Original Investigation

Novel Equations to Estimate Lean Body Mass in MaintenanceHemodialysis Patients

Nazanin Noori, MD, PhD,1 Csaba P. Kovesdy, MD,2 Rachelle Bross, RD, PhD,1,3,4

Martin Lee, PhD,5 Antigone Oreopoulos, MSc, PhD,6 Deborah Benner, RD,7

Rajnish Mehrotra, MD,1,9 Joel D. Kopple, MD,4,5,9 andKamyar Kalantar-Zadeh, MD, MPH, PhD1,5,8,9

Background: Lean body mass (LBM) is an important nutritional measure representing muscle mass andsomatic protein in hemodialysis patients, for whom we developed and tested equations to estimate LBM.

Study Design: A study of diagnostic test accuracy.Setting & Participants: The development cohort included 118 hemodialysis patients with LBM measured

using dual-energy x-ray absorptiometry (DEXA) and near-infrared (NIR) interactance. The validation cohortincluded 612 additional hemodialysis patients with LBM measured using a portable NIR interactance techniqueduring hemodialysis.

Index Tests: 3-month averaged serum concentrations of creatinine, albumin, and prealbumin; normalizedprotein nitrogen appearance; midarm muscle circumference (MAMC); handgrip strength; and subjective globalassessment of nutrition.

Reference Test: LBM measured using DEXA in the development cohort and NIR interactance in validationcohorts.

Results: In the development cohort, DEXA and NIR interactance correlated strongly (r � 0.94, P � 0.001).DEXA-measured LBM correlated with serum creatinine level, MAMC, and handgrip strength, but not withother nutritional markers. Three regression equations to estimate DEXA-measured LBM were developedbased on each of these 3 surrogates and sex, height, weight, and age (and urea reduction ratio for theserum creatinine regression). In the validation cohort, the validity of the equations was tested against theNIR interactance–measured LBM. The equation estimates correlated well with NIR interactance–measured LBM (R2 � 0.88), although in higher LBM ranges, they tended to underestimate it. Median (95%confidence interval) differences and interquartile range for differences between equation estimates andNIR interactance–measured LBM were 3.4 (�3.2 to 12.0) and 3.0 (1.1-5.1) kg for serum creatinine and 4.0(�2.6 to 13.6) and 3.7 (1.3-6.0) kg for MAMC, respectively.

Limitations: DEXA measurements were obtained on a nondialysis day, whereas NIR interactance wasperformed during hemodialysis treatment, with the likelihood of confounding by volume status variations.

Conclusions: Compared with reference measures of LBM, equations using serum creatinine level, MAMC, orhandgrip strength and demographic variables can estimate LBM accurately in long-term hemodialysis patients.Am J Kidney Dis. 57(1):130-139. © 2010 by the National Kidney Foundation, Inc.

INDEX WORDS: Hemodialysis; protein-energy wasting; lean body mass (LBM); serum creatinine; midarmmuscle circumference (MAMC); handgrip strength; near-infrared (NIR); bioelectrical impedance analysis (BIA);nutritional status.

Accurate assessment of nutritional status and bodycomposition in individuals with chronic kidney

disease, including long-term hemodialysis patients, iscrucial because malnutrition and wasting syndromesare among the strongest risk factors for morbidity andmortality.1-3 The International Society for Renal Nutri-tion and Metabolism (ISRNM) Expert Panel, which

From the 1Harold Simmons Center for Kidney Disease Researchand Epidemiology, Los Angeles Biomedical Research Institute atHarbor-UCLA Medical Center, Torrance, CA; 2Salem Veterans Af-fairs Medical Center, Salem, VA; 3Division of Bionutrition and 4Gen-eral Clinical Research Center, Los Angeles Biomedical Research Instituteat Harbor-UCLA Medical Center, Torrance; 5UCLA School of PublicHealth, Los Angeles, CA; 6University of Alberta, Edmonton,Canada; 7DaVita Nutrition, Irvine; 8Department of Pediatrics,Los Angeles Biomedical Research Institute at Harbor-UCLAMedical Center, Torrance; and 9David Geffen School of Medi-

cine at UCLA, Los Angeles, CA.

130

recently has proposed the term protein-energy wast-ing in lieu of other terms for uremic malnutrition inpatients with chronic kidney disease, defines protein-energy wasting as “loss of body protein mass and fuelreserves.”4 Decreased lean body mass (LBM) andmuscle mass are the main components of protein-energy wasting.5 Hence, accurate assessment of body

Received January 10, 2010. Accepted in revised form October11, 2010.

Address correspondence Kamyar Kalantar-Zadeh, MD, MPH, PhD,Harold Simmons Center for Chronic Disease Research and Epidemi-ology, Los Angeles Biomedical Research Institute at Harbor-UCLAMedical Center, and UCLA David Geffen School of Medicine andUCLA School of Public Health, 1124 W Carson St, C1-Annex,Torrance, CA 90502. E-mail: kamkal@ucla.edu

© 2010 by the National Kidney Foundation, Inc.0272-6386/$36.00

doi:10.1053/j.ajkd.2010.10.003

Am J Kidney Dis. 2011;57(1):130-139

LBM Equations in CKD

composition, including LBM, is the key to reliableevaluation of nutritional status in patients with chronickidney disease. Nevertheless, the optimal method fordetermining LBM in these patients is debatable. Al-though dual-energy x-ray absorptiometry (DEXA) isconsidered a reference method for assessing bodycomposition,6-8 very few dialysis clinics have directaccess to DEXA machines. Therefore, developing andtesting equations that can estimate LBM based onroutinely available clinical and nutritional measuresthat correlate with LBM is of paramount clinicalrelevance.9

Several previous studies, especially in peritonealdialysis patients, have examined the association ofLBM with other nutritional measures, including sev-eral anthropometric and biochemical values, subjec-tive global assessment, and normalized protein nitro-gen appearance (nPNA), also known as normalizedprotein catabolic rate (nPCR).10,11 Keshaviah et al12

found that LBM correlated with serum albumin, se-rum creatinine (SCr), and nPNA levels in peritonealdialysis patients, whereas Szeto et al13 and Heim-burger et al14 did not find good correlations of LBMwith nutritional indexes, including serum albuminlevel. To our knowledge, no study has been conductedto examine the validity of anthropometric or othernutritional measurements in predicting LBM in hemo-dialysis patients.15,16

In the present study, we examined the correlation ofDEXA-measured LBM with a number of nutritionalmarkers, including serum concentrations of creati-nine, albumin, and prealbumin, and anthropometricmeasurements, including midarm muscle circumfer-ence (MAMC), handgrip strength, nPNA, and subjec-tive global assessment in 118 randomly selected hemo-dialysis patients. In this so-called development cohort,we developed equations to estimate LBM based onthese measures and compared their consistency withDEXA-measured LBM as the reference standard.Subsequently, we tested the validity of the createdregression equations in a validation cohort of 612additional hemodialysis patients in whom LBM wasestimated using the portable near-infrared (NIR) inter-actance technique.

