Ultrasound as an approach to assessing body corn position 14

Transcription

1 Ultrasound as an approach to assessing body corn position 14 Marie T Faneii, PhD, RD and Robert J Kuczmarski, MPH, RD ABSTRACT Body composition is an important indicator of nutritional status. The most commonly used indirect method for estimating body fat is based on measurements of subcutaneous fat tissue. It has been suggested that ultrasonic measurements may be more precise than those of the caliper and therefore may yield more accurate measures of subcutaneous fat tissue. This study was designed to correlate ultrasonic and caliper measurements of subcutaneous fat with body density determined by hydrostatic weighing. Subcutaneous fat thickness was measured at seven body sites (triceps, biceps, subscapula, waist, suprailiac, thigh, and calf) with a Lange skinfold caliper and an ADR ultrasonic scanner, equipped with a display-screen, 7MHz transducer, and electronic calipers. Regression equations to predict body density, and hence body fat, were derived for each technique using a minimal number of body sites. The sample consisted of 124 white men, aged 18 to 30 yr. Mean body density determined by hydrostatic weighing was 1.07 g/ml (SD ± 0.01) and mean body fat was 12.7% (SD ± 5.8). Both ultrasonic and caliper measurements of waist, thigh, and triceps had the highest correlation with body density. Regression equations using these three sites in all possible two-site combinations were derived for each technique. The predictions of body density from these equations did not differ significantly. These results suggest that in free-living, nonobese, white men, body fat can be estimated with nearly the same degree of accuracy using either the caliper or ultrasonic technique. Am J Clin Nuir l984;39: KEY WORDS Body composition, densiotometry, ultrasonics Introduction Accurate measurements of body composition are essential for evaluating nutritional status and for planning the dietetic aspects of health care programs. Among the methods to estimate body fat, hydrostatic weighing is perhaps the most accurate, but also the most cumbersome, noninvasive technique currently available. A more practical method for both field surveys and routine clinical assessment is the measurement of subcutaneous fat thickness. This measurement is most commonly taken with a caliper as a double fold of skin and subcutaneous fat tissue, ie, skinfold. For lean subjects, caliper measurements of skinfolds correlate highly with direct measures of skin and subcutaneous fat thickness obtained by incisions (r = 0.83) (1) and by electrical conductivity (r = 0.81) (2). As with most techniques, there are limitations associated with the caliper method which may result in inaccurate estimates of subcutaneous fat thickness and, consequently, of total body fat. These limitations include the inability to control inter- and intrasubject variation in skinfold compressibility, the inability to palpate the fat-muscle interface, and the impossibility of obtaining interpretable measurements on very obese subjects (2-6). Variation in skinfold compression has been attributed to such factors as subcutaneous fat thickness, state of From the Department of Nutrition, School of Public Health, University of North Carolina, Chapel Hill, NC. in part by University Research Council Grant Funds and School of Public Health Biomedical Research Support Funds, University of North Carolina, Chapel Hill, NC. Presented in part at the Western Hemisphere Nutrition Congress VII, Miami, FL, August 10, Address reprint requests to: Marie T Fanelli, Department of Nutrition, 315 Pittsboro Street 325H, University of North Carolina, Chapel Hill, NC Received October 7, Accepted for publication November 29, The American Journal of Clinical Nutrition 39: MAY 1984, pp Printed in USA 1984 American Society for Clinical Nutrition 703

2 704 FANELLI AND KUCZMARSKI hydration, and distribution of fibrous tissue and blood vessels (3, 7). Ultrasound has been proposed as an alternative noninvasive technique to measure subcutaneous fat thickness because it may overcome some of the limitations of the caliper (6, 8, 9). Ultrasonic scanners are capable of measuring subcutaneous fat at depths of 100 mm or more without tissue compression and can reliably detect density interfaces with an accuracy of 1 mm. In the 1960 s and early 1970 s, ultrasonic measurements were demonstrated to be strongly correlated with direct measures of subcutaneous fat by electrical conductivity (r = 0.98) (2), needle puncture (r = 0.98) (10), and softtissue radiographs (r = 0.88) (6). These data suggested that ultrasonic measurements of subcutaneous fat may be more accurate than caliper measurements. Over the years, research defining the practical applications of ultrasound in anthropometry has not kept pace with advances in the electronics of ultrasonic technology. A possible explanation is that traditionally this