| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH COMMUNICATION |
1 From the Institute of Human Nutrition, University of Southampton, Southampton General Hospital, Southampton, United Kingdom (OAK, SAW, AAJ, and ME), and the MRC Epidemiology Resource Centre, University of Southampton, Southampton, United Kingdom (DIP and MP)
2 Supported in part by NIH grant RO1 HD 41107-01 (to DIP), by core funds from the Institute of Human Nutrition and Epidemiology Resource Centre, and a grant from Umm Al Qura University, Saudi Arabia, which supported OAK during his studies at the University of Southampton.
3 Address reprint requests to M Elia, Institute of Human Nutrition, University of Southampton, Southampton General Hospital, Tremona Road, Southampton SO16 6YD, United Kingdom. E-mail: elia{at}soton.ac.uk.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Objective: The objective was to assess body composition and fat distribution by using dual-energy X-ray absorptiometry (DXA).
Design: This was a case-control study of 64–72-y-old white men (n = 32) with a low (
: 2.76 kg) or high (: 4.23 kg) birth weight.
Results: Compared with the high-birth-weight group, after adjustment for weight and height, the low-birth-weight group had a higher percentage body fat (29.31% compared with 25.33%; P = 0.029) and fat mass (P = 0.039) but a lower fat-free soft tissue (56.32 compared with 59.22 kg; P = 0.024), muscle mass (27.25 compared with 29.22 kg; P = 0.022), and muscle-to-fat ratio. Low birth weight was also associated with a higher trunk-to-limb fat ratio after control for total fat mass (1.42 compared with 1.16; P = 0.005) or percentage body fat (P = 0.041). The same body mass index predicted a greater percentage body fat (P = 0.019) in the low- than in the high-birth-weight group, and the same ratio of trunk-to-limb skinfold thickness (or waist-to-hip ratio) predicted a higher trunk-to-limb fat ratio (P < 0.01).
Conclusion: Lifelong differences in adult body composition and fat distribution between the low- and high-birth-weight groups are consistent with programming in early life. The use of BMI to predict percentage body fat and the use of the trunk-to-limb skinfold thickness ratio (and waist-to-hip ratio) to predict the trunk-to-limb fat ratio measured by DXA can be misleading when low- and high-birth-weight groups are compared.
Key Words: Body composition • birth weight • fat • lean mass • muscle programming • body mass index • dual-energy X-ray absorptiometry • waist-to-hip ratio • fat distribution • skinfold thickness
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Fat distribution also predicts risk of cardiovascular disease and type 2 diabetes, independently of fat mass (FM) and percentage body fat. It is possible that early life programming of body composition affects fat distribution between body segments, but reference scanning body-composition techniques do not appear to have been used to address this issue. In contrast, the waist-to-hip circumference ratio or ratio of trunk to limb skinfold thickness (7) have frequently been used for this purpose, but the results have been variable, at least partly because of the limitations associated with the use of these techniques. For example, the waist-to-hip ratio (20) is influenced by skeletal shape and muscular development, and, in some individuals, by intestinal gas and fecal mass. In addition, in subjects with a low BMI, it has been suggested that waist circumference reflects the variation in abdominal lean tissue as much as it reflects variation in abdominal adiposity (8). Skinfold thicknesses reflect only subcutaneous fat at the measurement sites, rather than total fat (internal and subcutaneous fat) over entire body segments. Furthermore, if birth weight is related to percentage fat, which in turn is related to fat distribution, then the use of simple unadjusted (or incorrectly adjusted) anthropometric indexes of fat distribution may produce biased results and potentially incorrect conclusions.
The aims of this study were 2-fold: 1) to assess whether birth weight is related to the mass and proportion of tissues (lean fat, muscle fat), percentage body fat, and its distribution across body segments in older adults, as measured by dual-energy X-ray absorptiometry (DXA), and 2) to test the hypothesis that anthropometric markers of health risk, such as BMI, skinfold thicknesses, and waist-to-hip ratio can be misleading in comparisons of body composition and fat distribution between low- and high-birth-weight groups.