METHODS

Patient Population

We studied hemodialysis patients who participated in the Nutri-tional and Inflammatory Evaluation in Dialysis (NIED) Study.17

The original patient cohort was derived over 5 years from a pool ofmore than 3,000 hemodialysis outpatients in 8 DaVita maintenancedialysis facilities in the South Bay Los Angeles area (see the NIEDStudy website at www.NIEDStudy.org for more details).3,5,18-23

Included were outpatients who had been undergoing hemodialysistreatment for at least 8 weeks, were 18 years or older, and signed

the institutional review board–approved consent form. Participants

Am J Kidney Dis. 2011;57(1):130-139

with acute infections or an anticipated life expectancy less than 6months (eg, because of metastatic malignancy or advanced humanimmunodeficiency [HIV]/AIDS disease) were excluded.

From October 1, 2001, through December 31, 2006, a total of893 randomly invited hemodialysis patients from 8 DaVita dialysisclinics in the Los Angeles South Bay area signed the informedconsent form. Patients for whom the upper arm did not appearappropriate for midarm muscle measurements were excluded. In730 remaining patients, body composition was assessed using theportable NIR interactance technique, and triceps skinfold thick-ness and MAMC were measured in the dialysis clinic. One ofevery 5 of these patients also was invited randomly to come toHarbor-UCLA General Clinical Research Center during a nonhe-modialysis day to undergo additional tests, including DEXA andother body composition measures; 118 patients agreed and did so.This group was called the development cohort. The remainingpatients who did not attend additional testing at the GeneralClinical Research Center and thus underwent NIR interactance–,but not DEXA-measured LBM assessment (n � 612), were calledthe validation cohort (Fig 1). All participants refrained from eatingand drinking for at least 4 hours before the tests and did notconsume alcohol or exercise for 24 hours before testing.

AnthropometricMeasures

Participants were weighed while wearing a hospital gown with nofootwear. Body weight was measured to the nearest 0.1 kg on a GSEdigital platform scale, model 350 (GSE Scale Systems, www.gse-inc.com). Height was measured to the nearest 0.1 cm using a wall-mounted stadiometer (model S100; Ayrton Corp) with participantsstanding erect and arms hanging freely at their sides. Lange calipers(Cambridge Scientific Instruments, www.cambridgescientific.com)were used to measure triceps skinfold thickness.24 Triplicate measure-ments were obtained from the non–dialysis vascular access arm.Triceps skinfold thickness was used as an index test to estimate bodyfat.25,26 MAMC was calculated using a previously described equa-tion,27 by which triceps skinfold thickness multiplied by 3.142 wassubtracted from midarm circumference; all measurements were incentimeters.

After determining the participant’s hand dominance, dominanthandgrip strength was measured in the development cohort whilethey were seated with shoulder adducted and neutrally rotated,elbow flexed at 90°, forearm in neutral position, and the handle ofthe dynamometer adjusted at the second handle position, and thenasking the participant to hold the handle and squeeze as hard as heor she could.

Dual-EnergyX-rayAbsorptiometry

The reference test for assessment of body composition wasDEXA, performed using a Hologic Series Delphi-A Fan BeamX-ray Bone Densitometer with software version 12.4 (Hologic Inc,www.hologic.com).28 Measurements were performed as previ-ously described,6-8 with participants wearing a hospital gown withno metal snaps and all artifacts removed. Scans were analyzed todetermine lean mass, fat mass, bone mineral content, and total-body fluid percentage. The precision of body composition analysiswas determined by means of weekly quality control assessmentsusing a whole-body phantom and tissue calibration step phantomcomposed of soft tissue– and lean tissue–equivalent materials.6-8

NIR Interactance

In both the development (n � 118) and validation (n � 612)cohorts, portable NIR interactance technology was used in the 8participating dialysis clinics to estimate LBM. A commercial NIRinteractance sensor with a coefficient of variation of 0.5% for

total-body fat measurements (portable Futrex 6100; www.futrex.

131

dy m

Noori et al

com) was used. NIR interactance measurements were performedby placing a Futrex sensor on the non–vascular access upper armfor several seconds and entering the required data (date of birth,sex, weight, and height) from each patient. NIR interactancemeasurements of fat mass have correlated significantly with DEXA-measured fat mass in hemodialysis patients.29

Laboratory Tests

Predialysis blood samples and postdialysis serum urea nitrogensamples were obtained on a midweek day that coincided chrono-logically with the drawing of quarterly blood tests in DaVitafacilities. Urea reduction ratio and single-pool Kt/V were used torepresent the administered dialysis treatment dose.30 All routinelaboratory measurements were performed by DaVita Laboratories(www.davita.com) using automated methods. To decrease intrain-dividual variation, 3-month averaged values for laboratory mea-sures and urea reduction ratio during the study calendar quarterwere calculated and used in this study.

StatisticalMethods

Stepwise procedures were performed to select potential vari-ables for the regression equations. Pearson correlation coefficientsbetween DEXA-measured LBM (the reference test) and otherrelevant measures were examined first in the development cohortafter adjusting for case-mix. Case-mix variables included age, sex,race/ethnicity, diabetes, dialysis vintage, insurance (Medicare vsothers), marital status, modified Charlson comorbidity score, dialy-sis dose, and residual kidney function. We created 3 equations tocalculate LBM in the development cohort. To examine differencesbetween LBM estimated using our equations and NIR interactance–measured LBM in the validation cohort, we used both differenceplots with Pearson correlation tests, which is a graphical-statisticalapproach based on Bland-Altman analysis for comparison of afield method with a reference standard,31 and conventional Bland-Altman plots with Pitman test for trend.32 Unless otherwise stated,results are summarized as mean � standard deviation. Statisticalanalyses were carried out using Stata statistical software, version

Figure 1. Flow diagram of the development and validation cohandgrip strength; IRB, institutional review board; LBM, lean boSCr, serum creatinine.

10.0 (Stata Corp, www.stata.com).

