Assessment of Measures for Abdominal Adiposity in Persons with Spinal Cord Injury

Ultrasound in Med. & Biol., Vol. 37, No. 5, pp. 734–741, 2011Published by Elsevier Inc. on behalf ofWorld Federation for Ultrasound inMedicine &Biology

Printed in the USA. All rights reserved0301-5629/$ - see front matter

asmedbio.2011.02.002

d Original Contribution

ASSESSMENT OF MEASURES FOR ABDOMINAL ADIPOSITY IN PERSONSWITH SPINAL CORD INJURY

RACINE R. EMMONS,*y CAROL EWING GARBER,y CHRISTOPHER M. CIRNIGLIARO,*STEVEN C. KIRSHBLUM,zx ANN M. SPUNGEN,*{ and WILLIAM A. BAUMAN*{

*Department of Veterans Affairs, Rehabilitation Research and Development Center of Excellence for the MedicalConsequences of Spinal Cord Injury, James J. Peters VA Medical Center, Bronx, NY; yTeachers College, Columbia University,New York, NY; zKessler Institute for Rehabilitation, West Orange, NJ; xDepartment of Physical Medicine and Rehabilitation,University Medical School of New Jersey, Newark, NJ; and {Departments of Medicine & Rehabilitation Medicine, The Mount

Sinai School of Medicine, New York, NY

(Received 4 October 2010; revised 25 January 2011; in final form 3 February 2011)

ACenterPetersBronx,

Abstract—Ultrasoundmay be a useful tool to assess abdominal adiposity, but it has not been validated in the spinalcord injury (SCI) population. This study evaluated associations between abdominal ultrasound and other methodsto assess adiposity in 24 men with SCI and 20 able-bodied (AB) men. Waist (WC) and hip circumference (HC) andwaist-to-hip ratio (WHR) were measured. Trunk (TRK%), android (A%) and waist fat (W%) were determined bydual energy x-ray absorptiometry (DXA); ultrasonography determined abdominal subcutaneous (SF) and visceralfat (VF). The SCI group had greater TRK% (40.0 ± 9.6 vs. 32.0 ± 10.3), W% (47.0 ± 9.7 vs. 40.6 ± 9.4), A% (43.0 ±9.8 vs. 35.8 ± 10.6) andWHR (0.99 ± 0.1 vs. 0.92 ± 0.06) than the AB group.WC andWHR correlatedwith VF in theSCI group. These associations suggest that ultrasoundmay be a useful tool in clinical practice for the measurementof VF in weight loss programs and for the assessment of cardiometabolic disorders. (E-mail: [email protected]) Published by Elsevier Inc. on behalf of World Federation for Ultrasound in Medicine & Biology.

Key Words: Spinal cord injury, Abdominal adiposity, Abdominal ultrasonography.

INTRODUCTION

Individuals with excess body fat, primarily abdominal fat,are at an increased risk for metabolic abnormalities,which may predispose to cardiovascular disease (CVD).Approximately 65% of individuals with type 2 diabetesmellitus are overweight or obese (Bray 2004). Evidencesuggests that individuals with spinal cord injury (SCI)may be at a greater risk of CVD because of a higher prev-alence of dyslipidemia, insulin resistance, sedentary life-style and obesity compared with able-bodied (AB)individuals (Bauman and Spungen 2008; Gater 2007;Kocina 1997). It is estimated that approximately twothirds of the SCI population is obese (Gater 2007).

The accurate assessment of visceral (e.g., abdom-inal) fat is relevant in determining the risk of CVD inthe AB population. Body mass index (BMI; kg/m2) is

ddress correspondence to: Racine R. Emmons, Ed.D., VAof Excellence for the Medical Consequences of SCI, James J.VA Medical Center, Room 7A-13, 130 West Kingsbridge Rd.,NY 10468. E-mail: [email protected]

734

widely used clinically and in epidemiological studies toassess overall body habitus, although it is an indirectmeasure of adiposity. Because of variable decreases inmuscle mass and increases in fat mass after SCI,BMI may be even less accurate a measure of obesity inpersons with SCI because it will tend to underestimatethe degree of total body adiposity, and in a less predict-able manner (Jones et al. 2003; Spungen et al. 2003). Inaddition, BMI does not provide information on visceraladiposity, which is of particular importance in assessingcardiometabolic risk (Merino-Ibarra et al. 2005).

