Thiopental sodium

ABSTRACT. Background: A piglet model was used to determine the influence of frequently encountered situations in clinical studies of infants and young children on fan-beam dual-energy x-ray absorptiometry (DXA) measurements. Methods: DXA scans of piglets (640 g to 21,100 g) were acquired in the infant and adult mode and were analyzed with 1 infant, 1 pediatric, and 3 versions of adult software. Results: The effect of repositioning of the piglets from the center to the periphery of the scanning table on DXA measurements included an average difference of up to 0.5% for total weight, 5.0% for bone mineral content, 5.6% for bone mineral density, 1.3% for lean mass, and 21.9% for fat mass (p 10 kg allow systematic corrections of data from different scan modes and different software. Conclusions: Attention to details and consistency in the technique for scan acquisition and analysis are critical to the generation of meaningful data and to allow for detection of true differences in DXA measurements of small subjects. (Journal of Parenteral and Enteral Nutrition 28:328-333, 2004)

Dual energy x-ray absorptiometry (DXA) is generally accepted as the standard for measurement of bone mass and is the preferred means of measuring soft tissue composition in small subjects.1 The newer generation of fan-beam DXA technique was recently validated for use in small subjects, with documented significant advantages over the current pencil-beam DXA in the reduction of scan acquisition time and improved accuracy of bone mass and body composition measurements.2,3 We have documented that for the pencilbeam DXA technique, several practical issues associated with scan acquisition and analysis may lead to spurious results.4,5 However, it is not known whether the technical issues that adversely affected pencilbeam DXA measurements also might affect the validity of fan-beam DXA measurements because fan-beam DXA differs significantly in the distribution of photon beams, number and distribution of detectors, scan paths in the generation of the DXA images, and in the software algorithm for scan analysis. This study therefore aims to define the influence of a number of practical issues that may occur during data acquisition and analysis for fan-beam DXA-measured body composition in small subjects.

MATERIALS AND METHODS

Animals

Commercial domestic swine piglets (J&M Farms, Lansing, MI) were studied to determine the influence of variations in scan acquisition or analysis that may be encountered during clinical and experimental situations on DXA measurements. Each piglet and any covering (cotton blanket and diaper) used during the scan were weighed separately using an electronic scale (seca; Toledo Scale Company, Toledo, OH).

All piglets were studied as part of the comprehensive protocol on various aspects of body composition measurement, although the number of piglets scanned was varied, depending on the specific aspect studied. The study protocol was approved by the institutional review board for animal investigations at Wayne State University, Detroit, MI.

DXA Studies

Scan acquisition used a fan-beam densitometer (Hologic QDR4500A; Hologic Inc, Bedford, MA) in the infant and adult whole-body scan mode. Unless otherwise indicated, all piglets were scanned in the prone position with the extremities extended. Each piglet was placed at the center parallel to the long axis of the scanning table, with the head about 5 cm from the short axis. Piglets were sedated with pentobarbital and sodium thiopental. All scan acquisitions and analyses were performed according to the manufacturer's recommendations. Quality control scans were performed according to the manufacturer's recommendation using a manufacturer-supplied anthropomorphic spine phantom, and the long-term (>4 years) coefficients of variation for the determination of bone mineral content, bone area, and bone mineral density are 0.42%, 0.33%, and 0.31%, respectively.

The same investigator (MH) supervised all aspects of scan acquisition and analysis and determined that all scans were free from movement artifacts.4,5 All piglets were scanned with commercial infant whole body software (Hologic Inc) further modified, validated and cross-validated by the investigators as version vKH6. These procedures were performed using chemical analysis of piglets after DXA scanning.2,3 In vivo precision calculated according to the method of Gluer et al6 for DXA measured total weight, bone mineral content, bone area, bone mineral density, lean mass, and fat mass using the infant whole-body software vKH6 were 0.2%, 1.9%, 1.4%, 1.7%, 1.0%, and 13.1% respectively, from duplicate scans of 40 piglets with weights between 600 g and 21,100 g.