equipment was limited to use in hospitals where tissue resolution at near fields was not essential. It is only recently that high resolution transducers required for characterizing subcutaneous tissues have become available. The purpose of this investigation was to describe the relative merits and validity of ultrasound with respect to two other noninvasive techniques for estimating body density. Measures of subcutaneous fat obtained with a skinfold caliper and ultrasonic scanner were compared to body density determined by hydrostatic weighing. The following report presents 1) the correlations between caliper and ultrasonic measurements of subcutaneous fat at specified anatomical sites; 2) the correlations of both the caliper and the ultrasonic measurements of subcutaneous fat with densitometrically determined body density; and 3) regression equations developed to predict body density and subsequently, total body fat from caliper and from ultrasonic measurements of subcutaneous fat. Methods Subjects In response to advertisements placed in municipal and university newspapers, 124 white men, aged 18 to 30 yr. volunteered to participate in this study. Informed consent forms were signed by each subject before participation. All of the procedures used were in accordance with the ethical standards of the Institutional Board for Experimentation with Human Subjects of the School of Public Health at the University of North Carolina at Chapel Hill. Procedures Anthropometric measurements were taken for each subject dressed only in a swim suit in the following sequence: measurements of subcutaneous fat by caliper and then by ultrasound, stature measurement, weight measurement, and finally hydrostatic weighing. Subcutaneous fat thickness was measured by caliper and ultrasonic techniques at seven sites on the right side of the body. The sites and their anatomical descriptions are as follows: triceps, midpoint between the acromion and olecranon processes on the posterior aspect of the arm; biceps, midpoint of muscle belly; subscapula, inferior angle of the scapula; waist, midpoint between last rib and iliac crest at the midaxillary line; suprailiac, oblique fold on the iliac crest at the midaxillary line; thigh, anterior aspect of the thigh midway between the inguinal fold and midpoint of the patella; and calf, posterior aspect of the lower leg at the maximal girth. Each site was marked with a wax-based cosmetic pencil. To eliminate interobserver error, one investigator read all the caliper measurements and another read all the ultrasonic measurements. For both techniques, two independent measurements were taken at each site, and values were recorded to the nearest 0.5 mm. If the values were not within 1 mm of each other, a third measurement was taken. The two values that agreed within 1 mm were averaged and the mean was accepted as the representative value. All skinfolds were measured with a Lange skinfold caliper (Cambridge Scientific Instruments, Cambridge, MD) which was checked against a calibration gauge before each use. Skinfolds were taken in accordance with a standard anthropometric procedure (11). An ADR Real Time Scanner, model 2130, equipped with a 7.0 MHz transducer, display-screen, freeze-frame capacity, and electronic calipers (ADR Ultrasound, Tempe, AZ) was used to measure subcutaneous fat thickness (excluding skin thickness) at the same marked sites as skinfolds. A water-soluble transmission gel was applied to the array of the transducer. The transducer was then held manually about 0.5 mm above the marked site with the gel providing acoustic contact without depression of the dermal surface. To assure accurate depth readings the transducer was positioned until the ultrasonic beam was perpendicular to the tissue interfaces at each site. An angle of incidence other than 900 may result in a transmission parallax error (12). As noted, the ADR scanner is equipped with a Hewlett Packard display-screen for visual presentation of ultrasonic images. Using the near and far field gain controls, the image was focused and then frozen. Subcutaneous fat thickness was measured directly from the screen with the use of electronic calipers positioned at the skinfat and fat-muscle interfaces (Fig 1). Fat thickness depths less than 3 mm were measured directly from the