Anthropometric measures, such as waist circumfer-ence (WC) and waist-to-hip ratio (WHR), are easy andinexpensive to performandprovide surrogatemeasurementof abdominal adiposity and associated CVD risk. Of note,WC and WHR have been shown to be highly correlatedto more sophisticated and precise techniques, such ascomputed tomography (CT) and magnetic-resonanceimaging (MRI) in the AB population (Lee et al. 2008).However, these anthropometric methods are limited inthat they do not directly measure visceral adipose tissuethickness, but rather they include subcutaneous adipose

Adiposity in spinal cord injury d R. R. EMMONS et al. 735

tissue and only provide a measure of overall trunk girth(Bertin et al. 2000).

CT and MRI have been considered to be the ‘‘goldstandard’’ methods for measuring abdominal fat becausethey have the capability to precisely measure fat thick-ness in any body compartment, including the abdominalregion (Ribeiro-Filho et al. 2001, 2003); each has thecapacity to distinguish between subcutaneous andvisceral fat. However, these methods are costly toperform. MRI has a long acquisition time and a highamount of error because of motion artifact (Bertin et al.2000; Clasey et al. 1999). CT, although time-efficient,requires highly-specialized instrumentation and exposespatients to significant levels of ionizing radiation.

Dual-energy X-ray absorptiometry (DXA) iscommonly used to measure total and regional bodycomposition, including measures of bone mineralcontent, bone mineral density and lean and fat tissuemass; it has been proposed as a suitable alternative tomeasuring abdominal adiposity because it is relativelyinexpensive, rapid to perform and requires minimal coop-eration from the patient. DXA is limited because it cannotdistinguish between subcutaneous and visceral fat,cannot quantify the thickness of fat in a region of interestand does not account for hydration status, which may bealtered in persons with SCI.

Ultrasonography (US) is a method to measureabdominal fat thickness, being a relatively simple, inex-pensive and noninvasive method. Abdominal fatmeasurements by US have been shown to have a strongcorrelation to CT and MRI measurements in AB subjects(Guldiken et al. 2006; Hirooka et al. 2005; Ribeiro-Filhoet al. 2001, 2003). To date, there are no studies reportingthe use of abdominal US for the measurement ofsubcutaneous fat (SF) and visceral fat (VF) in the SCIpopulation. This study assessed the association amongabdominal US measurements and other methods ofmeasurement routinely used in clinical practice, as wellas body composition measurements used to determinethe degree of total body adiposity.

MATERIALS AND METHODS

SubjectsForty-four male subjects (20 AB and 24 SCI)

between 20 and 70 years of age were recruited for partic-ipation in this study from an existing subject database,physician referral, medical records or during a clinicalvisit to the Department of Veterans Affairs RehabilitationResearch and Development National Center of Excel-lence for the Medical Consequences of Spinal CordInjury James J. Peters Veterans Affairs Medical Center(JJPVAMC), Bronx, NY; or Kessler Institute for Rehabil-itation (KIR), West Orange, NJ. Individuals with SCI

were included if they had a chronic injury ($6 months),regardless of completeness or etiology of injury, andwere clinically stable. Subjects were excluded fromparticipation if they had any type of metal implant, hadbeen diagnosed by a physician with a pressure ulcer, dia-betes, coronary heart disease and/or any other chronicillness, or lacked the mental capacity to provide informedconsent. The SCI group included 8 subjects with para-plegia and 16 subjects with tetraplegia. Using theAmerican Spinal Injury Association Impairment Scale(AIS) to classify neurologic impairment, 19 subjectswere previously diagnosed with motor complete injury(AIS A or B) and 5 subjects with motor incomplete injury(AIS C and D). The study was approved by the Institu-tional Review Boards at the JJPVAMC and KIR.

ProceduresSubjects underwent one day of testing at the

JJPVAMC or KIR. After providing informed consent inaccordance with the policies of the Institutional ReviewBoards at each of the institutions, subjects completedanthropometric measurements, abdominal US imagingand a total-body DXA scan.

Anthropometric and body compositionmeasurements. Subjects wore a lightweight shirt andpants or shorts without metal parts during the bodycomposition and anthropometric assessment. All extra-neous clothing, shoes and metal (e.g., watches, rings,necklaces) were removed before testing.

