| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH |
From the Department of Reproductive Biology, Case Western Reserve University, MetroHealth Medical Center, Cleveland, Ohio.
Address reprint requests to: Patrick M. Catalano, MD, Case Western Reserve University, MetroHealth Medical Center, Department of Reproductive Biology, 2500 MetroHealth Drive, Cleveland, OH 44109, E-mail: pcatalano{at}metrohealth.org
 |
Abstract |
|---|
Top
 Abstract Materials and MethodsResultsDiscussionReferences |
|---|
Methods: Twenty healthy pregnant women at 30 weeks’ gestation had estimates of body composition using hydro-densitometry, with corrections for residual lung volume, and total body water using 18O (development group). Total H2 body water was estimated from 18O abundances measured by gas-isotope-ratio mass spectrometry. Maternal age, height, weight, and seven skinfold sites were correlated with fat mass using stepwise regression analysis. The anthropometric model to estimate fat mass was then tested prospectively in a second group of 20 subjects and correlated with underwater weighing and total body water measurements (validation group). Statistical analysis used 
2, paired t and Wilcoxon sign-rank tests.
Results: There were no statistically significant differences in maternal demographics between groups. The fat mass of development group subjects using underwater weighing and total body water was 22.7 ± 7.6 kg. Using the development group, a model was derived that explained 91% of the variance in fat mass by underwater weighing and total body water using maternal weight and triceps, subscapular, and suprailiac skinfolds (r2 = 0.91, P < .001). When tested prospectively in the validation group, the correlation remained statistically significant (r2 = 0.89, P < .001). There was no statistically significant (P = .88) difference between the anthropometric estimates of fat mass and underwater weighing and total body water measurements (95% confidence interval -2.476, 2.748 kg of fat mass).
Conclusion: This anthropometric model can be used to predict maternal fat mass in late gestation.
Estimation of body composition is important for assessing nutrition and metabolism. Knowing a subject’s body composition allows an investigator to more accurately classify obesity status compared with either weight alone of body mass index (weight/height2). Estimating body fat allows for more precise classification of individuals for issues related to assessment of nutritional interventions. The body can be divided theoretically into two compartments, fat mass and fat-free mass. Although the terms fat-free mass and lean body mass have been used interchangeably, fat-free mass is the preferred term. Lean body mass refers to body mass minus nonessential or excess lipids. Lean body mass differs from fat-free mass by a small quantity (2–5%) of essential lipid in the central nervous system and other organs. In general, water constitutes approximately 72% of fat-free mass in nonpregnant women, whereas fat mass is considered essentially devoid of water. Those characteristics, with estimates of body density using a density for fat-free mass of 1.1 g/m2 and a density for fat mass of 0.9 g/m2, are the basis for most methods of estimating body composition in living human beings such as densitometry and isotope dilution.
Because on the increases in total body water during pregnancy, estimating body composition (fat mass and fat-free mass) is best done using multicompartment modeling, which includes densitometry with underwater weighing and isotope dilution to estimate total body water. During pregnancy, measures of density tend to overestimate fat mass by lowering body density closer to that of fat. Measures of total body water tend to underestimate fat mass by overestimating water compartments and therefore fat-free mass. Those research methods are expensive and not readily available to most investigators, so the primary purpose of this study was to estimate body composition of pregnant women by using a two-compartment model, ie, density and total body water, then develop a model to estimate fat mass using anthropometric measurements in late gestation.
|
Materials and Methods |
|---|
|
Top
AbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Forty pregnant women were recruited at 30 weeks’ gestation through advertisements in the hospital. This was a consecutive series with a success rate of approximately 75% in recruitment. The initial measurements were made on 20 subjects to derive the model, then prospectively tested in another 20 subjects. All subjects were healthy nonsmokers who had no clinical evidence of edema. Each subject had a normal 50-g, 1-hour glucose (under 135 mg/dL) screening test for gestational diabetes, and all pregnancies were singletons. All subjects completed the entire protocol. The same 40 women were studied to collect data estimating the hydration constant for fat-free mass in pregnancy, reported previously.1
Subjects were admitted at 7:00 AM after an overnight fast. Demographic information and obstetric histories were collected then. Subcutaneous adipose tissue was measured using skinfold calipers. Total body water was estimated using plasma samples collected at 3, 4, and 5 hours after ingestion of H218O. Body density was calculated using hydrodensitometry with correction for residual lung volume. Subjects were given a light breakfast and then discharged.