132

RESULTS

Table 1 lists general characteristics of the studypopulation in both the development (n � 118) andvalidation cohorts (n � 612). Patients in the valida-tion cohort were slightly older and included fewermen and African Americans, but more Hispanicscompared with the development cohort. Mean vintagetime and triglyceride level were higher in the develop-ment and validation cohorts, respectively.

Table 2 lists both regression coefficients and corre-lation coefficients of DEXA-measured LBM withrelevant nutritional measures in the substudy of 118patients in the development cohort using linear regres-sion equations with LBM as the outcome. Model 1 isbased on unadjusted (Pearson) correlations of eachmeasure separately, model 2 includes all 7 surrogatesin the model at the same time, and model 3 alsoincludes case-mix variables.

We used multiple linear regression analyses withleast squares methods to develop the most parsimoni-ous equations to predict LBM. Stepwise proceduresled to the selection of 3 demographic variables (weight,height, and sex), MAMC, SCr level, and handgripstrength. Hence, we created 3 equations (Box 1) usingeach of these 3 variables separately in combinationwith the selected demographic variables of sex, height,and weight. For the SCr-based equation, we alsoincluded urea reduction ratio because SCr level maybe affected by dose of hemodialysis treatment.

Figure 2 shows the distribution of LBM valuesusing 4 different methods, that is, measured directly

. Abbreviations: DEXA, dual-energy x-ray absorptiometry; HGS,ass; MAMC, midarm muscle circumference; NIR, near infrared;

horts

using DEXA and estimated using each of the 3 de-

Am J Kidney Dis. 2011;57(1):130-139

LBM Equations in CKD

scribed regression equations in 118 hemodialysis pa-tients in the development cohort. As shown in Fig 2,the SCr-, MAMC-, and handgrip strength–based esti-mates of LBM showed similar mean values and varia-tions compared with each other and also with thedirect assessment of LBM using DEXA. NIR interac-tance–measured LBM correlated closely with DEXA-measured LBM in the development cohort (r � 0.94;P � 0.001). Therefore, NIR interactance–measuredLBM was used in the validation cohort as the refer-ence standard.

Tables 3 and 4 list comparisons of the performance

Table 1. Demographic and Clinical Characteristics ofHemodialysis Patients in the Development and Validation

Cohorts on Body Composition Measurement

DevelopmentCohort (n � 118)

ValidationCohort (n � 612)

Age (y) 49 � 11 54 � 15

Men (%) 57 53

Diabetes (%) 52 53

Race/ethnicity (%)African American 40 30Hispanic 38 53

Weight (kg) 74.5 � 18.4 72.3 � 19.0

Height (inch) 65.3 � 4.1 65.1 � 4.3

Body mass index (kg/m2) 27.0 � 6.0 26.6 � 6.2

Lean body mass (kg)By DEXA 49.8 � 9.9 NABy NIR interactance 55.3 � 10.5 52.2 � 11.6

Dialysis vintage (mo) 41.1 � 32.9 30.7 � 33.7

Dialysis dose (Kt/V) 1.7 � 0.3 1.6 � 0.3

nPNA or nPCR (g/kg/d) 1.11 � 0.22 1.06 � 0.24

Laboratory measurementsBlood hemoglobin (g/dL) 12.2 � 0.7 12.0 � 1.0Serum albumin (g/dL) 4.0 � 0.3 3.9 � 0.4Serum creatinine (mg/dL) 10.8 � 3.0 10.1 � 3.3Prealbumin

(transthyretin) (mg/dL)30.6 � 9.6 28.1 � 9.6

Total iron-bindingcapacity (mg/dL)

210.9 � 35.3 206.6 � 40.0

Total cholesterol (mg/dL) 147.2 � 41.1 150.1 � 42.3LDL cholesterol (mg/dL) 80.9 � 28.9 82.6 � 34.5HDL cholesterol (mg/dL) 36.1 � 13.4 35.2 � 13.6Triglycerides (mg/dL) 148.2 � 122.6 163.2 � 105.7Serum urea nitrogen

(mg/dL)63.0 � 16.2 63.3 � 15.1

Note: Values are presented as mean � standard deviationor percentage. Conversion factors for units: hemoglobin andalbumin in g/dL to g/L, �10; serum creatinine in mg/dL to�mol/L, �88.4; total, LDL, and HDL cholesterol in mg/dL tommol/L, �0.02586; triglycerides in mg/dL to mmol/L, �0.01129;serum urea nitrogen in mg/dL to mmol/L, �0.357.

Abbreviations: DEXA, dual-energy x-ray absorptiometry; HDL,high-density lipoprotein; LDL, low-density lipoprotein; NA, notapplicable; NIR, near-infrared; nPCR normalized protein cata-bolic rate; nPNA, normalized protein nitrogen appearance.

of our equations against NIR interactance–measured

Am J Kidney Dis. 2011;57(1):130-139

LBM in the development and validation cohorts,respectively. Analyses were repeated within the 2mutually exclusive strata of greater and less than themedian LBM to compare performance of the equa-tions within different ranges of LBM. All 3 equationstended to underestimate LBM, especially in the higherranges of LBM greater than its median value. Table 5and Fig 3 illustrate difference plot–based analyses andprovide correlation test results between NIR interacta-nce–measured LBM in the validation sample of 612participants and LBM estimates derived from SCr andMAMC regression equations in the development co-hort of 118 participants. Compared with NIR interac-tance–based LBM, both equations appeared accuratein predicting LBM, although consistent measurementbias in the form of underestimating LBM was ob-served. In women, both equations had smaller meandifferences in estimated LBM (mean differences, 2.1with NIR interactance and 2.3 with SCr and MAMC,respectively). Both equations tended to underestimateLBM in participants with higher LBM. Differenceplot analyses confirmed these findings (Fig 3). Notethat handgrip strength was assessed in only the devel-opment cohort and not the validation cohort; hence,its regression equation could not be examined furtherin the validation cohort.

To further verify the validity of the developedregression equations, we compared NIR interactance–measured LBM with the MAMC and SCr regressionequations in the validation cohort of 612 hemodialysispatients. Table 6 and Fig 4 show correlation coeffi-cients and scatterplots between NIR interactance–measured LBM and LBM estimates from each of the2 regression equations, respectively. Correlation coef-ficients were similarly high in the validation cohortcompared with the development cohort of 118 partici-pants.