DXA scans were performed (LUNAR ProdigyAdvance, GE Lunar, Madison, WI) to determine totalbody percent fat (TBF%), regions of trunk percent fat,body length and total body weight. Because of the possi-bility of contractures and spasms in the men with SCI,particular effort was made to position each subject asprecisely as possible. Total body scan time lasted approx-imately 5 min. Regions of interest (ROIs) were adjustedin the total body scan for each image; body segmentswere analyzed by one investigator at both sites, as recom-mended by the manufacturer.

Three specific ROIs were collected for the analysisof trunk adiposity. The android region (A%) was extrap-olated from the total body scan and was defined by themanufacturer inferiorly at the pelvis cut line, superiorlyabove the pelvic cut line by 20% of the distance betweenthe pelvis and neck cut lines and laterally at the arm cutlines (Fig. 1). Two custom ROIs were analyzed to inves-tigate the different areas of abdominal fat accumulation(Fig. 2). The suprailiac trunk region (to permit determina-tion of trunk region percent fat [TRK%]) was definedsuperiorly as one-third the distance from the superioraspect of the iliac crest to the knee joint, from the superioriliac crests and inferiorly as a horizontal line at the level

Fig. 1. Android region (A%) from a total-body DXA scan. (a) A% is outlined in the dashed box inferiorly at the pelvis cut line,superiorly above the pelvic cut line by 20% of the distancebetween the pelvis and neck cut lines and laterally at the arm

cut lines.

Fig. 2. Trunk fat custom regions of interest. (a) Suprailiac trunkfat region (TRK%) height was measured as one-third thedistance between the top of the iliac crest and knee joint byDXA (Walton et al. 1995), represented by the solid-lined box.(b) Waist fat region (W%) was measured as a 1-cm slice atthe level of the iliac crests, represented by the dashed-line box.

736 Ultrasound in Medicine and Biology Volume 37, Number 5, 2011

of the superior iliac crest. Laterally, it spanned thedistance to encompass the soft tissue, with careful consid-eration not to include the hands and arms (Walton et al.1995). The waist region (to permit determination of waistregion percent fat [W%]) was defined superiorly 1 cmabove the superior aspect of the iliac crests, inferiorlyas a horizontal line at the level of the superior iliac crestsand laterally to encompass the soft tissue.

Daily quality assurance (QA) testing was performedto determine the stability of the densitometer. The preci-sion was determined at both study sites for 150 phantom

scans, in which the reference range was 30% fat; thecoefficient of variation was less than 1%.

Body weight was recorded as the total body massmeasured via the DXA scan. Height (body length) wasdetermined from the total body scan using electronic cali-pers to measure the length from the top of the skull to thecalcaneus. BMI was calculated from measures from theDXA scan as weight in kilograms divided by body lengthin meters squared.

While each subject remained in the supine positionon the DXA table, a flexible measuring tape was wrappedaround each subject’s waist at the level of the umbilicus tomeasure WC. HC was measured at the level of the iliaccrest. Three measurements were taken at each site andthe average was recorded. The ratio of the WHR wascalculated.

Abdominal ultrasound. Subjects remained supine ona table for approximately 20 minutes while the US exam-ination was performed by a trained sonographer. Scans

Adiposity in spinal cord injury d R. R. EMMONS et al. 737

were interpreted by one sonographer for both testingsites. All scans were obtained using the GE Logiq BookXP (GE Healthcare, Buckinghamshire, UK) using a 3C-RS 2-5 MHz curvilinear transducer (GE Healthcare) tomeasure SF and VF from the methods described byRiberio-Filho et al. (2003) and Hirooka et al. (2005).SF was measured with the transducer set at 5 MHz intransverse position with minimal pressure applied so asnot to compress the adipose tissue. The area 1 cm abovethe umbilicus along the xipho-umbilical line wasmeasured upon exhalation as the distance between theskin-fat interface and the outer edge of the linea alba(Fig. 1); 3 measures of this distance were recorded andaveraged. VF was measured in transverse as the distancebetween the linea alba and the anterior aorta, 1 cm abovethe umbilicus along the xipho-umbilical line (Fig. 3); 3measures were recorded and averaged.

