCT/ONS	Colorectal cancer	“N = 105 Age = 54.2 (46.8–65.3) %F = 36.0	CT	Trained investigators and radiologist	XXX
Caccialanza, 2015^74	7D/exploratory methodological/PNS	Hypophagic cancer	N = 30 Age = 63 (13.7) %F = 50	SF-BIA Functional	Nutrition and dietetics service in the clinical units	NRS-2002
Caccialanza, 2019^75	7D/clinical RT/PNS	Hypophagic cancer	N = 118 Age = 59.9 (14.7) %F = 34.4	SF-BIA Functional	Clinical nutrition and dietetics unit	NRS-2002
Casaer, 2013^76	9D/retrospective/early PNS	Traumatic brain injury	N = 15 Age not stated %F = 60	CT	CT evaluated by investigators, validated by expert radiologist	XXX
Creutzberg, 2003^77	8W/intervention trial/ONS	Chronic obstructive pulmonary disease	“N = 64 Age = 65 (9) %F = 23.4	SF-BIA Functional	Operator not described	XXX
De Benedetto, 2018^78	8W/double blind, placebo-controlled RCT/ONS	Chronic obstructive pulmonary disease	N = 45 Age = 73 (7) %F = 24.4	Functional SF-BIA	Operator not described	XXX
Della Valle, 2018^79	6M/prospective intervention trial/ENS	Head and neck cancer	N = 35 Age = 60 (55–67) %F = 42.9	SF-BIA	Dietitian	XXX
Federico, 2001^80	60D/case control/ONS	Gastrointestinal cancer	N = 60 Age = 55 (46–61) %F = 38.3	SF-BIA	Operator not described	XXX
Gonzalez-Granda 2019^81	~22D/prospective RT/ENS and PNS	Mechanically ventilated intensive care unit	N = 40 Age = ~57 %F = 40	MF-BIA	Operator not described	NUTRIC
Ha, 2010^82	3M/RCT/individualized nutrition support	Older adult stroke patients at nutrition risk	N = 124 Age = 79 %F = 51.6	BIS Anthropometrics	Trained investigators	MUST
Hoekstra, 2011^83	3M/controlled trial/multidisciplinary nutrition care	Older adult hip fracture	N = 127 Age = ~80 %F = 75.6	SF-BIA	Nurse (assumed)	MNA
Hyltander, 1991^84	10W/RCT/PNS	Testicular cancer	N = 33 Age = ~34 %F = 0	Anthropometrics	Operator not described	HSNAT
Inglis, 2021^85	24W/retrospective double blind RCT/ONS	Prostate cancer	N = 59 Age = 67.6 (5.4 SD) %F = 0.0	SF-BIA Functional	Operator not described	XXX
Ishikawa, 2016^86	4W/open-label RCT/ONS	Esophageal cancer	N = 33 Age = ~67 %F = 18.2	MF-BIA	Operator not described	Asian Working Group for Sarcopenia criteria
Kim, 2019^87	8W/RCT/ONS	Pancreatic cancer	N = 34 Age = 65.2 %F = 52.9	MF-BIA	Dietitian	PG-SGA
Krüger, 2016^88	≥3D/RCT/PNS	Pancreatic cancer	N = 100 Age = 64.9 %F = 43	MF-BIA	Operator not described	NRS-2002
Löser, 2021^89	Varied duration/prospective RCT/intensive nutrition counselling	Head and neck squamous cell carcinoma	N = 61 Age = 63 (20–89) %F = 27.9	SF-BIA	Physicians and dietitians	MUST NRS-2002
Malafarina, 2017^90	XXX/multicenter RCT/ONS	Geriatric	N = 92 Age = 85.4 (6.3) %F = 73.8	SF-BIA Functional	Functional assessments performed by physical therapist; SF-BIA operator not described	MNA-SF
Mazzuca, 2019^91	6M/single blinded multicenter PC RCT/ONS	Colorectal cancer	N = 47 Age = ~67 %F = 38.3	SF-BIA CT	SF-BIA performed by dietitian; CT analyst not described	MUST MNA
McCurdy, 2019^92	6–8W/prospective cohort/intake recommendations	Head and neck cancer	N = 41 Age = 57.8 (10.8) %F = 22	CT	Operator not described	ESPEN guidelines for dietary intake
Mohamed, 2019^93	~107D/retrospective cohort/ENS	Esophageal cancer	N = 15 Age = 61.3 (12.8 SD) %F = 27	CT	Operator not described	XXX
Murphy, 2011^94	≥6W/open-label contemporaneous control/ONS	Non–small cell lung cancer	N = 40 Age not stated %F = 47.5	CT	Operator not described	PG-SGA
Nakamura, 2020^95	10D/prospective RCT/ENS	Intensive care unit	N = 50 Age = 71.8 (12.4 SD) %F = 42.3	CT	CT analyzed by radiology technician	XXX
Nakamura, 2021^96	10D/prospective RCT/high protein	Intensive care unit	N = 117 Age = 68.3 (14.3 SD) %F = 41.7	CT	CT analyzed by radiology technician	HSNAT
Nakano, 2021^97	10D/historical control intervention trial/ENS	Intensive care unit	N = 101 Age = 70.9 (14.5) %F = 30.4	CT	CT analyzed by radiology technician	MUST
Norman, 2006^98	8W/double blind, placebo-controlled RCT/ONS	Colorectal cancer	N = 30 Age = ~63 %F = 35.1	SF-BIA Anthropometrics Functional	Operator not described	XXX
Obling, 2018^99	24W/open-label RCT/HPN	Gastrointestinal cancer	N = 47 Age = 66.9 %F = 36	MF-BIA Functional Anthropometrics	Dietitian and primary investigator	NRS-2002
Olveira, 2016^100	24W/parallel treatment RCT/ONS	Bronchiectasis	N = 30 Age = 56.1 (13) %F = 60	DXA Anthropometrics MF-BIA Functional	Operator not described	XXX
Pelzer, 2010^101	18W/cohort trial/additional PNS	Pancreatic cancer	N = 32 Age = 62 (47–75 range) %F = 43.75	SF-BIA	Operator not described	XXX
Phang, 1996^102	7–21D/prospective intervention trial/ENS or PNS	Mechanically ventilated intensive care unit	N = 54 Age = ~58 (17) %F = 40	SF-BIA	Operator not described	SGA
Rigaud, 2010^103	2M/retrospective/low-sodium diet	Anorexia nervosa refeeding	N = 218 Age = 22.1 (4.2) %F = 98	MF-BIA Anthropometrics	Dietitian (assumed)	XXX
Rimini, 2021^104	3M/prospective clinical trial/monitoring	Advanced gastric cancer	N = 40 Age mostly over 70 %F = 40	CT	Operator not described	XXX
Rinninella, 2021^105	Varied duration/retrospective clinical trial/NCP	Older adult colorectal surgery	N = 302 Age = 80.5 (4.1 SD) %F = 49	SF-BIA	Dietitian	NRS-2002
Ritch, 2019^106	2M/pilot RCT/ONS	Urological cancer	N = 31 Age = 69 %F = 26	CT DXA	Certified densitometrist for DXA; CT analyst not described	Sarcopenia
Ryan, 2009^107	26D/double blind RCT/ENS	Esophageal cancer/surgical	N = 53 Age = ~63.5 %F = 9.4	MF-BIA	Dietitian	XXX
Shirai, 2017^108	XXX/retrospective/ONS	Gastrointestinal cancer	N = 37 Age = 72.3 (8.4) %F = 29.7	MF-BIA	Operator not described	≥5% pre-illness body weight loss
van der Werf, 2020^109	Varied duration/single blind RCT/individualized NCP	Colorectal cancer	N = 105 Age = 64.0 (13.0 SD) %F = 30.0	CT Functional	CT analysis performed by trained investigator	XXX
Wan, 2020^110	~10M/retrospective interventional/ENS	Superior mesenteric artery syndrome	N = 26 Age = 25.0 (11.8 SD) %F = 61.5	MF-BIA	Operator not described	SGA HSNAT
Wang, 2002^111	10D/prospective observational/early ENS or early PNS	Enterocutaneous fistula	N = 61 Age = 41.9(±13.5) %F = 12.4	MF-BIA	Operator not described	XXX
Willemsen, 2020^112	~32 M/prospective cohort/ENS	Head and neck squamous cell carcinoma	N = 137 Age = 59.0 (8.0 SD) %F = 32	SF-BIA Functional	Dietitian, medical oncologist, and radiation oncologist	Dutch guidelines
Yeh, 2018^113	14D/retrospective/ENS + PNS	Intensive care unit/surgical	N = 140 Age = 60.1 (17.4) %F = 51	CT	Operator not described	HSNAT
Zhuang, 2020^114	~30D/prospective cohort/intake recommendations	Head and neck cancer	N = 287 Age = 52.7 (14.1) %F = 34.8	MF-BIA	Trained investigators	ESPEN guidelines for dietary intake