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
), to undergo assessments of anthropometric and body-composition characteristics. All subjects were born in Hertfordshire, where detailed birth records were routinely kept. The subjects were chosen randomly from among those below the 25th centile of birth weight [<3.18 kg (7 lb); n = 16] or above the 75th centile [>3.86 kg (8.5 lb); n = 16] of birth weight. Only 55% of those who were contacted by letter (after permission was granted by their general practitioners) agreed to have a home visit and participate in the body-composition studies. Written informed consent was obtained from each subject, and ethical approval for the study was obtained from local research committees. This study was part of a larger study that included other measurements of energy metabolism and fuel selection (data not reported in this article), which for practical reasons limited the sample size. All the required measurements were obtained from each subject participating in the study. Weight was measured to the nearest 0.1 kg with a Seca 708 electronic weighing scale, and height was measured to the nearest 0.1 cm with a Seca electronic stadiometer (Seca Ltd, Medical Scales and Measurement Systems, Birmingham, United Kingdom). Information on socioeconomic status (both currently and at the time of birth), smoking status, alcohol consumption, and physical activity (Table 2) was obtained by using questionnaires that were administered to the subjects on the study day. The study was conducted according to the South-West Hampshire Local Ethics Committee.
|
|
Whole body weight (in kg), obtained from the sum of the weights of body segments reported by DXA, was highly correlated (r = 0.997) with measured weight. Waist and hip circumferences were measured according to standard procedures (21). Skinfold thicknesses (biceps, triceps, subscapular, and suprailiac) were measured according to the method of Durnin and Womersley (22) and were used to calculate percentage body fat. Midupper arm circumference was measured midway between the acromion and olecranon, and, together with triceps-skinfold thickness, was used to calculate arm muscle area (23):
 | (1) |
Whole-body muscle mass was estimated by using the equation of Kim et al (24), which was established by relating DXA measurements of appendicular (limb) FFST with whole-body measurements of muscle mass obtained by using magnetic resonance imaging in healthy adults. The equation for men is as follows:
 | (2) |
The inter- and intraobserver CVs for body fat and fat-free mass (FFM) (as a percentage of body weight) for repeated measurements in the same subjects is 
1% for DXA and air-displacement plethysmography (3–4% of actual FM, 1.4% of actual FFM). For BMI, the CV is <1% (interobserver and intraobserver CVs). For DXA the CVs (as a percentage of actual tissue or body component) were as follows: 2–3% for fat in the trunk and in the limbs, 2.0% for the trunk-to-limb fat ratio, 0.9% for bone mineral, 1% for FFST and 1.5% of estimated skeletal muscle mass, and 0.9% for bone mineral content. The intraobserver CVs for measurements with the same scan were 1.35% for the bitrochanteric breadth (manual procedure) and 3.5% for abdominal FM. The intraobserver variation for skinfold thicknesses (5–10%) (23, 26, 27) yield CVs for skinfold ratios of 3–10% (3.5–7.0% for the ratio of subscapular + suprailiac/biceps + triceps) when calculated with the use of propagation of errors (28) and correlation coefficients between numerator and denominator obtained in our study (r = 0.40–0.77)). Interobserver variations were generally greater (26, 27, 29). The CVs for midarm circumference and for waist-to-hip ratio were <2%.
Statistical analysis
One-way analysis of variance was used to compare the mean differences in weight, height, FFST, FM, muscle mass, and the ratios of nonlimb to limb, trunk to limb, and abdomen to limb FM between both the low- and high-birth-weight groups. Analysis of covariance was also used to explore the effect of birth weight (low- and high-birth-weight groups as fixed factors) on body composition and fat distribution, with and without adjustment for covariates such as body mass index (BMI) and weight plus height. Chi-square tests were used to assess categorical data, such as socioeconomic status. Total observed variance was taken to be the sum of biological variance, variance due to observer error, and variance due to instrument error, which was assumed to be negligible (26). The statistical analyses were carried out by using the STATISTICAL PACKAGE FOR THE SOCIAL SCIENCES (version 12.0; SPSS Inc, Chicago, IL). Continuous data are expressed as means ± SEs.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). Although subjects in the low-birth-weight group were significantly shorter (8.0 cm) and lighter (8.3 kg) as adults than were the subjects in the high-birth-weight group, their BMIs were not significantly different (Table 1).