3 ESTIMATES OF BODY FAT FROM ULTRASONIC MEASURES 705 or swim suit. Underwater weighings were continued until the heaviest weight reading was repeated twice. This value was used in the calculation of body density. Water temperature was recorded immediately after each subject was weighed. Residual lung volumes were measured independently in a plethysmograph (iaeger Body Test, Germany) at the Pulmonary Function Laboratory of North Carolina Memorial Hospital, Chapel Hill. Body density (Db) was calculated from the formula of Goodman and Buskirk (15): Wa Db - Wa-Ww - V. Where: Wa = weight of subject in air, kg; Ww = weight of subject in water minus weight of equipment, kg; Dw = density of water at X#{176}C, g/ml; V = residual lung volume, I. Percentage of body fat was calculated from body density according to the formula of Sin (16): % body fat = (4.95/Db ) x 100. Dw Statistical analysis FIG 1. Thigh site as measured by ultrasound. Arrows indicate position of electronic calipers. Arrow I indicates skin-fat interface; arrow 2, fat-muscle interface; and arrow 3, muscle-bone interface. The left highlights the fat layer between skin and muscle (11 mm) while the right points out the muscle between subcutaneous fat and the femur (31 mm). screen with a draftsman s caliper (HB, Germany), because the electronic calipers could not be positioned to measure to these depths. Vertical and horizontal calibration of the equipment was routinely checked according to the manufacturer s instructions (13). The percentage compression associated with the caliper technique was determined for each body site. Percentage compression was calculated by the following equation which was developed for this study: U-0.5C % compression = x 100 where: U = mean fat thickness (mm) measured by ultrasound; C = mean skinfold thickness (mm) measured by caliper. Stature was measured against a vertical height board to the nearest 0.1 cm. Each subject stood with feet flat and eyes looking straight forward, keeping his heels, buttocks, and shoulders in contact with the board while the headboard was lowered and stature recorded. All men were weighed to the nearest 0.1 kg on a double beam balance. The accuracy of this balance was checked and verified with standard test weights. Hydrostatic weighing was conducted according to the procedures described by Pollock et al (14). Each man entered the hydrostatic weighing tank and was completely submerged. Precautions were taken to assure that no air was trapped by the subject s body hair Mean values, SDs, and ranges for age, height, weight, body density, residual lung volume, percentage body fat, percentage compression, and caliper and ultrasonic measurements of subcutaneous fat at each body site were calculated. Using the Statistical Analysis System package for zero-order Pearson correlation analysis for the null hypothesis, Ho = 0, coefficients of correlation between body density, as determined by hydrostatic weighing, and subcutaneous fat thickness measured at each body site with caliper and ultrasound were also calculated(17). Multiple regression analyses, using body density as the dependent variable, and subcutaneous fat thickness as the independent variables, were performed separately for caliper and ultrasonic measurements (17). Regression analysis was used to identify the body sites that provided the best estimation of body density from caliper and from ultrasonic measurements of subcutaneous fat and to develop equations for predicting body density from these selected sites. Results The physical characteristics of the men are presented in Table 1. Mean body density was 1.07 g/ml which by Siri s formula, corresponds to a mean body fat of 12.7%. The frequency distribution of the body density values is reported in Table 2. Approximately 53% (n = 23) had more than 17% body fat. Thus, this sample was comprised mainly of relatively lean men. Mean values of subcutaneous fat measured by the caliper and ultrasonic techniques and correlation coefficients between the measurements are presented in Table 3. Measurements by the two techniques corre-

4 706 FANELLI AND KUCZMARSKI TABLE 1 Physical characteristics of subjects (n = 124) Variable Age (yr) Ht(cm) Wt(kg) Body density (g/ml)* Residual lung volume (1) Body fat(%)t * Derived from hydrostatic weighing. Mean SD Range t Calculated from Siri s formula, using the body density value derived from hydrostatic weighing. TABLE 2 Frequency distribution of body density and percentage body fat Body densities g/m! Total sample Body % fatt * Derived from hydrostatic weighing. t Calculated from Sin s formula. Triceps 10.1 ± ± t Biceps 3.6 ± ± 1.2 O.703t Subscapula 10.5 ± ± t Waist 16.0 ± ± t Suprailiac 15.1 ± ± 3.8 O.734t Thigh 11.4± ± t Calf 8.9 ± ± t * Caliper measurement is the doublefold of skin and subcutaneous fat, while ultrasonic measurement is the single thickness of subcutaneous fat. t p < n Frequency TABLE 3 Mean subcutaneous fat thickness (mm) as determined by caliper and ultrasonic techniques and correlations of measurements by technique and site Site Caliper0 Ultrasound x±sd lated significantly (p < ) at all sites. The measurements from the caliper and ultrasonic techniques taken over the triceps, waist, and thigh sites correlated more highly than those taken over the other sites. Compression generally ranged between 10 and 40% for fat thickness values recorded with the caliper for each body site. Of the seven sites measured, the thigh, waist, and triceps sites demonstrated the least amount of variance in percentage compression. The data were