• Note: Sample characteristics: “indicates that sample characteristics are based on the intervention group.
• Abbreviations: ASPEN MCC, American Society for Parenteral and Enteral Nutrition Malnutrition Consensus Criteria; BCA, body composition assessment; BIS, bioimpedance spectroscopy; BIVA, bioelectrical impedance vector analysis; CS, cross-sectional; CT, computed tomography; D, days; ENS, enteral nutrition supplement; FSA, functional status assessment; GLIM, Global Leadership Initiative on Malnutrition; GNRI, Geriatric Nutrition Risk Index; HPN, home parenteral nutrition; HSNAT, hospital-specific nutrition assessment tool; LOS, length of stay; M, months; MF-BIA, multifrequency bioelectrical impedance analysis; MNA, Mini Nutritional Assessment; MNA-SF, Mini Nutritional Assessment Short Form; MUST, Malnutrition Universal Screening Tool; NA, nutrition assessment; NCP, nutrition care plan; NRI, Nutrition Risk Index; NRS-2002, Nutrition Risk Screening 2002; NUTRIC, nutrition risk in critically ill; ONS, oral nutrition supplement; PG-SGA, Patient-Generated Subjective Global Assessment; PG-SGA-SF, Patient-Generated Subjective Global Assessment Short Form; PNS, parenteral nutrition supplement; RCT, randomized controlled trial; RT; randomized trial; SF-BIA, single-frequency bioelectrical impedance analysis; SGA, Subjective Global Assessment; W, weeks; XXX, no NA comparator (or not stated by authors).