In the development cohort, interaction terms withsex showed P � 0.20 for MAMC and SCr and P �0.14 for handgrip strength and were considered notmeaningful. Calculated root mean square errors werethe smallest in equations based on SCr level (3.43),handgrip strength (3.46), and MAMC (3.50; all P �0.001). Inclusion of all 3 predictors in the sameregression equation did not improve root mean squareerror (3.45). We also examined the inclusion of qua-dratic terms, which did not improve the gain inregression equation.

DISCUSSION

In the development cohort of 118 long-term hemodi-alysis patients, we examined correlations between sev-eral nutritional measures and LBM measured usingDEXA and found that SCr level, MAMC, and handgrip

strength had the highest correlations with LBM. We then

133

Noori et al

developed 3 regression equations based on SCr level,MAMC, and handgrip strength to estimate LBM. Whenvalidated against NIR interactance–measured LBM, theSCr and MAMC equations yielded accurate estimates ofLBM with reasonable concordance on the basis of bothdifference plots and Bland-Altman analyses in the valida-tion cohort of 612 hemodialysis patients. Compared withNIR interactance, both equations appeared accurate inpredicting LBM; however, they tended to underestimateLBM in participants with higher LBM.

Assessment of body composition, which classicallyis divided into fat and fat-free mass, is an importanttask for providing required nutritional care to patientswith chronic kidney disease. Compared with body fat,which stores energy in the form of adipose tissue,fat-free mass includes muscle and visceral proteinsand consists predominantly of water, protein, and

Table 2. Regression and Pearson Correlation Coefficient

Model 1(unadjusted)

Regressio

MAMC 1.73 (1.39 to 2.07)a

Handgrip strength 0.39 (0.22 to 0.56)a

SCr 1.14 (0.53 to 1.75)a

Serum albumin �1.81 (�8.14 to 4.86)Serum prealbumin 0.01 (�0.20 to 0.21)SGA �0.79 (�1.80 to 0.22)nPNA �9.59 (�17.46 to �1.72)a

Correlatio

MAMC 0.69b

Handgrip strength 0.40b

SCr 0.33b

Serum albumin �0.05Serum prealbumin 0.00SGA �0.15nPNA �0.22d

Note: Analysis was performed using 7 selected measures opatients. Values in parentheses are 95% confidence intervals. Moincludes all 7 surrogates in the model, and model 3 includes aobserved among DEXA-measured LBM values and MAMC, hand

Abbreviations: DEXA, dual-energy x-ray absorptiometry; LBMnormalized protein nitrogen appearance; SCr, serum creatinine;

aSummary estimate is statistically significant (P � 0.05).bP �0.001; cP � 0.01 to 0.001; dP � 0.05 to 0.01.

Box 1. LBM Estimation Equations

LBMSCr � 0.34 � SCr (mg/dL) � 5.58 � {1 if female; 0 if male} �0.30 � weight (in kg) � 0.67 � height (in inches) � 0.23 �URR � 5.75

LBMHGS � 9.09 � HGS (unit) � 5.15 � {1 if female; 0 if male} �0.33 � weight (in kg) � 0.74 � height (in inches) � 29.06

LBMMAMC � 0.28 � MAMC (cm) � 5.52 � {1 if female; 0 if male} �0.28 � weight (in kg) � 0.82 � height (in inches) � 35.30

Abbreviations: HGS, handgrip strength; LBM, lean bodymass; MAMC, midarm muscle circumference; SCr, serum creat-

inine; URR, urea reduction ratio.

134

minerals. Conventionally referred to as LBM, thisbody compartment is heterogeneous and its measure-ment is affected by abnormalities in fluid and electro-lyte distribution commonly observed in kidney patientpopulations.8

Body mass index, which is an attempt to adjustbody weight for height, is the most commonly usedsurrogate of fat or LBM, as well as nutritional status.The major limitation of body mass index is its inad-equacy to discriminate among the variations in the

EXA-Measured LBM With Measures of Nutritional Status

Model 2(all 7 measures included)

Model 3 (all 7 measures andcase-mix variables included)

efficients

1.19 (1.15 to 1.83)a 1.05 (0.59 to 1.51)a

0.32 (0.19 to 0.45)a 0.33 (0.16 to 0.44)a

0.64 (0.18 to 1.09)a 1.54 (0.64 to 2.43)a

3.90 (�8.60 to 0.71) �2.15 (�8.19 to 3.89)0.16 (�0.31 to �0.01)a �0.11 (�0.31 to 0.00)0.42 (�1.07 to 0.22) �0.44 (�1.26 to 0.38)0.48 (�6.32 to 5.36) �3.55 (�12.36 to 5.26)

efficients

0.68b 0.57c

0.46b 0.52d

0.28c 0.46d

�0.17 �0.10�0.22d �0.16�0.13 �0.16�0.01 �0.12

ritional status in the development cohort of 118 hemodialysisincludes each surrogate separately without adjustment, model 2

urrogates plus case-mix variables. Statistical correlations werestrength, and SCr level, but not with other nutritional markers.an body mass; MAMC, midarm muscle circumference; nPNA,subjective global assessment.

Figure 2. Box plots of lean body mass (LBM) measureddirectly using dual-energy x-ray absorptiometry (DEXA) andestimated indirectly using serum creatinine (SCr) level, midarmmuscle circumference (MAMC), and handgrip strength (HGS)using 3 different regression equations (see text) in the develop-ment cohort of 118 hemodialysis patients. Lower and upper box

s of D

n Co

����

n Co

f nutdel 1ll 7 sgrip, le

SGA,

boundaries are 25th and 75th percentiles, the line within the boxis the median, and whiskers extend to minimum and maximum.

Am J Kidney Dis. 2011;57(1):130-139

les.

LBM Equations in CKD

different constituents of body composition, that is, fatmass versus LBM. Therefore, accurate evaluation ofnutritional status needs a precise quantitative assess-ment of at least the 2 mentioned components of bodymass. Currently, DEXA is considered a reliable refer-ence method for body composition analysis and assess-ment of LBM in adult patients with chronic kidneydisease. Not withstanding its inability to differentiatebetween edema-related and muscle-associated wa-ter,8,33-35 DEXA measurements are based on 3-com-partmental models, that is, total-body minerals, fat-

Table 3. Performance of SCr-, Handgrip Strength–, and MAM

All Patients (n � 11

Median difference (kg)a

SCr equation 4.2 (3.5 to 4.7)MAMC equation 3.7 (3.4 to 5.1)Handgrip strength equation 4.1 (3.5 to 5.2)

IQR for differences (kg)b

SCr equation 3.9 (�0.8 to 10.4MAMC equation 4.3 (�1.3 to 10.8Handgrip strength equation 4.3 (�0.4 to 10.0

RMSESCr equation 2.9MAMC equation 2.6Handgrip strength equation 2.7

Note: Analysis was performed in the development cohort of 11across the mutually exclusive strata of greater and less than megiven in parentheses for median difference and IQRs.