Statistical analysisThe data were analyzed using SPSS Version 13,

(SPSS, Inc., Chicago, IL). An a priori alpha level wasset at p# 0.05. Demographic and descriptive data are re-ported as mean 6 standard deviation (SD). Independentsample t-tests determined the differences between groups(AB and SCI) and each of the independent variables (age,weight, height, BMI, TBF%, TRK%,W%, A%,WC, HC,WH, SF, VF). Pearson correlation coefficients were usedto test associations between US measures of SF and VFand body composition assessments of BMI, TBF%,TRK%, W%, A%, WC and WH.

RESULTS

Characteristics of the subjects are shown in Table 1.Subjects ranged in age from 23–64 years, and in the SCIgroup, duration of injury ranged from 2–36 years.Subjects with SCI were significantly taller than AB

Fig. 3. Abdominal fat measurements by ultrasound. (a) Subcutalicus along the xipho-umbilical line and measured upon exhalaouter edge of the linea alba. (b) Visceral fat (VF) measurement

linea alba and the anterior aorta, 1 cm above the

subjects (1.79 6 0.07 m vs. 1.74 6 0.08 m, respectively,p 5 0.04), but no significant differences in body mass orBMI were noted.

Measures of body compositionMeasures of body composition are shown in Table 2.

There were no significant differences among groups forBMI, HC, SF, and VF. TBF% was significantly greaterin subjects with SCI than in the AB group (35.9 67.5% vs. 25.8 6 7.2%, p , 0.05). Although not reachingsignificance, WC was trending to be greater in the SCIthan the AB group (99.8 6 14.8 cm vs. 91.3 6 13.6cm, respectively, p 5 0.06). The group with SCI hadapproximately 36% total body fat compared with 26%in the AB group as measured by DXA, and those withSCI had greater TRK%, W%, A% and WHR than ABsubjects.

Correlations between abdominal US measures andother assessments of body composition are shown inTable 3. In subjects with SCI, SF approached a significantcorrelation with TRK%; in AB subjects, SF correlatedwith BMI (r 5 0.51, p 5 0.02) and WC. BMI trendedtoward a correlation with VF in subjects with SCI. Signif-icant correlations were noted between VF and BMI (r 50.58, p5 0.01) and VF and TBF% (r5 0.80, p, 0.01) inthe AB group. All measures of abdominal fat in ABsubjects were highly correlated with VF; however, onlyA%, WC (Fig. 4) and WHR were significantly correlatedin subjects with SCI.

DISCUSSION

This is the first study to report the use of US for theassessment of abdominal fat compartments in individualswith SCI. Measures of VF were found to significantlycorrelate with measures of WC, WHR and A% measuredby DXA in AB and SCI men. These results support the

neous fat (SF) measured as the area 1 cm above the umbi-tion as the distance between the skin-fat interface and theby ultrasound was measured as the distance between theumbilicus along the xipho-umbilical line.

Table 1. Characteristics of the subjects

AB SCI

n 5 20 n 5 24

Age (y) 39 6 11 44 6 10Height (m) 1.7 6 0.08 1.79 6 0.07*Weight (kg) 84.1 6 17.8 83.2 6 18.3BMI (kg/m2) 27.6 6 4.8 26.2 6 5.4DOI (years) 19 6 11Complete/Incomplete 19/5

Values are expressed as mean 6 SD.AB = able-bodied; SCI = spinal cord injury; BMI = body mass index;

DOI 5 duration of injury.* p , 0.05 between AB and SCI.

Table 3. Correlation coefficients between abdominal USmeasures and other measures for evaluation of abdominal

adiposity

Subcutaneous fat Visceral fat

AB SCI AB SCI

r r r r

TRK% 0.34 0.39 0.78* 0.33W% 0.44 0.29 0.67* 0.28A% 0.39 0.34 0.78* 0.42*WC 0.52* 0.14 0.68* 0.55*WHR 0.35 0.05 0.51* 0.55*

AB 5 able-bodied; SCI 5 spinal cord injury; BMI 5 body massindex; TBF% 5 % total body fat measured by DXA; TRK% 5 trunkregion % fat (Walton 1995); W% 5 waist region % fat; A% 5 androidregion % fat; WC 5 waist circumference; WHR 5 waist-to-hip ratio.

* p , 0.05 within groups.