As can be seen in Table 2, of the 47 studies that evaluated NA, 19 used the SGA,^{24, 26, 28, 31, 32, 37, 38, 40, 41, 43, 47, 50-52, 54, 55, 61, 64, 65} three used the EDC,^{25, 28, 45} one used the ASPEN/AND MCC,⁴⁷ and one used the GLIM criteria³⁵ as a reference standard.

As can be seen in Table 3, of the 47 studies that evaluated RNI, five used the SGA,^{69, 87, 94, 102, 110} whereas the ASPEN/AND MCC, EDC, and GLIM criteria were not used at all in these studies.

Overall, bioimpedance was the most commonly used bedside technique. The most frequently employed protocol for this form of assessment was to have patients lay supine with limbs abducted. The two main BCA parameters that were assessed along with NA methods were phase angle (PhA) at 50 kHz and fat-free mass (FFM) estimated through a variety of bioimpedance approaches (Tables S3 and S7). A total of 18 studies included the SGA as the NA tool along with bioimpedance techniques; associations were evaluated in all but four. Of these 14 studies, seven used PhA/standardized PhA (SPhA) and seven used FFM/the FFM index (FFMI) by a variety of bioimpedance approaches. PhA/SPhA differentiated between patients with and without the SGA defined malnutrition in two studies,^{38, 50} was associated with the SGA class in four studies,^{41, 47, 52, 54} and independently predicted SGA malnutrition in another.³⁷ PhA also differentiated between patients with and without malnutrition using the ASPEN MCC.⁴⁷ FFM estimated by a single-frequency bioelectrical impedance analysis (SF-BIA) equation was able to differentiate between patients with and without SGA-identified malnutrition in one study,⁴⁰ and FFMI (using FFM from the same equation) showed an association with SGA-identified malnutrition in another report.⁵⁵ For more information, see Table S3.