Abbreviations: IQR, interquartile range; LBM, lean body massroot mean square error; SCr, serum creatinine.

aMedian difference refers to NIR interactance–measured LBMbIQR refers to the distance between the 25th and 75th percenti

Table 4. Performance of SCr- and MAMC-Based E

All Patients (n � 612)

Median difference (kg)a

SCr equation 3.4 (�3.2 to 12.0)MAMC equation 4.0 (�2.6 to 13.6)

IQR for differences (kg)b

SCr equation 3.0 (1.1 to 5.1)MAMC equation 3.7 (1.3 to 6.0)

RMSESCr equation 4.10MAMC equation 4.27

Note: Analysis was performed in the validation cohort of 612across the mutually exclusive strata of greater and less than the mgiven in parentheses for median difference and IQRs. Note that in

Abbreviations: IQR, interquartile range; LBM, lean body massroot mean square error; SCr, serum creatinine.

aMedian difference refers to NIR interactance–measured LBM

bIQR refers to the distance between the 25th and 75th percentiles.

Am J Kidney Dis. 2011;57(1):130-139

free soft mass or LBM, and fat tissue mass.1-3 Theequipment is not inexpensive and requires trained person-nel to operate. The DEXA machine is not a practical toolfor routine use in patients with chronic kidney diseasebecause of its technical complexity, its space-occupyingscanner (it requires participants to be in the supineposition), exposure to radiation, relatively high cost, andneed for trained and licensed personnel.6

To our knowledge, no prior study has developed orvalidated different regression methods to estimateLBM in long-term hemodialysis patients, although

ased Equations Relative to NIR Interactance–Measured LBM

Patients With Estimated LBM

<50 kg (n � 59) >50 kg (n � 59)

2.3 (1.4 to 3.1) 5.5 (4.3 to 6.4)2.2 (1.3 to 3.0) 5.8 (3.9 to 6.7)2.5 (2.0 to 2.9) 5.6 (4.8 to 7.0)

2.2 (�0.9 to 5.1) 4.0 (�2.5 to 11.3)2.4 (�1.5 to 7.0) 3.9 (�1.4 to 11.1)1.7 (�2.5 to 5.7) 3.8 (0.2 to 10.4)

1.7 3.02.0 2.71.9 2.8

odialysis patients. Values are calculated for all participants andLBM value, which was 50 kg. The 95% confidence intervals are

MC, midarm muscle circumference; NIR, near-infrared; RMSE,

s estimated LBM.

tions Relative to NIR Interactance–Measured LBM

Patients With Estimated LBM

<51 kg (n � 306) >51 kg (n � 306)

1.4 (�5.5 to 7.2) 5.4 (�0.7 to 15.9)1.7 (�4.3 to 7.6) 6.3 (�0.2 to 16.2)

1.7 (0.2 to 3.1) 4.8 (3.0 to 7.1)2.0 (0.2 to 3.4) 5.7 (4.1 to 8.0)

2.98 4.263.17 4.23

dialysis patients. Values are calculated for all participants andn LBM value, which was 51 kg. The 95% confidence intervals arevalidation cohort, handgrip strength was not measured.MC, midarm muscle circumference; NIR, near-infrared; RMSE,

s estimated LBM.

C-B

8)

)))

8 hemdian

; MA

minu

qua

hemoediathe

; MA

minu

135

value

Noori et al

several studies have done so in peritoneal dialysispatients.7,13,36 In the present study, we used DEXA asthe reference method for assessing LBM and com-pared different measures of nutritional status in esti-mating LBM in hemodialysis patients. We found that

Table 5. Difference Plot Analyses Comparing NIR Interactanc

Limits of Agreement M

Women (n � 298)SCr-estimated LBM �5.7 to 9.9MAMC-estimated LBM �5.8 to 10.4

Men (n � 314)SCr-estimated LBM �3.8 to 12.9MAMC-estimated LBM �3.0 to 14.0

All (n � 612)SCr-estimated LBM �5.1 to 11.9MAMC-estimated LBM �5.0 to 12.8

Note: Difference plot analyses are based on a modified Blahemodialysis patients.

Abbreviations: CI, confidence interval; LBM, lean body mass; Mcreatinine.

aPearson correlation between difference and NIR interactance

Figure 3. Difference plots between lean body mass (LBM)using near-infrared (NIR) interactance as reference standardand the 2 equations based on serum creatinine (SCr) level andmidarm muscle circumference (MAMC) in the validation cohort of612 hemodialysis patients. Medium dashed line is the difference,long dashed lines are limits of agreement (mean � 2 standard

deviations), and short dashed lines are 95% confidence intervals(CIs) for the difference.

136

serum albumin and prealbumin levels did not corre-late well with DEXA-measured LBM. Although se-rum albumin is measured routinely in most dialysispatients, it is an insensitive indicator of nutritionalstatus, especially because it may take several monthsof sustained visceral protein depletion for hypoalbu-minemia to develop.37 Serum albumin also may be amarker of systemic inflammation.38,39 Other visceralproteins have been used, including prealbumin,14

which has a shorter half-life than albumin and closecorrelation with nutritional status and is a good predic-tor of clinical outcomes.40 However, prealbumin alsodid not correlate well with LBM in our study. Duringrecent years, subjective global assessment41 has beenused increasingly to assess nutritional status in dialy-

Table 6. Mean Values for LBM and Correlation CoefficientsBetween NIR Interactance–Measured LBM and Estimated LBM

Total(n � 612)

Women(n � 298)

Men(n � 314)

LBM (kg)NIR interactance-

measured LBM52.2 � 11.6 44.6 � 8.5 59.1 � 9.6

SCr-estimated LBM 48.7 � 9.8 42.5 � 7.3 54.6 � 7.9MAMC-estimated LBM 48.3 � 9.6 42.3 � 7.3 54.0 � 8.0

Correlations (r)a,b

SCr-estimated LBM 0.93 0.88 0.91MAMC-estimated LBM 0.93 0.87 0.90

Note: Values are shown as mean � standard deviation forLBM measurements or estimates. Analyses are based on all 612hemodialysis patients in the validation cohort.