738 Ultrasound in Medicine and Biology Volume 37, Number 5, 2011

use of US as a measure of abdominal adiposity in personswith SCI. Correlations between VF and BMI, TBF%,TRK% and W% measured by DXA were observed inAB subjects only.

Many measures and indices of body compositionused clinically, such as BMI, anthropometrics andDXA, have limitations. The underestimation of TBF%by BMI calculation in the SCI population has been welldocumented. It is estimated that for any given unit ofBMI, individuals with SCI have a 10–15% greaterTBF% compared with AB counterparts (Spungen et al.2003; Weaver et al. 2007). In our study, we observedthat men with SCI had approximately 10% greaterTBF% than AB men when measured by DXA;however, the BMI values in both groups were similar.We also found that VF measured by US was correlatedwith BMI in the AB group and trended towarda correlation in the SCI. Men with SCI may not beviscerally obese but have a greater amount of fat belowthe level of lesion compared with AB men, whichelevates their overall TBF%. Underestimation may bea result of potential measurement error of body weightand height. Ideally, height is measured while standing;

Table 2. Comparison of body composition measures inable-bodied and SCI subjects

AB SCI

n 5 20 n 5 24

TRK% 32.0 6 10.3 40.0 6 9.6*W% 40.6 6 9.4 47.0 6 9.7*A% 35.8 6 10.6 43.0 6 9.8*WC (cm) 91.3 6 13.6 99.8 6 14.8HC (cm) 99.5 6 10.1 101.2 6 9.2WHR 0.92 6 0.06 0.99 6 0.10*SF (cm) 1.74 6 0.91 1.42 6 0.50VF (cm) 5.39 6 1.90 5.25 6 1.95

Values are expressed as mean 6 SD.AB 5 able-bodied; SCI5 spinal cord injury; TRK% 5 trunk region

% fat (Walton 1995);W%5waist region% fat; A%5 android region%fat; WC 5 waist circumference; HC 5 hip circumference; WHR 5waist-to-hip ratio; SF 5 subcutaneous fat; VF 5 visceral fat.* p , 0.05 between AB and SCI.

however, this is not usually a practical approach toobtaining a height measurement in the SCI population.An alternative method to record height is to measuresupine length in those with SCI. In addition, individualswith chronic SCI have greater fat mass and less fat-freemass per unit BMI than AB persons (Jones et al. 2003;Spungen et al. 2003). Gupta et al. (2006) attempted todescribe the prevalence of overweight and obesity ina population with SCI through retrospective measure-ment of BMI without any other measure of body compo-sition in 387 male and female subjects with paraplegiaand tetraplegia; 28% of the subjects studied had a normalBMI, whereas 66% were classified as overweight orobese. Therefore, individuals with SCI who may havea normal BMI value may actually have unhealthyamounts of total body fat that are associated with beingoverweight or even obese in the general population,which may subject them to greater risk for cardiovasculardisease. Confirming previous reports, the findings fromour current study indicate that the prevalence and magni-tude of obesity is underrepresented by measurement ofBMI alone.

Greater WC measures and WHR values are associ-ated with increased CVD risk and risk factors such as dys-lipidemia, obesity and diabetes, and these measures arerecommended for clinical application to assess CVDrisk (National Heart, Lung, and Blood Institute 1998).Circumference measures have been used to assess bodycomposition in the SCI population. Maki et al. (1995)measured the WCs of 46 men with paraplegia and tetra-plegia with duration of injury greater than 6 months andfound that WC measures correlated directly with serumtriglyceride and indirectly with serum high-density lipo-protein (HDL) cholesterol concentrations. In anotherstudy, Demirel et al. (2001) measured the WCs of 69men and women with paraplegia and tetraplegia and 52age- and gender-matched AB subjects, and did not find

Fig. 4. Relationship between waist circumference (WC) andvisceral fat (VF). WC correlated with VF in able-bodied (AB)men and men with spinal cord injury (SCI) (r 5 0.55, p ,0.01). B represents AB men and : represents men with SCI.The solid line represents men with SCI, and the dashed line

represents AB men.

Adiposity in spinal cord injury d R. R. EMMONS et al. 739

a significant difference in measures between the groups.WC in SCI subjects were, however, associated withhigher plasma glucose, serum total and LDL cholesteroland lower serum HDL cholesterol concentrations.