The second most frequently used bedside assessment technique was anthropometry. The main anthropometric assessments used for NA were mid-upper arm circumference (MUAC), arm muscle area (AMA), and calf circumference (CC; Tables S1 and S5). CC was able to differentiate between patients with and without SGA-identified malnutrition, and low CC was also associated with an increased risk of malnutrition by the SGA in the same study.⁶⁴ In another study, CC was positively correlated with the EDC.⁴⁵ MUAC was significantly different between patients with and without SGA-identified malnutrition in one study,³⁸ and showed no association with the SGA in another.⁵² Finally, AMA was shown to decrease with the SGA rating in the one study that evaluated it statistically.⁵² For study details, see Table S1.

A number of studies that we reviewed included computed tomography (CT) and/or dual-energy x-ray absorptiometry (DXA) as the reference BCA technique; many of these utilized the assessment of skeletal muscle cross-sectional area at the level of L3 (CTL3; Tables S2 and S6). CTL3 showed fair agreement with the SGA⁴³ and was able to differentiate between patients with and without GLIM-identified malnutrition.³⁵ For study details, see Table S2.

Finally, almost every study that included an assessment of functional status utilized handgrip strength (HGS). Many of the protocols called for measurements using hydraulic hand dynamometers, with the patient seated, using the dominant arm bent at 90 degrees, measured multiple times for an average value (Tables S4 and S8). HGS was able to distinguish between patients with and without malnutrition identified by the GLIM criteria³⁵ and the SGA.^{51, 52} On the other hand, HGS was not associated with a risk of malnutrition when assessed by the EDC.⁵⁹ For study details, see Table S4.

DISCUSSION

We conducted a comprehensive review on the bedside assessment of nutrition status through the use of BCA and FSA tools. In the studies we reviewed, the most common bedside BCA and FSA tools used in clinical NA, along with the four most widely accepted sets of diagnostic criteria (Table 1), are (1) bioimpedance, specifically 50 kHz PhA/SPhA and FFM/FFMI (by the Kyle equation);¹¹⁵ (2) anthropometric methods, specifically MUAC, AMA, and CC; (3) CT assessment of CTL3; and (4) HGS. The individuals conducting these assessments were not consistently identified in the reviewed articles. From the information available, these assessments were performed primarily by trained investigators and dietitians/nutritionists from the hospital nutrition unit; however, some measures were performed by nurses, physicians, radiologists, and physical therapists (see Tables 2 and 3). Because of the necessary focus on research studies in our systematic review, we have gained perspective on the tools used in clinical research settings, however, we are unable to draw definitive conclusions regarding the incorporation of BCAs and FSAs by clinicians in daily practice. The formation of a global consensus on criteria for the diagnosis of malnutrition is an evolving process, with the most recent development being the proposition of the GLIM criteria. The top five criteria put forth by the GLIM consortium include nonvolitional weight loss, low body mass index, reduced muscle mass, reduced food intake/assimilation, and disease burden/inflammation. Within the GLIM framework, there is strong evidence to support reduced muscle mass as a phenotypic indicator of malnutrition; however, no consensus exists on how best to measure it clinically.¹¹⁶ A number of validated BCA methods for assessing muscle are identified by GLIM for clinical use, although they recognize that there is limited availability of these tools in most settings globally.¹¹⁶ Anthropometry and physical examination are identified as acceptable alternatives.¹¹⁶ FSA tools (eg, HGS) are recommended as supportive measures.¹¹⁶ A majority of the studies we reviewed utilized bioimpedance and anthropometric methods to assess muscle mass in relation to nutrition status. For the most part, these assessment methods consistently showed associations with the four most widely accepted NA diagnostic frameworks. Although these were the most commonly reported assessment methods, our research team is not advocating the use of any particular technique over another.