For one group of piglets, additional scans were performed using a commercial adult whole-body (AWB) software version 8.26a:3 (Hologic Inc), although the scans were analyzed with 2 other versions of AWB software (8.24a:3 and 8.24a:3*) and a commercial experimental pediatrics software (8.24a:3*) from the same manufacturer.

The fan-beam DXA technique consists of multiple photon beams passing through a slit collimator coupled to a multidetector linear array. This allows the acquisition of scan images with 3 or 4 parallel sweeps. For the densitometer used in this study, the scan acquisition occurs in 3 parallel sweeps where both the table and the C arm move along the longitudinal axis of the scanner. To test the effect on fan-beam DXA measurements from variations in the location of the subject in relation to the scanning paths, a white sheet of paper covering the width of the total scan field was placed on the DXA table. The midpoint at the width was marked on the paper, and 2 lines were drawn, dividing the paper longitudinally into 3 zones. The middle zone was 19 cm wide, as determined by preliminary scanning using inanimate objects. The width of 2 outer zones was about 24 cm each. The outer zone closest to the C arm at the home position of the scanner was assigned as path A, middle zone was assigned path B, and opposite outer zone was assigned path C. Three points (midpoint and left and right quarters) were marked on each piglet after measuring the distance between the shoulders with an adjustable caliper. For this series of scans, each piglet was placed in path B, making sure that the midpoint mark on paper and midpoint mark on the animal were aligned, and the resultant scans were used as reference scans to determine the effect of the subject location on DXA measurements. Additional scans for the same piglets were performed with progressively greater proportion (25%, 50%, and 75%) of each animal, shifting from path B to path A. The scans were repeated on the same group of piglets with the same proportion of each animal, shifting from path B to path C.

The effect of the posture of the subject was tested with piglets scanned in the prone, supine, and at the side with left and right limbs superimposed. The effect of layers of covering was tested with piglets wearing a diaper and covered with 1, 3, and then 5 cotton blankets. Additional scans were performed on piglets with no covering, with 1 blanket placed on top, followed by having the same blanket wrapped around the animal.

The effect on DXA measurements from IV infusion of about 17 mL/kg of parenteral nutrition solution (2.5 g amino acids with 12.5% dextrose containing vitamin and mineral admixture) or from gavage feeding of similar volume of milk-based infant formula (Similac; Ross Products Division, Abbott Laboratories, Columbus, OH) was also determined.

In addition, we used the commercial AWB software for scan acquisition, followed by analysis with 4 different versions of commercial software: 3 adult and 1 pédiatrie software versions, as described above. These studies were designed to determine the relationship among the 4 commercial software products with the infant whole-body vKH6 software and to test the feasibility for a transition from the use of infant to adult software as the subject grows.

Statistical Analysis

Each of the DXA measured parameters (total weight, bone mineral content, bone area, bone mineral density, lean mass, and fat mass) was analyzed separately. General linear model for repeated measures was used for comparisons. The reference scans, as determined for each series of scans, were used as the baseline data and compared with data from other sets of scans using appropriate contrasts. Pearson correlation coefficient was used to determine correlation among various measurements.

Principal components analysis was used to determine the colinearity of the DXA measurements from 3 adult software versions. In addition, general linear model for repeated measures with appropriate contrast was used to determine the relationship among DXA measurements from the 3 AWB versions. Analysis of covariance using body weight category as a fixed factor was used to determine the existence of a weight-dependent interaction in the relationship between measurements from adult or pediatric software with infant software for the same pigs. Regression analysis was used to determine the ability of the measurements from adult and pediatric software to predict measurements from infant software. T tests were used to compare the predicted and residual values from adult and pediatric software prediction equations for infant software measurements. In addition, principal components analysis was used to determine the colinearity of the DXA measurements from adult, pediatric, and infant software.

Unless otherwise stated, all values are mean ± SD. All statistical tests were performed with SPSS 11.5 for Windows (SPSS Inc, Chicago, IL) at an adopted significance level of 0.05.