Abbreviations: LBM, lean body mass; MAMC, midarm musclecircumference; NIR, near-infrared; SCr, serum creatinine.

aPearson correlation coefficients (r) between NIR interactance–measured LBM and each of the 2 estimates of LBM using theSCr- and MAMC-based LBM estimating equations in 612 long-term hemodialysis patients of the validation cohort.

easured LBM With SCr- and MAMC-Based Estimates of LBM

Difference (95% CI) Correlationa (r) Correlation P

.1 (1.6 to 2.6) 0.45 �0.001

.3 (1.8 to 2.8) 0.47 �0.001

.6 (4.1 to 5.0) 0.62 �0.001

.5 (5.0 to 6.0) 0.59 �0.001

.4 (3.0 to 3.7) 0.58 �0.001

.9 (3.6 to 4.3) 0.61 �0.001

tman test and were performed in the validation cohort of 612

, midarm muscle circumference; NIR, near-infrared; SCr, serum

s.

e–M

ean

22

45

33

nd-Al

AMC

bAll P � 0.001.

Am J Kidney Dis. 2011;57(1):130-139

LBM Equations in CKD

sis patients.10,41-44 Subjective global assessment cor-relates well with many nutritional markers in thesepatients10,14,42,43 and has a high predictive value formortality.44 However, in our study, subjective globalassessment did not correlate with DEXA-measuredLBM. Although subjective global assessment reflectsoverall nutritional status, LBM may be more represen-tative of the somatic protein pool.

We found that SCr level was among the 3 bettercorrelates of LBM, and its regression equation thatwas combined with demographics and urea reductionratio was accurate enough to estimate LBM in hemo-dialysis patients. SCr level is affected by musclemass, kidney function or dialysis adequacy, and di-etary protein (meat) intake.45 The dietary variationcan be mitigated if averaged values over a long periodare used, as we did in our study by using 3-monthaveraged SCr values. We also included 3-month aver-aged urea reduction ratio in the regression equation sothat variations in SCr levels based on changes indialysis dose and adequacy can be compensated for.Because SCr is measured at least monthly in alldialysis patients in the United States and most othercountries, we believe our equations can be used conve-

Figure 4. Scatterplots, regression lines, and 95% confidenceintervals (CIs) reflecting correlations between near-infrared (NIR)interactance–measured lean body mass (LBM) and LBM esti-mated using (A) midarm muscle circumference (MAMC) and (B)serum creatinine (SCr) level in the validation cohort of 612hemodialysis patients. Shaded areas reflect 95% CIs.

niently to estimate LBM in these patients.

Am J Kidney Dis. 2011;57(1):130-139

We also found that MAMC yielded a reliable esti-mate of LBM. Although use of anthropometric meth-ods is an indirect and insensitive means of evaluationwith several inherent errors, including the influence ofhydration status, findings in our study pertaining toMAMC are consistent with some prior studies thatused DEXA as the reference standard.14 MAMC tradi-tionally has been used as a convenient and noninva-sive method for estimating LBM despite its limitedreproducibility and precision caused by high intra-and interobserver measurement variability.8,46,47 Inour study, we found good correlation between MAMCand LBM, especially in men. We also found thathandgrip strength, which is a convenient assessmentmethod for upper-extremity muscle strength, corre-lated well with DEXA-measured LBM. Handgripstrength is a simple test for assessment of musclestrength in dialysis outpatients,42,48 but its utility toestimate LBM has been studied in only patients with-out chronic kidney disease.14 In our study, handgripstrength had a strong correlation with LBM, espe-cially in women. Evidence suggests that handgripstrength may be a good measure of nutritional statusand predictor of mortality and complications in surgi-cal patients.49,50 Second, handgrip strength also hascorrelated closely with other nutritional parameters,for example, protein index (assessed using neutronactivation) in surgical patients51 and fat-free mass(assessed using anthropometry) in patients with chronicheart failure.52 Third, handgrip strength has beenreported to improve with nutritional supplementa-tion,53,54 indicating that its variation is a function ofnutritional interventions. Finally, some studies foundthat handgrip strength is lower in malnourished dialy-sis patients.42,48

Our study should be qualified for its relatively largeproportions of African Americans and Hispanics andfor potential selection bias because of exclusion ofpatients for whom the upper arm deemed inappropri-ate for NIR interactance or anthropometric measure-ment. Furthermore, all 3 equations tended to underes-timate LBM, especially in the higher ranges of LBMgreater than its median value (see Tables 3 and 4).Another potential limitation is that we used DEXA inthe development cohort, but not in the validationcohort, for which NIR interactance was used instead.However, NIR interactance measurements of lean andfat mass correlate closely with DEXA.29 Differencesin characteristics of people in the development andvalidation cohorts also could affect results. We shouldnote that variation in fluid status may affect DEXAmeasurements, whereas NIR interactance and anthro-pometric measurements are less prone to this con-founder. Our NIR interactance–based validation stud-

ies show that DEXA-based regression equations were

137

Noori et al

adequately valid, augmenting the robustness of ourdeveloped equations. We did not compare DEXA orfield methods with underwater weighing or air dis-placement techniques; however, these elaborate andcumbersome techniques are rarely used in dialysispatient studies.

In conclusion, in long-term hemodialysis patients,our novel equations to estimate LBM based on SCrlevel, MAMC, or handgrip strength appear valid andyield accurate estimates of DEXA- or NIR interac-tance–measured LBM, although in higher LBM ranges,they may underestimate it. SCr level, MAMC, andhandgrip strength are practical and inexpensive assess-ments that can be used for routine assessment ofnutritional status or in clinical or epidemiologic stud-ies, bearing their limitations in mind. Given emergingstudies that indicate the association of greater musclemass with better survival in hemodialysis patients,55,56

additional studies using these or other reference stan-dards and equations are needed to verify the accuracyand reliability of our developed regression equations.

ACKNOWLEDGEMENTSWe thank the hard-working collaborating dietitians in 8 DaVita

dialysis clinics in the Los Angeles South Bay area and DaVitateammates in these facilities, as well as DaVita Clinical Research.

Support: This study was supported by grants K23-DK61162 andR21-DK078012 from the National Institute of Diabetes, Digestiveand Kidney Diseases at the National Institutes of Health (NIH).Additional sources of funding include research grants from WatsonPharmaceuticals, DaVita Clinical Research, and Harold Simmons(to Dr Kalantar-Zadeh) and General Clinical Research Centergrant # M01-RR00425 from the National Centers for ResearchResources, NIH.

Financial Disclosure: The authors declare that they have norelevant financial interests.