Results from previous studies support the associa-tion between WC and risk factors for CVD in the SCIpopulation. However, a number of issues pertaining toWC measures need to be addressed. Multiple methodolo-gies and inter- and intra-investigator variation may affectthe validity of this assessment tool (Lee et al. 2008). Themost appropriate measurement site in those with SCI hasyet to be identified, as well as the effects of positioning(supine, sitting, standing), or the effects of potentialconfounding variables in the SCI population, such asloss of muscle tone, abdominal distention and spasticity(Buchholz and Burgaresti 2005). WC measurementsonly provide an indirect index of abdominal adipositybecause it simply measures trunk circumference withno differentiation possible between fat, muscle, bone,and internal organs.

Many studies in the SCI population have usedWC toassess abdominal fatness in relation to CVD risk factors.Liang et al. (2007) reported WC values similar to those inour study. In 185 men with SCI, the prevalence of centralobesity and depressed serum HDL concentration in thesubjects with SCI was greater than the age-matched ABsubject. Conversely, Maruyama et al. (2008) measured

central obesity by CT and WC in 44 male subjects withparaplegia and age-matched AB controls and found thatmen with SCI had greater VF than AB persons, but nodifference in WC was found. Higher measures of VFwere associated with increased incidence of insulin resis-tance and elevated plasma leptin levels, as well as higherlevels of inflammatory markers, such as high-sensitivityC-reactive protein.

In our study, men with SCI had greater TBF%measured by DXA. However, there were no significantdifferences in abdominal fat measures between AB andSCI men. The difference in overall body fatness may bea result of skeletal muscle atrophy below the level oflesion after injury (Maggioni et al. 2003; Spungen et al.2003). Abdominal ROIs by DXA analysis have beencorrelated with abdominal CT and provide a surrogatemeasure of fat in the abdominal region; however, theabdominal region measured by DXA does notdifferentiate VF from SF. Clasey et al. (1999) measuredabdominal fat by CTand DXA in 76 AB men and womenand found that total abdominal regional fat measured byDXA correlated with VF measures by CT. In a study of3075 nondisabled elderly men and women, abdominaladiposity was assessed by DXA and CT. Trunk regionalmeasures by DXA were highly correlated to CT, witha 10% difference between methods (Snijder et al.2002). Glickman et al. (2004) reported a difference ofapproximately 26% when measuring abdominal fat byDXA and CT in 27 AB men and women. These resultsmay differ, in part, because of differences between scan-ning technologies.

US is able to visualize and quantify individual fatlayers and compartments and provide a measurement ofVF thickness. Studies by Ribiero-Filho et al. (2001,2003) and Hirooka et al. (2005) show a high correlationbetween visceral fat measurements acquired by CT andUS in ABmen and women. CT has the capability to deter-mine fat area, and various reports have noted a visceral fatvolume .100 cm2 for ‘‘viscerally obese,’’ although thiscriteria has not yet been standardized in the AB or SCIpopulations (Despres and Lamarche 1993). Because ofthe high costs of performing CT scans and the associatedrisk of ionizing radiation, CT scanning is not a commonor feasible method of measuring visceral fat in a clinicalsetting. Although there currently are no reports on theassociation between CT measures of SF and VF to thosemeasures by US in persons with SCI, our findings suggestthat US may be a clinically useful tool to assess abdom-inal fat in the SCI population.

Over the past 20 years, studies in AB subjects haverepeatedly demonstrated that the amount of VF highlycorrelates to an altered metabolic risk factor profile inmen and women (Despres 2006). Gender and geneticfactors highly influence fat mass distribution and,

740 Ultrasound in Medicine and Biology Volume 37, Number 5, 2011

combined with environmental factors, may influence theprogression to insulin resistance, diabetes mellitus anddyslipidemia (Montague and O’Rahilly 2000). The chal-lenge lies in the accurate determination of VF volumes,assessed clinically, that are associated with CVD riskfactors in persons with SCI, and then to derive a criterionfor treatment of obesity, as well as the secondary medicalconsequences associated with an increased risk of CVD.Thus, abdominal US is easy to perform and a usefulmethod to quantitate VF mass in persons with SCI, andit may prove to be of value clinically in the assessmentof cardiac risk.