Each subject’s body density was determined by an underwater weighing procedure with simultaneous correction for residual lung volume. Each subject was first weighed on a calibrated metabolic scale (Toledo Scales Corp., Worthington, OH) in her bathing suit, and weight was adjusted for the weight of the bathing suit. Subjects then entered the underwater weighing tank that has a cot with four calibrated load cells (Precision Biomedical Systems, State College, PA) to measure weight when the subject was under water. While breathing room air through a mouthpiece (the nasal passage was closed with a nose clip), the subject submerged. The weight of the subject was taken at the time of maximum expiration. Residual lung volume at the time the weight of the subject was measured underwater was estimated using a nitrogen washout technique,2 which was accomplished by having subjects switch from breathing room air to 100% oxygen using a specially designed two-chamber mouthpiece. Subjects were than allowed to come up from under that water and breathe until filling a 120-L tissot. After oxygen washout, a reading was taken from the nitrogen analyzer (MedScience, St. Louis, MO) and residual lung volume was calculated. That was used to adjust for the weight of the subject while submerged, because the greater the residual lung volume the less subjects weighed under water.
Total body water was calculated by having each woman ingest 125 mg of H218O (Isotec Inc., Miamisburg, OH) per kilogram of body mass after collection of baseline plasma samples. Venous plasma samples were collected at 3, 4, and 5 hours after ingestion of H218O for determination of total body water. Abundances of 18O in the baseline and postdose samples were measured by gas-isotope-ratio mass spectrometry and were used to calculate total body water, which was the average of the 3-, 4-, and 5-hour postdose 18O isotope-dilution spaces (NO), calculated as follows:
 |
where d is the amount of H218O given to each subject in grams, A is the amount of laboratory water in grams used in the gravimetric dilution of x (g) of the isotope water; Ex is the change in isotopic abundance (% O) of the laboratory water after the addition of the isotopic water, and EO is the 18O enrichments in plasma above the baseline values at 3, 4, or 5 hours after the isotope was ingested.
Measurements of density and total body water were available for all subjects, so body composition was calculated using both factors, as described by Siri3:
 |
Fat mass was measured in kilograms. Refer to reference 3 for the complete derivation of the constants in the equation.
Subcutaneous fat deposits were measured by one investigator (PMC) at seven different sites on the left side of the subject using Harpenden skinfold calipers (British Indicators, Sussex, England). The sites include the biceps, triceps, subscapula, suprailiac, costal, thigh, and suprapatellar. The biceps skinfold was measured on the anterior aspect of the arm, midway between the acromion process and the antecubital fossa as determined by a tape measure. The triceps skinfold was measured in the midline of the posterior portion of the arm at the midpoint between the lateral projection of the acromion process and olecranon. The subscapular skinfold was measured at a 45° angle at a site just below the inferior angle of the scapula. The suprailiac skinfold was measured at the midpoint between the anterior superior iliac spine and the lowest rib, as measured by a tape measure. The costal skinfold measurement was taken at the midaxillary line at the level of the lowest rib. The thigh skinfold measurement was located in the midline of the anterior thigh at the midpoint between the inguinal crease and proximal border of the patella. The suprapatellar skinfold measurement was taken in the midsagittal plane on the anterior aspect of the thigh, above the patella on the proximal edge. Each measurement was done three times and held for 20–30 seconds each and averaged when a stable reading was obtained.4 Skinfold measurements were reproducible. In a study of ten subjects by the same examiner (PMC) in women who ranged from lean to obese, the mean difference in repeat triceps skinfold measurements was 0.63 mm and for subscapular skinfold measurements 0.68 mm. The correlation between the two measurements was linear; for the triceps, r = 0.99, P < .001, and for subscapular, r = 0.98, P < .001.
We evaluated 40 subjects based on the number of women needed to estimate the hydration constant for fat-free mass in pregnancy that was reported previously.1 In the original power analysis, we assumed the outcome was a correlation coefficient and that a one-sample correlation was normal. We also assumed that the null hypothesis correlation was equal to zero. The alternative hypothesis correlation ranged from 0.45–0.60, with a significance of .05 and power of 0.8. Using those conditions, we needed a sample of 20 for a correlation coefficient of 0.6 or a sample of 37 for a correlation of 0.45. Based on our previous results, we believed that there would be sufficient statistical power to detect differences between anthropometric and combined total body water and density measures of fat mass.
The data are expressed as mean ± standard deviation (SD). Differences between the two groups of subjects were estimated using unpaired t tests, Wilcoxon sign-rank test, and 2 analysis, as appropriate. Stepwise linear regression analysis was used to develop a model. We used best-fit estimates of fat mass, ie, dependent variable with multiple independent variables, including age, weight, height, and the seven skinfold measurements. P < .05 was considered statistically significant. All statistical analyses were done using Statview II and Statview 4.5 statistical packages (Abacus Concepts Inc., Berkeley, CA).
|
Results |
|---|
|
Top
AbstractMaterials and MethodsResultsDiscussionReferences |
|---|
. There was no significant difference in age, race, height, weight (pregravid and at time of study), gravidity, parity, and gestational age at time of study between groups. There also was no significant difference in any measurement of skinfold thickness between groups (triceps, biceps, subscapula, costal, suprailiac, thigh, and suprapatellar), as shown in Table 2.