REFERENCES1. Kovesdy CP, Kalantar-Zadeh K. Why is protein-energy wast-

ing associated with mortality in chronic kidney disease? SeminNephrol. 2009;29:3-14.

2. Kalantar-Zadeh K, Kilpatrick RD, Kuwae N, et al. Revisitingmortality predictability of serum albumin in the dialysis popula-tion: time dependency, longitudinal changes and population-attributable fraction. Nephrol Dial Transplant. 2005;20:1880-1888.

3. Rambod M, Kovesdy CP, Bross R, Kopple JD, Kalantar-Zadeh K. Association of serum prealbumin and its changes overtime with clinical outcomes and survival in patients receivinghemodialysis. Am J Clin Nutr. 2008;88:1485-1494.

4. Fouque D, Kalantar-Zadeh K, Kopple J, et al. A proposednomenclature and diagnostic criteria for protein-energy wasting inacute and chronic kidney disease. Kidney Int. 2008;73:391-398.

5. Rambod M, Bross R, Zitterkoph J, et al. Association ofmalnutrition-inflammation score with quality of life and mortalityin hemodialysis patients: a 5-year prospective cohort study. Am JKidney Dis. 2009;53:298-309.

6. Donadio C, Halim AB, Caprio F, Grassi G, Khedr B, Mazzan-tini M. Single- and multi-frequency bioelectrical impedance analy-

ses to analyse body composition in maintenance haemodialysis

138

patients: comparison with dual-energy x-ray absorptiometry. PhysiolMeas. 2008;29:S517-524.

7. Negri AL, Barone R, Veron D, et al. Lean mass estimation bycreatinine kinetics and dual-energy x-ray absorptiometry in perito-neal dialysis. Nephron Clin Pract. 2003;95:c9-14.

8. Kamimura MA, Avesani CM, Cendoroglo M, Canziani ME,Draibe SA, Cuppari L. Comparison of skinfold thicknesses andbioelectrical impedance analysis with dual-energy x-ray absorpti-ometry for the assessment of body fat in patients on long-termhaemodialysis therapy. Nephrol Dial Transplant. 2003;18:101-105.

9. Delwaide PA, Crenier EJ. Body potassium as related to leanbody mass measured by total water determination and by anthropo-metric method. Hum Biol. 1973;45:509-526.

10. Young GA, Kopple JD, Lindholm B, et al. Nutritionalassessment of continuous ambulatory peritoneal dialysis patients:an international study. Am J Kidney Dis. 1991;17:462-471.

11. Bhatla B, Moore H, Emerson P, et al. Lean body massestimation by creatinine kinetics, bioimpedance, and dual energyx-ray absorptiometry in patients on continuous ambulatory perito-neal dialysis. ASAIO J. 1995;41:M442-446.

12. Keshaviah PR, Nolph KD, Moore HL, et al. Lean bodymass estimation by creatinine kinetics. J Am Soc Nephrol. 1994;4:1475-1485.

13. Szeto CC, Kong J, Wu AK, Wong TY, Wang AY, Li PK. Therole of lean body mass as a nutritional index in Chinese peritonealdialysis patients—comparison of creatinine kinetics method andanthropometric method. Perit Dial Int. 2000;20:708-714.

14. Heimburger O, Qureshi AR, Blaner WS, Berglund L, Sten-vinkel P. Hand-grip muscle strength, lean body mass, and plasmaproteins as markers of nutritional status in patients with chronicrenal failure close to start of dialysis therapy. Am J Kidney Dis.2000;36:1213-1225.

15. Lo WK, Prowant BF, Moore HL, et al. Comparison ofdifferent measurements of lean body mass in normal individualsand in chronic peritoneal dialysis patients. Am J Kidney Dis.1994;23:74-85.

16. de Fijter WM, de Fijter CW, Oe PL, ter Wee PM, DonkerAJ. Assessment of total body water and lean body mass fromanthropometry, Watson formula, creatinine kinetics, and bodyelectrical impedance compared with antipyrine kinetics in perito-neal dialysis patients. Nephrol Dial Transplant. 1997;12:151-156.

17. Colman S, Bross R, Benner D, et al. The Nutritional andInflammatory Evaluation in Dialysis patients (NIED) Study: over-view of the NIED Study and the role of dietitians. J Ren Nutr.2005;15:231-243.

18. Kalantar-Zadeh K, Kopple JD, Kamranpour N, FogelmanAM, Navab M. HDL-inflammatory index correlates with pooroutcome in hemodialysis patients. Kidney Int. 2007;72:1149-1156.

19. Bross R, Zitterkoph J, Pithia J, et al. Association of serumtotal iron-binding capacity and its changes over time with nutri-tional and clinical outcomes in hemodialysis patients. Am J Neph-rol. 2009;29:571-581.

20. Shantouf R, Kovesdy CP, Kim Y, et al. Association of serumalkaline phosphatase with coronary artery calcification in mainte-nance hemodialysis patients. Clin J Am Soc Nephrol. 2009;4:1106-1114.

21. Rambod M, Kovesdy CP, Kalantar-Zadeh K. Malnutrition-inflammation score for risk stratification of patients with CKD: is itthe promised gold standard? Nat Clin Pract Nephrol. 2008;4:354-355.

22. Rambod M, Kovesdy CP, Kalantar-Zadeh K. Combinedhigh serum ferritin and low iron saturation in hemodialysis pa-tients: the role of inflammation. Clin J Am Soc Nephrol. 2008;3:

1691-1701.

Am J Kidney Dis. 2011;57(1):130-139

LBM Equations in CKD

23. Raj DS, Shah VO, Rambod M, Kovesdy CP, Kalantar-Zadeh K. Association of soluble endotoxin receptor CD14 andmortality among patients undergoing hemodialysis. Am J KidneyDis. 2009;54:1062-1071.

24. Lohman T, Roche A, Martorell R. Anthropometric Standard-ization Reference Manual. Champaign, IL: Human Kinetics Books;1988.

25. Roche AF, Sievogel RM, Chumlea WC, Webb P. Gradingbody fatness from limited anthropometric data. Am J Clin Nutr.1981;34:2831-2838.

26. Durnin JV, Womersley J. Body fat assessed from total bodydensity and its estimation from skinfold thickness: measurementson 481 men and women aged from 16 to 72 years. Br J Nutr.1974;32:77-97.

27. Gibney MJEM, Ljungqvist J. Clinical nutrition: anthropo-metric assessment of body composition. 3rd Ed. Oxford, UK:Blackwell; 2005:20.