examined to determine whether there was greater compressibility with increasing skinfold thickness. No consistent trends were detected. To determine which body sites, by technique, were the most accurate in predicting total body fat, the subcutaneous fat thicknesses obtained by caliper and ultrasound at each site were individually compared with body density derived from hydrostatic weighing (Table 4). All correlations were negative and highly significant (p <0.0001). The caliper measurements of subcutaneous fat for most sites had slightly higher correlations with body density than with ultrasound. Of the sites measured, the triceps, waist, and thigh sites demonstrated the strongest correlations with body density, regardless of the measurement technique. For the caliper technique, the triceps site appeared to be the best single predictor of body density (r = 0.749); for the ultrasonic technique, the waist was the best single predictor (r = 0.736). Two different procedures were used to develop regression equations to predict body density. First, forward stepwise multiple regression analysis was used with body density as the dependent variable and subcuta- TABLE 4 Correlation coefficient (r) between body density and subcutaneous fat thickness measured at seven body sites with caliper and ultrasound Site Caliper Ultrasound Triceps Biceps Subscapula * Waist Suprailiac Thigh Calf -0.59l * p <

5 ESTIMATES OF BODY FAT FROM ULTRASONIC MEASURES 707 neous fat thickness at each of the seven sites as the independent variable. For both the caliper and ultrasonic techniques, there was a small improvement in the correlation coefficient values with the inclusion of a second body site in the equation. Using three or more body sites did not significantly increase the accuracy of the prediction of body density over the use of two sites. Therefore, it appears that values of subcutaneous fat at two body sites are adequate for the prediction of body density. In this study of relatively lean, healthy men, the two skinfolds among the seven measured that provided the best prediction of body density were the triceps and waist. The multiple correlation using these two sites together was r = and the unbiased standard error of estimate (SEE) was (Table 5). Ultrasonic measurements of subcutaneous fat at the waist and thigh sites gave the best prediction of body density. The multiple correlation using two sites simultaneously was r = and the SEE = (Table 5). Thus, the best prediction for the ultrasonic technique was slightly better than that of the best prediction for the caliper technique, as indicated by the slightly higher r value and slightly lower SEE value. The data were subsequently analyzed using the SAS General Linear Models Procedure, with body density serving as the dependent variable and subcutaneous fat thickness at two designated sites as the independent variables. The sites selected for this model were the triceps, waist, and thigh because they individually demonstrated strong correlations with body density. The formulas for prediction of body density from the various two-site combinations along with their r and SEE values are presented in Table 5. The mean values for percentage body fat generated from the formulas using caliper measurements are not significantly different from those derived from ultrasound measurements. Correlation coefficients were calculated to evaluate the relationship between the predicted density values by measurement technique for each two-site combination. The correlations were r = for triceps-waist sites, r = for thigh-waist sites, and r = for thigh-triceps sites. These findings suggest that the predictive accuracy of the six formulas is very similar. Discussion The mean percentage body fat calculated from Siri s formula using densitometrically derived body density was compared to previously reported mean values for men of similar ages. The mean of 12.7% body fat in the present study was less than the mean values of 15.0, 14.6, and 13.4% reported by Durnin and Womersley (18), Wilmore and Behnke (19), and Pollock et al (14), respectively, but exceeded the mean value of 10.3% reported by Sloan (20). These inconsistencies may be associated with the overall state of physical fitness of the subjects represented in the various study samples. TABLE 5 Multiple regression equations for prediction of body density (Db) using caliper or ultrasonic measurements of subcutaneous fat Multiple regression equations#{176} r SEE Body fart (04) x ± SD A. Caliper measurements (mm) = X Db = (X,) (X2) Db = (X2) (X3) Db = (X1) (X3) B. Ultrasonic Measurements (mm) = Y Db = (Y2) (Y3) Db = (Y) (Y2) Db = (Y1) (Y3) * Key to variables in equations: I = triceps; 2 = waist; 3 = thigh ± ± ± ± ± ± 4.7 t Calculated from Siri s formula using predicted body density value. Mean body fat calculated from Siri s formula using body density derived from hydrostatic weighing equals 12.7% (SD ± 5.8%).