LimitationsInter-investigator variability is one of the major

limitations of US. To overcome this limitation, sonogra-phers at both study locations were trained simultaneouslyand periodically monitored for measurement reliability.Visualization of the anatomical landmarks was chal-lenging in some subjects because of gas bubbles in theabdominal cavity. In some subjects with SCI, landmarkswere difficult to find because of the nature of the injury(i.e., stabilization hardware, scar tissue, scoliosis, etc.).In some subjects, differentiation between fatty andnonfatty tissue was challenging, which would influencethe overall thickness.

One of the main disadvantages of US for themeasurement of abdominal adiposity compared withother diagnostic modalities is that US is a simplemeasurement of thickness of adipose tissue in one loca-tion, compared with DXA, CT, and MRI, that has thecapacity to measure cumulative fat area. US does notcontrol for abdominal fat distribution, like the otherimaging techniques and WC. Although the VF amountswere similar, WC measures trended to be greater, andabdominal fat measures by DXA were significantlygreater in the SCI group; suggesting that fat may bedistributed differently in those with SCI compared withAB subjects. Currently, there are no diagnostic cutoffvalues representing obesity or CVD risk by visceral fatmeasured by US. Although the cost associated with USis far less than that associated with CT and MRI, US isstill more labor intensive and expensive to perform thananthropometrics, skinfolds and bioimpedance, but it canprovide valuable clinical information that these methodscannot. However, US may not easily be performedoutside of a clinical setting.

CONCLUSION

Abdominal US may be a useful tool in clinical prac-tice for the measurement of visceral fat in persons withSCI, and it would be anticipated to provide useful infor-mation in addition to that obtained with anthropometrics

and DXA. US may be of value in monitoring changes invisceral fat in individuals participating in weight lossprograms because of the cost-effectiveness and ease ofuse of US compared with that of CT and MRI. However,prospective epidemiological studies are needed in the SCIpopulation to quantitate visceral fat mass and developcut-off values by CT, MRI and US that predict cardiovas-cular risk.

Acknowledgment—This work was supported by the Department ofVeterans Affairs Rehabilitation Research & Development Center ofExcellence for the Medical Consequences of Spinal Cord Injury(#B4162C).

REFERENCES

Bauman WA, Spungen AM. Coronary heart disease in individuals withspinal cord injury: Assessment of risk factors. Spinal Cord 2008;47:466–476.

Bertin E, Marcus C, Ruiz JC, Eschard JP, Leutenegger M. Measurementof visceral adipose tissue by DXA combined with anthropometry inobese humans. Int J Obes 2000;24:263–270.

Bray GA. Medical consequences of obesity. J Clin Endocrinol Metab2004;89:2583–2589.

Buchholz AC, Burgaresti J. A review of body mass index and waistcircumference as markers of obesity and coronary heart diseaserisk in persons with chronic spinal cord injury. Spinal Cord 2005;43:513–518.

Clasey JL, Bouchard C, Teates CD, Riblett JE, Thorner MO,Hartman ML, Weltman A. The use of anthropometric and dual-energy X-ray absorptiometry (DXA) measures to estimate totalabdominal and abdominal visceral fat in men and women. ObesRes 1999;7:256–264.

Demirel S, Demirel G, Tukek T, Erk O, Yilmaz H. Risk factors for coro-nary heart disease in patients with spinal cord injury in Turkey.Spinal Cord 2001;39:134–138.

Despres JP. Is visceral obesity the cause of the metabolic syndrome?Ann Med 2006;38:52–63.

Despres JP, Lamarche B. Effects of diet and physical activity onadiposity and body fat distribution: Implications for the preventionof cardiovascular disease. Nutr Res Rev 1993;6:137–159.

Gater DR. Obesity after spinal cord injury. PhysMed Rehabil Clin NorthAm 2007;18:333–351.

Glickman SG, Marn CS, Supiano MA, Dengel DR. Validity and reli-ability of dual-energy X-ray absorptiometry for the assessment ofabdominal adiposity. J Appl Physiol 2004;97:509–514.

Guldiken S, Tuncbilek N, Okten OO, Arikan E, Tugrul A. Visceral fatthickness determined using ultrasonography is associated withanthropometric and clinical parameters of metabolic syndrome. IntJ Clin Pract 2006;60:1576–1581.