RESULTS

Subject Location With Respect to Scan Path

Thirteen pigs with weights 1950 to 21,100 g were studied to determine the effect of variations in the location of the subject in relation to the scan paths on fan-beam DXA measurements. Of these, 9 pigs weighing

Subject Posture

Nine pigs with weights of 640 g to 21,100 g were scanned in the prone, supine, and at the side with left and right limbs superimposed. DXA scans of pigs in prone position were used as reference. Regardless of the position of the pigs at scan acquisition, all measurements for each DXA parameter were highly correlated (r = .94 to 1.00, p

Covering

The effect of direct overlay of 1, 3, or 5 layers of cotton blankets on DXA measurements was determined in 11 pigs with weights 640 g to 21,100 g (Table I). DXA scans of pigs covered with 1 blanket were used as the reference. All measured DXA parameters, regardless of the number of blankets overlying the pigs, were highly significantly related (r = .97 to 1.00, p

IV or Gavage Feeding

Six pigs with weights 2250 g to 14,500 g had a repeat scan performed after an IV infusion of 121 ± 63 (mean ± SD) mL of parenteral nutrients. The repeat scan was performed at 6 ± 5 minutes after completion of IV infusion. Results show that DXA measurements pre- and post-IV parenteral nutrients were highly correlated for all parameters (r = .99 to 1.00, p

Another 6 pigs with weights 2305 g to 17300 g had a repeat scan performed after a gavage feeding of 111 ± 64 (mean ± SD) mL of infant formula. The repeat scan was performed at 17 ± 19 minutes after completion of gavage feeding. Results show that DXA measurements pre- and postgavage feeding of milk were highly correlated within each parameter (r = .99 to 1.00, p

Predictability of Infant Software Measurements From Pediatric and Adult Software

Twenty-six pigs with weights 640 g to 21,100 g were scanned in the infant and adult mode followed by analysis with the infant, pediatric, and 3 versions of adult software as described above. Principal components factor analysis indicated that composite scores for bone area, bone mineral content, bone mineral density, lean mass, fat mass, and total weight derived from each of the 3 adult software versions accounted for 99.9% to 100% of the total variance for each DXA measured parameter. Thus, measurements from any one of the adult software versions are interchangeable with those from the other 2 adult software versions. However, general linear model for repeated measures with contrast indicated that total weight, lean mass, and fat mass was significantly lowered when the scans were analyzed with v8.24a:3* by an average of about 54 g (p

There was an interaction between the pig weights (below or above 10 kg) and the predictive equations of adult and pediatric software for infant software measurements. When only the larger pigs with weights >10 kg (ra = 12) were considered, all DXA measured parameters were highly significantly correlated (r = .86 to 1.00, p

DISCUSSION

The technical challenges of obtaining DXA measurements in small subjects are being overcome by multiple investigators for the pencil-beam7-9 and the newergeneration fan-beam2,3 technique. However, knowledge of the technical issues that may affect DXA results is critical to the generation of meaningful data because any error would be proportionally greater in the measurement of small subjects. Our study represents the first report using a piglet model to determine the influence of a number of practical issues that may occur during data acquisition and analysis for fanbeam DXA measured body composition in clinical studies of infants and young children and in experimental studies of small animals. Because the physics of the DXA technique is not a priori based on knowing the species being examined, our data are applicable to both animal and human studies. It is also important to note that for clinical studies of unsedated infants and young children, the possible occurrence of motion artifact4 potentially could exaggerate any changes from other technical issues with DXA scans described in our piglet model.