6 708 FANELLI AND KUCZMARSKI Subcutaneous fat thickness measured by the caliper technique had slightly higher correlations with body density when compared with the ultrasonic technique for five of the seven sites. Sloan (20) also found body density to be more highly correlated with the caliper technique than with ultrasound. The correlation coefficient values reported by Sloan for both techniques were somewhat higher than those reported here. Nevertheless, in the present study, ultrasonic measurements showed good agreement with caliper measurements. For the triceps site, the correlation between these two techniques (r = 0.807) was higher than that observed by Haymes et al (r = 0.64) (6) but similar to that reported by Bullen et al (r = 0.80) (10). The mean subcutaneous fat thicknesses measured by ultrasound were greater than one-half of the mean caliper values, indicating a compression effect. The appropriateness of skinfold measurements in the assessment of nutritional status needs further investigation if the caliper technique is shown to significantly underestimate actual fat thickness. For this sample of white men, the variation in compression over the triceps, waist, and thigh sites did not appear to be a source of error in the prediction of body density. However, this may not be true for other populations. Therefore, more research is needed to determine the extent to which skinfold compression makes an appreciable difference in estimating body fat in samples of nonlean individuals. The multiple correlation coefficients for caliper measurements from two body sites with body density compared favorably with those from other studies. Similar body density predictions were obtained by using any one of the three possible skinfold combinations presented in Table 5. Previous studies have reported the following combinations to yield the highest multiple correlation coefficients: thigh and abdomen (18), thigh and chest (14), and thigh and subscapula (20). It should be noted that all of these combinations include an extremity and trunk site. It has been suggested that such combinations are good predictors of anatomical fat distribution which may be an important diagnostic determinant of one s susceptibility to chronic diseases (21). Even more interesting is the absence of the triceps site from these combinations, recalling that the triceps site is continually reported to be the best single predictor of body density and overall adiposity (16, 22). The regression equation derived in this study using the ultrasonic measurements of subcutaneous fat at the thigh and waist sites predicted body density as accurately as Sloan s equation using ultrasonic measurements at the suprailiac and thigh sites (20). This is evidenced by the almost identical multiple correlation coefficients, that is r = in this report and r = reported by Sloan. The equation derived in the present study was cross-validated on the data from men examined by Sloan. This equation yielded the same value for body density that Sloan had obtained by hydrostatic weighing. This was expected since the ages and physical characteristics of Sloan s population were similar to that of the present sample. Therefore, it appears that the formula developed from this research may be used interchangeably with Sloan s equation for other samples of white men, aged 18 to 30 yr, with body density values ranging between 1.03 and 1.10 g/ml. Other studies involving different samples of subjects are encouraged and needed to test the formula s general applicability. The results of this study differ from those of Borkan et a! (23) who found the Lange skinfold caliper to be more effective than ultrasound in assessing subcutaneous fat. A partial explanation may be that the Body Composition Meter used by Borkan et al was relatively less sensitive than the ADR Real Time Ultrasonic Scanner used in this study. The readout from the Body Composition Meter uses a series of light-emitting diodes spaced at 1-mm increments. In theory, a single diode should appear at each interface. However, it has been found that in practice many diodes may be lit at once which introduces measurement errors and uncertainties (Fanelli M, unpublished observations). Future development of ultrasonic devices should include a design that helps the user to more easily identify the wave reflected from a tissue interface. A displayscreen such as that found on the ADR scanner helps minimize measurement error be-