In a recent guidance document from GLIM, a call to action was issued for the validation of the key criteria for diagnosing adult malnutrition.¹¹⁷ Criterion validity was identified as a preferred form of validation that involves comparing the GLIM criteria against a reference standard for malnutrition (eg, SGA). A number of the studies in this systematic review compared a BCA and/or FSA measure against various accepted NA diagnostic frameworks. From our review, it appears that there is some evidence, however limited, supporting the concurrent validity of PhA, FFM by single-frequency bioimpedance analysis using the Kyle¹¹⁵ equation, CC, and HGS against the SGA and other traditionally accepted NA diagnostic frameworks.

This review has shed light on the BCA and/or FSA tools used for NA in clinical research settings. The assessment of body composition and functional status in daily inpatient clinical practice is challenging in light of numerous physical and physiologic factors associated with acute and chronic disease and treatments that could adversely impact measurements, for example, altered fluid status, wounds, fever, venous fluid, and nutrition administration. In addition, it is important to note that even if all of these confounding factors were controlled for, differences in the device used for assessment may produce different results, especially with bioimpedance, as the technology and algorithms for estimating body composition vary between devices.¹⁹ Because we found a wide variety of devices being used for clinical BCA, we have provided a brief description of the devices utilized in each study in the respective Supporting Information tables (Tables S1–S8). Furthermore, each assessment technique has its own strengths, limitations, and underlying assumptions that must be met to obtain valid measurements in a particular situation.¹¹⁷ For example, although FSA tools, such as HGS, are recommended as a supplementary assessment for the identification of malnutrition,¹¹⁶ HGS is a patient-motivated measure,¹¹⁸ so if the patient is not putting forth their greatest effort during the assessment, the measurement will produce results that could potentially indicate nutrition risk when the patient is actually well nourished. Ultimately, the effective use of reduced muscle mass as a core GLIM criterion will require clinicians to appropriately select validated bedside assessment tools, to apply consensus-based measurement protocols, and to interpret muscle measures using suitable reference data. To facilitate this vision, additional research is needed with available BCA technologies in clinical settings. The GLIM consortium has identified DXA, CT, magnetic resonance imaging, bioimpedance methods, and ultrasound as the most likely methods to assess reduced muscle mass for the diagnosis of malnutrition.^{119, 120} Of these, DXA, CT, and magnetic resonance imaging are considered by many to be reference BCA techniques and are not often available for bedside use by clinicians. However, there is growing interest by clinicians in the evaluation of CT scans to derive muscle measures, both in research and in clinical assessment, and CTL3 is likely to be the best reference parameter to use for validation of bedside techniques, at least in select populations for whom CT scans are conducted as part of diagnostic care and where reference cut-points have been established, including liver failure,¹²¹ critical illness,¹²² and cancer.^{15, 123, 124}

There are some limitations to this systematic review. First, although every effort was made to capture a comprehensive view of the literature involving BCA and FSA tools specifically being used in clinical inpatient settings to either assess nutrition status against an NA standard or to evaluate RNI, it is possible that important papers were missed. In addition, in some instances during the process of screening papers for inclusion, there were some uncertainties regarding the research setting (ie, inpatient vs outpatient); in these instances, reviewers came to consensus using best judgment. Furthermore, it is beyond the scope of this review to discuss validity, reliability, technical, and physiologic factors (eg, hydration status) influencing BCA and FSA measures in clinical populations. These important factors should be considered when selecting methods and interpreting results in daily practice and have been reviewed elsewhere.^{19, 117, 125-129} Finally, given that our aim was to describe the use of BCA and FSA tools in inpatient settings as they relate to NA or monitoring RNI, a wide range of study designs, populations, and outcome variables were incorporated into our review. Although this gave us a broad perspective on the utilization of these various tools in clinical populations, this wide-ranging approach made it difficult to concisely summarize studies.