Our data demonstrated that the fan-beam DXA densitometer transmitted photons to the panel of detectors across a width of at least 19 cm, and a whole-body scan can be acquired with 3 longitudinal sweeps along the long axis of the instrument. This is in contrast to the limited distribution of photon and detectors used by the pencil-beam DXA, which necessitates a slower rectilinear pattern of scan acquisition. Thus, a whole-body scan of an older infant can be completed in

The manufacturer of the type of fan-beam DXA densitometer used in this study has modified the hardware and software to minimize the potential for projection magnification errors. These include rotation of the C arm that contains the photon detectors and automatic adjustment of table heights during scanning at different paths and additional modifications to the software algorithm. These modifications appear to compensate adequately for the study of bone mass and body composition in adults.10 In the study of small humans such as infants and young children, it is possible that many subjects could be placed within a single path, and it is theoretically possible to lower the scanning time to

Our data also demonstrated that a change in the posture of the piglet from prone to supine can affect several fan-beam DXA measurements. This finding differs from our previous report that pencil-beam DXA measurements were not significantly different between scans acquired in prone and supine positions.4,5 Unlike the prone position, the extremities of piglets in the supine position cannot be in contact with the scanning table, and the distance between each extremity and the scanning table increases with increased size of the piglets. Fan-beam DXA with its associated magnification error would become greater as the distance between the body part and the scanning table (the level of photon source) increases. In contrast, pencil-beam DXA scans are acquired in a rectilinear fashion with minimal magnification error. Furthermore, the piglets used for the current fan-beam DXA studies were about 3 times the weight of the piglets used for pencil-beam DXA studies,7-9 thereby increasing the potential for magnification error. The difference in fan-beam DXA measurements between the prone and side postures might be due in part to superimposing one extremity over the opposite extremity, thus decreasing the bone area and possibly affecting other bone measurements. It is also possible that a true difference in the DXA measured area exists between the side vs prone posture. In any case, the critical issue in clinical studies is the documentation whether the infant was bundled with extremities closed to the body or whether the infant was allowed to lie in a posture with extremities extended away from the body, because unlike that for the piglets, all parts of the human body would be in contact with the scanning table, regardless of whether the subject is placed in a supine or prone position.

The use of clothing or covering is a necessity in all clinical studies for infants and young children to minimize any stress to thermoregulation. Our data indicated that even a single blanket could affect DXA measurements when it is wrapped around the small subject, thereby reflecting the equivalent of multiple layers of blanket. The latter situation typically reflects the clinical practice in the study of smallest infants. The increase in DXA total weight is translated into changes in the components of DXA measurement, primarily in lean tissue mass. Thus, consistency in the type and amount of covering used would be critical to minimize any spurious result. In any case, the effect of blankets on fan-beam DXA measurements decreases proportionally as the subject grows.

We studied the effect of parenteral or enterai feeding to determine the extent these potential clinical practices might affect DXA measurements. Although DXA technique was not designed and is not expected to detect specific nutrient components delivered to the subjects, nevertheless, it is interesting that an IV bolus of parenteral nutrient infusate can significantly affect DXA measured total weight and lean mass. This is not surprising because the parenteral nutrients contained amino acids; minerals and electrolytes can mimic lean tissue. In contrast, gavage feeding of milk formula did not appear to have any major effect on DXA measurements. The extent of effect that gastric emptying, digestion and absorption processes, and presence of milk and digestive product in the gut lumen may have on the changes in DXA measurement is not known. The changes in bone area and bone mineral density were small and did not affect bone mineral content measurement. Thus, the common practice of using milk feeding to pacify the infant and to allow the infant to fall asleep before scanning does not seem to greatly affect DXA measurements. In any case, the effect of milk feeding on fan-beam DXA measurements also decreases proportionally as the subject grows.

The data from adult software tested were essentially identical. The statistically significant difference in several DXA parameters is the result of extremely small differences in scan analysis amongst the 3 adult software versions. Thus, changes to the version number of the commercial DXA software does not necessarily equate with significant changes in DXA measurements, although large measurement differences also may occur with change in software version (11). It is imperative for the manufacturer to state whether there are any significant changes in DXA measurements with each change in the software version number to allow for meaningful interpretation of DXA studies.