7 ESTIMATES OF BODY FAT FROM ULTRASONIC MEASURES 709 cause it enables the investigator to identify visually these tissue interfaces. With regard to near field chatter with ultrasound, a recent communication suggests that much of these near field effects has been eliminated through advances in electronic design (24). The findings from this study suggest that the caliper and ultrasonic techniques are equally effective in predicting body density and, hence, total body fat of lean men. Both techniques can be performed with minimal inconvenience to the subjects. In addition to its ability to assess subcutaneous fat tissue, the ultrasonic method permits direct measurements of muscle tissue. Furthermore, permanent records of underlying fat and muscle thicknesses can be obtained with ultrasonic scanners equipped with camera attachments. The ability to monitor and document changes in fat and muscle tissues would enable clinicians to assess, and then if necessary, modify diet and nutritional support of patients. Reviewing these pictorial records with patients may well improve their compliance with prescribed therapy. fl The authors thank Joy Wood for her assistance with the computer programming of statistical analyses, Dr. Robert McMurray of the UNC Exercise Physiology Department for use of his laboratory facilities, and the ADR Ultrasound Company for the loan of its equipment. References 1. Lee MMC, Ng CK. Postmortem studies of skinfold caliper measurement and actual thickness of skin and subcutaneous tissue. Hum Biol 1965;37: Booth RAD, Goddard BA, Patton A. Measurement of fat thickness in man: a comparison of ultrasound, Harpenden calipers and electrical conductivity. Br J Nutr I 966;20: Himes JH, Roche AF, Siervogel RM. Compressibility of skinfolds and the measurement of subcutaneous fatness. Am J Clin Nutr 1979;32:l Garrow is. New approaches to body composition. Am J Clin Nutr l982;35: Brozek J, Kinzey W. Age changes in skinfold compressibility. J Gerontol 1960; 15: Haymes EM, Lundergren HM, LoomisiL, Buskirk ER. Validity of the ultrasonic technique as a method of measuring subcutaneous adipose tissue. Ann Hum Biol l976;3:245-5l. 7. Clegg EJ, Kent C. Skinfold compressibility in young adults. Hum Biol l967;39: Sanchez CL, Jacobson HN. Anthropometry measurements, a new type. Am J Clin Nutr 1978;3l: Whittingham PDGV. Measurement of tissue thickness by ultrasound. Aerospace Med I 962;33: Bullen BA, Quaade F, Olesen F, Lund SA. Ultrasonic reflections used for measuring subcutaneous fat in humans. Hum Biol l965;37: National Center for Health Statistics. HANES II. Examination staff procedures manual for the health and examination survey, Rockville, MD: National Center for Health Statistics, Donald I, Brown TG. Demonstration of tissue interfaces within the body by ultrasonic echo sounding. Br i Radiol 1961 ;34: ADR Real Time Ultrasound Scanner, model 2130, operator s manual. Tempe, AZ: ADR Ultrasound, Pollock ML, Hickman T, Kendrick Z, Jackson A, Linnerud AC, Dawson G. Prediction of body density in young and middle-aged men. J AppI Physiol l976;40: Goldman RF, Buskirk ER. Body volume measurement by underwater weighing: description of a method. In: Brozek I, Henschel A, eds. Techniques for measuring body composition. Washington, DC: National Academy of Science, 196 1: Lohman TF. Skinfold and body density and their relation to body fatness: a review. Huth Biol 198 l;53: Ray AA, ed. SAS s user s guide: Statistics ed. Cary, NC: SAS Institute Inc Durnin JVGA, Womersley J. Body fat assessed from total body density and its estimation from skinfold thickness: measurements on 481 men and women aged 16 to 72 years. Br J Nutr l974;32: Wilmore JH, Behnke AR. An anthropometric estimation of body density and lean body weight in young men. J AppI Physiol l969;27: Sloan AW. Estimation of body fat in young men. i AppI Physiol l967;23:3l Mueller WH, Stallones L. Anatomical distribution of subcutaneous fat: skinfold site choice and construction of indices. Hum Biol 198l;53:32l Seltzer CC, Mayer J. Greater reliability of the triceps skinfold over the subscapular skinfold as an index of obesity. Am J Clin Nutr 1967;20: Borkan GA, Hults DE, Cardarelli i, Burrows BA. Comparison of ultrasound and skinfold measurements in assessment of subcutaneous and total fatness. Am J Phys Anthropol l982;58:307-l Toyokawa H, Kimura N, Marui E. Validity in identification of the target wave at thigh. Jap J Public Health (in press).