|
|
shows the statistical model that correlates best with fat mass (kg) = (weight • 0.33529) + (triceps • 0.65664) - (subscapula • 0.4373) + (suprailiac • 0.43461) - 13.0538, with weight measured in kilograms and skinfold sites measured in millimeters. Table 3 shows the results of the stepwise regression analysis. The best model, ie, r2 = 0.91 included the suprailiac, subscapula, and triceps skinfolds and maternal weight. Maternal age, height, and the other skinfold measurements did not improve our model. For example, if one included all additional skinfold measurements (suprapatellar, costal, thigh, and biceps), the adjusted r2 was 0.90.
|
|
.
|
|
Discussion |
|---|
|
Top
AbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Besides total body water and density, another compartment that could be assessed directly, but which we elected not to, is the mineral compartment. Although it could be assessed directly using dual photon x-ray devices, it would require exposure of pregnant women and their fetuses to radiation, albeit very small doses. Because of the potential safety concerns and the potential for very little, if any, change in maternal bone mineralization status in pregnancy ( Little KD, Clapp JF, Gott PJ. Bone density changes during pregnancy and lactation in exercising women. Med Sci Sport Exerc 1993;25:5154 [Abstract]), we used a two-compartment model as our reference method.
We acknowledge several limitations of our study design. All our studies were conducted at or near 30 weeks’ gestation. The accuracy of the anthropometric model in early pregnancy is not known but is currently being studied in our laboratory. None of our study subjects had evidence of clinical edema or polyhydramnios (as estimated by ultrasound). Increases in total body water, commonly found in those conditions during pregnancy, will adversely affect estimates of body composition using our anthropometric model. Our subjects were predominantly white women with normal glucose tolerances. Although race and glucose intolerance are unlikely to affect the model’s ability to predict maternal body composition, additional studies are needed.
Although there was a significant correlation between the anthropometric estimate of maternal body composition using Forsum’s model and the two-compartment model of Siri, the correlation was stronger (r2 = 0.89 versus r2 = 0.44) with our model. In any model used to estimate body composition in late gestation, one cannot readily distinguish estimates of maternal from fetal body composition. Although ultrasound has been used to assess fetal body composition, as with estimates of fetal weight, the estimates of fetal body composition have not been accurate. The mean ± absolute percent error in the estimates of fat mass between sonographic measurements and total body electrical conductivity measurements was 11.8 ± 9%. The 95% CI for ultrasound estimates of fat mass was 7.4, 16.2%.6 This anthropometric model does not distinguish maternal fat mass from fetal fat mass.
The results of this study support our hypothesis that an anthropometric model can be used to estimate maternal body composition in late pregnancy. The tools required are inexpensive (estimated cost of the Harpenden skinfold caliper is $299) and can be readily mastered after a brief training period. Additionally, this method poses no risk to mother or fetus and can be completed in less than 10 minutes. Estimation of body composition in late pregnancy might be useful in assessing the nutritional status of women in late gestation. Changes in total body water in pregnancy, ie, increased plasma volume and amniotic fluid, mean that increases in weight alone might not indicate differential changes in maternal body composition or nutritional status.
|
Footnotes |
|---|
Received November 10, 1999. Received in revised form February 9, 2000. Accepted February 25, 2000.
|
References |
|---|
|
Top
AbstractMaterials and MethodsResultsDiscussionReferences |
|---|
2. Akers R, Buskirk ER. An underwater weighing system utilizing "force cube" transducers. J Appl Physiol 1969;26:649–52.
3. Siri WE. Body composition from fluid spaces and density: An analysis of methods. In: Brozek J, Hanschel A, eds. Techniques for measuring body composition. Washington, DC: National Academy of Science-National Research Council, 1961:223–44.
4. Harrison GG, Buskirk ER, Lindsay Carter JE, Johnston FE, Lohman TG, Pollock ML, et al. Skinfold thickness and measurement technique. In: Lohman TG, Roche AF, Martorel R, eds. Anthropometric standardization reference manual. Champaign, Illinois: Human Kinetics Books, 1988:55–70.
5. Forsum E, Sadurskis A, Wager J. Estimation of body fat in healthy Swedish women during pregnancy and lactation. Am J Clin Nutr 1989;50:465–73.
6. Crane SS, Avallone DA, Thomas AJ, Catalano PM. Sonographic estimation of fetal body composition with gestational diabetes mellitus at term. Obstet Gynecol 1996;88:849–54.[Abstract]
This article has been cited by other articles:
|
|
 |
|
  T Ueland, T Dalsoren, N Voldner, K Godang, T Henriksen, and J Bollerslev Retinol-binding protein-4 is not strongly associated with insulin sensitivity in normal pregnancies. Eur. J. Endocrinol., July 1, 2008; 159(1): 49 - 54. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