We have demonstrated that both the fan-beam DXA pediatric and adult software were able to generate measurements when used for small subjects, although there was an interaction effect from the piglet weight. In any case, the fan-beam DXA measurements generated from the infant, pediatric, and adult software are highly correlated, presumably in part related to the validation of the fan-beam DXA infant software,2,3 with a much greater range of piglet weights and body composition compared with that for the pencil-beam DXA infant software.7-9 In this study, the strongly predictive relationships of DXA measurements among the infant, pediatric, and adult software would support that data generated from subjects with weights >10 kg using the same fan-beam densitometer are comparable. This is of importance in longitudinal growth studies, and systematic corrections in the form of linear transformation are possible to allow comparison of clinical and experimental data in the growing subject beyond 10 kg in weight. In contrast to the adult or infant software, the pediatric software has never been independently validated with chemical analysis or other cross-calibration studies. Our data also demonstrated that the use of pediatric software offered no significant advantages over the direct transition to the adult software.

There is no perfect technique for the simultaneous measurement of bone mass and soft tissue composition, although DXA is probably the most extensively studied. It is generally accepted as the gold standard for the measurement of bone mass and is the preferred means to measure multiple components of soft tissue composition in small subjects.1 However, this study demonstrated that any error in DXA measurements would be proportionally greatest in the smallest subjects. Attention to details and consistency in the technique for scan acquisition and analysis are critical to the generation of meaningful data and to allow for the detection of true differences in DXA measured bone and body composition of small subjects.

REFERENCES

1. Koo WWK. Body composition measurements during infancy. Ann N Y Acad Sci. 2000;904:383-392.

2. Koo WK, Hammami M, Hockman EM. Use of fan beam dual energy X-ray absorptiometry to measure body composition of piglets. J Nutr. 2002;132:1380-1383.

3. Chauhan S, Koo WW, Hammami M, Hockman EM. Fan beam dual energy X-ray absorptiometry body composition measurements in piglets. J Am Coll Nutr. 2003;22:408-414.

4. Koo WW, Walters J, Bush AJ. Technical considerations of dual energy x-ray absorptiometry-based bone mineral measurements for pediatric studies. J Bone Mineral Res. 1995;10:1998-2004.

5. Koo WW, Hockman EM, Hammami M. Dual energy X ray absorptiometry measurements in small subjects: conditions affecting clinical measurements. J Amer Coll Nutr. 2004;23:212-219.

6. Gluer CC, Blake G, Lu Y, Blunt BA, Jergas M, Genant HK. Accurate assessment of precision errors: how to measure the reproducibility of bone densitometry techniques. Osteopor Int. 1995;5:262-270.

7. Koo WW, Massom LR, Walters J. Validation of accuracy and precision of dual energy x-ray absorptiometry for infants. J Bone Miner Res. 1995;10:1111-1115.

8. Picaud JC, Rigo J, Nyamugabo K, Milet J, Senterre J. Evaluation of dual energy X ray absorptiometry for body composition assessment in piglets and term human neonates. Am J Clin Nutr. 1996;63:157-163.

9. Fusch C, Slotboom J, Fuehrer U, et al. Neonatal body composition: dual energy X ray absorptiometry, magnetic resonance imaging, and three dimensional chemical shift imaging versus chemical analysis in piglets. Pediatr Res. 1999;46:465-473.

10. Albanese CV, Diessel E, Genant HK. Clinical applications of body composition measurements using DXA. J Clin Densit. 2003; 6:75-85.

11. Koo WWK, Hammami M, Shypailo RJ, Ellis KJ. Bone and body composition measurements of small subjects: discrepancies from software for fan-beam dual energy x-ray absorptiometry. J Am Coll Nutr. In press.

Mouhanad Hammami*; Winston W. K. Koo*; and Elaine M. Hockman[dagger]

Departments of *Pediatrics, Obstetrics and Gynecology and [dagger]Computing and Information Technology, Wayne State University, Detroit, Michigan

Received for publication September 16, 2003.

Accepted for publication April 23, 2004.

Correspondence: Dr. Winston Koo, Hutzel Hospital, Carman and Ann Adams Department of Pediatrics, 4707 St. Antoine Blvd, Detroit, MI 48201. Electronic mail may be sent to wkoo@wayne.edu.

Copyright American Society for Parenteral and Enteral Nutrition Sep/Oct 2004
Provided by ProQuest Information and Learning Company. All rights Reserved