Gupta N, White KT, Sandford PR. Body mass index in spinal cordinjury—A retrospective study. Spinal Cord 2006;44:92–94.

Hirooka M, Kumagi T, Kurose K, Nakanishi S, Michitaka K,Matsuura B, Horiike N, Onji M. A technique for the measurementof visceral fat by ultrasonography: Comparison of measurementsby ultrasonography and computed tomography. Int Med 2005;44:794–799.

Jones LM, LeggeM, Goulding A. Healthy body mass index values oftenunderestimate body fat in men with spinal cord injury. Arch PhysMed Rehabil 2003;84:1068–1071.

Kocina P. Body composition of spinal cord injured adults. Sports Med1997;23:48–60.

Lee K, Lee S, Kim Y, Kim Y. Waist circumference, dual-energy X-rayabsorptiometry measured abdominal adiposity, and computed tomo-graphically derived intra-abdominal fat area on detecting metabolicrisk factors in obese women. Nutrition 2008;24:625–631.

Liang H, Chen D, Wang Y, Rimmers JH, Braunschweig CL. Differentrisk factor patterns for metabolic syndrome in men with spinal

Adiposity in spinal cord injury d R. R. EMMONS et al. 741

cord injury compared with able-bodied men despite similar preva-lence rates. Arch Phys Med Rehabil 2007;88:1198–1204.

Maggioni M, Bertoli S, Margonato V, Merati G, Veicsteinas A,Testolin G. Body composition assessment in spinal cord injurysubjects. Acta Diabetol 2003;40:S183–S186.

Maki KC, Briones ER, Langein WE, Inman-Felton A, Nemchausky B,Welch M, Burton J. Associations between serum lipids and indica-tors of adiposity in men with spinal cord injury. Paraplegia 1995;33:102–109.

Maruyama Y, Mizuguchi M, Yaginuma T, Kusaka M, Yoshida S,Yokoyama K, Kasahara Y, Hosoya T. Serum leptin, abdominalobesity, and the metabolic syndrome in individuals with chronicspinal cord injury. Spinal Cord 2008;46:494–499.

Merino-Ibarra E, Artieda M, Cenarro A, Goicoechea J, Calvo L,Guallar A, Civeira F. Ultrasonography for the evaluation of visceralfat and the metabolic syndrome. Metab Clin Exp 2005;54:1230–1235.

Montague CT, O’Rahilly S. The perils of portliness: Causes and conse-quences of visceral adiposity. Diabetes 2000;49:883–888.

National Heart, Lung, and Blood Institute. Clinical Guidelines on theIdentification, Evaluation, and Treatment of Overweight andObesity in Adults. Washington, DC: National Heart, Lung, andBlood Institute; 1998.

Ribeiro-Filho FF, Faria AN, Azjen S, Zanella M, Ferreira SR. Methodsof estimation of visceral fat: Advantages of ultrasonography. ObesRes 2003;11:1488–1494.

Ribeiro-Filho FF, Faria AN, Kohlmann O, Ajzen S, Ribeiro AB,Zanella M, Ferreira SR. Ultrasonography for the evaluation ofvisceral fat and cardiovascular risk. Hypertension 2001;38(part 2):713–717.

Snijder MB, Visser M, Dekker JM, Seidell JC, Fuerst T, Tylavsky F,Cauley J, Lang T, Nevitt M, Harris TB. The prediction of visceralfat by dual-energy X-ray absorptiometry in the elderly: A compar-ison with computed tomography and anthropometry. Int J Obes2002;26:984–993.

Spungen AM, Adkins RH, Stewart RA,Wang J, Pierson RN,Waters RL,Bauman WA. Factors influencing body composition in persons withspinal cord injury: A cross-sectional study. J Appl Physiol 2003;95:2398–2407.

Walton C, Lees B, CrookD,WorthingtonM, Godsland IF, Stevenson JC.Body fat distribution, rather than overall adiposity, influences serumlipids and lipoproteins in healthy men independently of age. Am JMed 1995;99:459–464.

Weaver FM, Collins EG, Kurichi J, Miskevics S, Smith B, Rajan S,Gater DR. Prevalence of obesity and high blood pressure in veteranswith spinal cord injuries and disorders. Am J Phys Med Rehabil2007;86:22–29.