Although bioimpedance techniques are the most widely applied bedside tools, there is growing interest in ultrasound for the clinical assessment of muscle. How well these tools can facilitate the diagnosis of malnutrition remains to be firmly established. To move us forward in this regard, clinical nutrition intervention studies should ideally assess muscle using multiple bedside tools, such as bioimpedance and ultrasound techniques, anthropometry, and FSA, such as HGS, in addition to at least one reference BCA and one of the four traditionally accepted NA diagnostic frameworks, in order to better understand the concurrent validity of these bedside methods. Furthermore, collaboration among disciplines is indispensable for a comprehensive assessment of nutrition status. For example, it is not likely to be feasible for dietitians to incorporate skills such as CT image analysis into their daily practice because of time constraints. Engaging radiology departments to provide muscle measures from CT scans by physician order is one potential opportunity to standardize NA in groups in which CT is an option. The inclusion of clinical outcome data would also allow us to assess the predictive validity of these measures. Finally, incorporating repeated measures by a particular bedside method in clinical intervention studies would facilitate the analysis of interobserver and intraobserver variation, and thus provide invaluable information on the precision and sensitivity of the method to detect changes in response to an intervention.

In a recent publication, the GLIM workgroup identified key areas for future research, including the need to come to consensus on optimal protocols for measurement and interpretation of muscle mass and function.¹²⁰ They further called out the need to identify appropriate reference cut-points that can be used to interpret a single measurement to make the diagnosis of malnutrition through the GLIM criteria.¹¹⁹ A recent editorial by Gonzales¹²⁹ highlights the need for clinicians using bioimpedance techniques to appreciate the importance of selecting reference cut-points that are appropriate to the device/algorithm and population that best matches the patient being assessed. This is an important consideration for other BCA methods as well, and additional research is needed on appropriate reference data for the clinical application of all techniques. The reader is directed to a number of excellent guidance documents and reviews for a full discussion of considerations.^{19, 117, 129-132}

CONCLUSION

It is widely accepted that muscle is a vital component of nutrition status, and it is recognized as a core phenotypic criterion in all four of the major NA diagnostic frameworks. Our focus in conducting this review was to survey the literature to gain perspective on the ways BCA and FSA tools are being used to evaluate muscularity as a component of nutrition status in clinical adult inpatient settings. Bioimpedance methods are the most widely used bedside BCA tools, and HGS is the most widely used FSA tool; however, these methods are being used with a variety of protocols, algorithms, and interpretation practices in highly heterogeneous populations. For all bedside techniques, there is a need for validation studies, the development of globally standardized assessment protocols, and the interpretation of measures against suitable reference data in clinical inpatient settings. Furthermore, longitudinal research studies should incorporate BCA and FSA tools using standardized protocols, as well as reference techniques for muscle assessment, to capture important changes in nutrition status. An interdisciplinary collaborative approach is likely to be essential to make advancements in these areas in both research and clinical practice. With progress as an international professional community, we can close in on a more globally standardized process for the assessment of nutrition status.

AUTHOR CONTRIBUTIONS

Gerdien C. Ligthart-Melis, Kirsten A. Berk, Joanne F. Olieman, and Carrie P. Earthman contributed to the conception/design of the research. All authors contributed to the acquisition, analysis, or interpretation of the data. Luke O. Smith drafted the manuscript. All authors critically revised the manuscript, agree to be fully accountable for ensuring the integrity and accuracy of the work, and read and approved the final manuscript.

ACKNOWLEDGMENTS

The authors wish to thank Wichor Bramer, PhD and Marjolein Udo, MSc from the Erasmus MC Medical Library for developing and updating the search strategies and Elles van der Louw, PhD, RD for her contributions to the design and final review of the manuscript.

CONFLICT OF INTEREST

None declared.