28. Park JC, Kovesdy CP, Duong U, et al. Association of serumalkaline phosphatase and bone mineral density in maintenancehemodialysis patients. Hemodial Int. 2010;14(2):182-192.

29. Bross R, Chandramohan G, Kovesdy CP, et al. Comparingbody composition assessment tests in long-term hemodialysispatients. Am J Kidney Dis. 2010;55:885-896.

30. Miller JE, Kovesdy CP, Nissenson AR, et al. Association ofhemodialysis treatment time and dose with mortality and the roleof race and sex. Am J Kidney Dis. 2010;55:100-112.

31. Petersen PH, Stockl D, Blaabjerg O, et al. Graphical inter-pretation of analytical data from comparison of a field method withreference method by use of difference plots. Clin Chem. 1997;43:2039-2046.

32. Bland JM, Altman DG. Statistical methods for assessingagreement between two methods of clinical measurement. Lancet.1986;1:307-310.

33. Formica C, Atkinson MG, Nyulasi I, McKay J, Heale W,Seeman E. Body composition following hemodialysis: studiesusing dual-energy x-ray absorptiometry and bioelectrical imped-ance analysis. Osteoporos Int. 1993;3:192-197.

34. van den Ham EC, Kooman JP, Christiaans MH, et al. Bodycomposition in renal transplant patients: bioimpedance analysiscompared to isotope dilution, dual energy x-ray absorptiometry,and anthropometry. J Am Soc Nephrol. 1999;10:1067-1079.

35. Locatelli F, Fouque D, Heimburger O, et al. Nutritionalstatus in dialysis patients: a European consensus. Nephrol DialTransplant. 2002;17:563-572.

36. Dong J, Li YJ, Lu XH, Gan HP, Zuo L, Wang HY. Correla-tions of lean body mass with nutritional indicators and mortality inpatients on peritoneal dialysis. Kidney Int. 2008;73:334-340.

37. Abrahamsen B, Hansen TB, Hogsberg IM, Pedersen FB,Beck-Nielsen H. Impact of hemodialysis on dual x-ray absorptiom-etry, bioelectrical impedance measurements, and anthropometry.Am J Clin Nutr. 1996;63:80-86.

38. Kaysen GA, Stevenson FT, Depner TA. Determinants ofalbumin concentration in hemodialysis patients. Am J Kidney Dis.1997;29:658-668.

39. Yeun JY, Kaysen GA. Factors influencing serum albumin indialysis patients. Am J Kidney Dis. 1998;32(suppl 4):S118-125.

40. Sreedhara R, Avram MM, Blanco M, Batish R, Mittman N.Prealbumin is the best nutritional predictor of survival in hemodi-

alysis and peritoneal dialysis. Am J Kidney Dis. 1996;28:937-942.

Am J Kidney Dis. 2011;57(1):130-139

41. Detsky AS, McLaughlin JR, Baker JP, et al. What issubjective global assessment of nutritional status? JPEN J Par-enter Enteral Nutr. 1987;11:8-13.

42. Qureshi AR, Alvestrand A, Danielsson A, et al. Factorspredicting malnutrition in hemodialysis patients: a cross-sectionalstudy. Kidney Int. 1998;53:773-782.

43. Enia G, Sicuso C, Alati G, Zoccali C. Subjective globalassessment of nutrition in dialysis patients. Nephrol Dial Trans-plant. 1993;8:1094-1098.

44. McCusker FX, Teehan BP, Thorpe KE, Keshaviah PR,Churchill DN. How much peritoneal dialysis is required forthe maintenance of a good nutritional state? Canada-USA(CANUSA) Peritoneal Dialysis Study Group. Kidney Int Suppl.1996;56:S56-61.

45. Mayersohn M, Conrad KA, Achari R. The influence of acooked meat meal on creatinine plasma concentration and creati-nine clearance. Br J Clin Pharmacol. 1983;15:227-230.

46. Oe B, de Fijter CW, Oe PL, Stevens P, de Vries PM.Four-site skinfold anthropometry (FSA) versus body impedanceanalysis (BIA) in assessing nutritional status of patients on mainte-nance hemodialysis: which method is to be preferred in routinepatient care? Clin Nephrol. 1998;49:180-185.

47. Woodrow G, Oldroyd B, Smith MA, Turney JH. Measure-ment of body composition in chronic renal failure: comparisonof skinfold anthropometry and bioelectrical impedance withdual energy x-ray absorptiometry. Eur J Clin Nutr. 1996;50:295-301.

48. Jones CH, Newstead CG, Will EJ, Smye SW, Davison AM.Assessment of nutritional status in CAPD patients: serum albuminis not a useful measure. Nephrol Dial Transplant. 1997;12:1406-1413.

49. Webb AR, Newman LA, Taylor M, Keogh JB. Hand gripdynamometry as a predictor of postoperative complications reap-praisal using age standardized grip strengths. JPEN J ParenterEnteral Nutr. 1989;13:30-33.

50. Klidjian AM, Archer TJ, Foster KJ, Karran SJ. Detection ofdangerous malnutrition. JPEN J Parenter Enteral Nutr. 1982;6:119-121.

51. Windsor JA, Hill GL. Grip strength: a measure of theproportion of protein loss in surgical patients. Br J Surg. 1988;75:880-882.

52. McParland C, Resch EF, Krishnan B, Wang Y, Cujec B,Gallagher CG. Inspiratory muscle weakness in chronic heartfailure: role of nutrition and electrolyte status and systemic myop-athy. Am J Respir Crit Care Med. 1995;151:1101-1107.

53. Gray-Donald K, Payette H, Boutier V, Page S. Evaluation ofthe dietary intake of homebound elderly and the feasibility ofdietary supplementation. J Am Coll Nutr. 1994;13:277-284.

54. Efthimiou J, Fleming J, Gomes C, Spiro SG. The effect ofsupplementary oral nutrition in poorly nourished patients withchronic obstructive pulmonary disease. Am Rev Respir Dis. 1988;137:1075-1082.

55. Kalantar-Zadeh K, Streja E, Kovesdy CP, et al. Obesityparadox and mortality-predictability of surrogates of body size andmuscle mass in hemodialysis patients. Mayo Clin Proc. 2010. Inpress.

56. Noori N, Kopple JD, Kovesdy CP, et al. Mid-arm musclecircumference and quality of life and survival in maintenancehemodialysis patients [Epub ahead of print October 14, 2010]. Clin

J Am Soc Nephrol. doi:10.2215/CJN.02080310.

139