Body composition and fat distribution measured by DXA
Compared with the high-birth-weight group, the low-birth-weight group was found to have significantly less FFM (8 kg), less FFST (7 kg) in the whole body (Table 1) and in the limbs (limb FFM, 3.8 kg; limb FFST, 3.6 kg), and less muscle in the whole body (4.4 kg), but not significantly less FM. Because these differences could be due to the smaller adult size (weight and height) of the low-birth-weight group, the data were reexamined after control for weight plus height (Table 3) (weight + height2 or weight and 1/height2 gave identical or almost identical results and significance values; data not shown). However, both FFM and FFST remained significantly lower in the low-birth-weight subjects than in the high-birth-weight subjects, by 3 kg (muscle mass by 2 kg), whereas FM was significantly lower by 3 kg, corresponding to 4% of body weight (Table 3). The effect of other potential variables, such as socioeconomic status, physical activity, and smoking status (Table 2) were also examined, by adding them as covariates (no more than one at a time) in the model, but because they were evenly matched between groups and had no significant effect, they were removed from the model for simplicity. The same applied to the other analyses reported below.
|
Fat distribution assessed by DXA was found to differ between the low- and high-birth-weight groups, in favor of a more central distribution in the low-birth-weight group (Table 1 for unadjusted values and Table 4 for values adjusted for weight plus height and for total-body FM). This finding was due to a combination of a greater trunk (and nonlimb fat) and a lower limb FM in the low-birth-weight group (Table 4). The trunk-to-limb fat ratios were also significantly greater in the low-birth-weight group after adjustment for percentage body fat: 1.37 ± 0.055 and 1.21 ± 0.055, respectively (P = 0.041). The ratio of abdominal fat to limb FM was also greater in the low-birth-weight group than in the high-birth-weight group (0.29 ± 0.013 and 0.027 ± 0.013; P = 0.314), but there was substantial variability, and the upper and lower abdominal boundaries could only be defined to the nearest pixel (1.3035 cm), which corresponded to 15% of the distance between these 2 boundaries.
|
), and BMI (29.31% and 25.33%, respectively; Table 3).
Percentage body fat, estimated from skinfold thicknesses in all subjects (27.3 ± 0.62%; n = 32), was not significantly different from measurements obtained by DXA (27.3 ± 0.95%; n = 32) or air-displacement plethysmography (27.9 ± 1.38%). Although skinfold thicknesses predicted a lower percentage body fat in the low- than in the high-birth-weight group, the difference was not significant, either before or after adjustment for BMI or weight plus height. More than 90% of the variability in all skinfold-thickness ratios and waist-to-hip circumference was estimated to be due to biological variability and <10% to measurement error. The mean ratios of trunk (subscapular + suprailiac) to limb skinfold thicknesses (biceps + triceps) were not different between the low- and high-birth-weight groups (80–100% of the values obtained in the high-birth-weight group, both before (Table 1) and after adjustment for BMI and weight plus height (data not shown). This was also shown by analysis of covariance, in which the skinfold-thickness ratio was used as a covariate. This analysis showed that any given trunk-to-limb skinfold-thickness ratio was associated with a significantly greater ratio of whole trunk to whole limb fat obtained by DXA in the low- than in the high-birth-weight group (Table 5), even when further adjustments were made for BMI (Table 5). The same was found to apply to the waist-to-hip ratio (Table 5).
|
). |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Body composition and fat distribution measured by DXA
The low-birth-weight group was found to have a greater amount (3 kg) and proportion (4%, after control for weight plus height) of fat, a more central fat distribution (after control for the amount of or percentage of body fat), and a lower muscle mass and muscle to fat ratio (after control for weight, or weight plus height, or BMI), all of which contribute to an increased risk of cardiovascular disease and diabetes. Although the ratio of abdominal to limb FM was greater in the low than in the high-birth-weight group, the differences were not significant, possibly because of anatomical variation and the inability to define upper and lower boundaries of the region of interest by <1.3035 cm (1 pixel, or 15% of the distance between the boundaries). It is unlikely that significant differences in fat and lean tissue masses between the groups were due to methodologic problems associated with DXA because of the concordance of results with those obtained by air-displacement plethysmography, which is based on very different principles for measuring body composition.
The observed association of birth weight with body composition (lean and FM) in older adults is consistent with the results of another DXA study in England, which suggested that body composition is programmed (2). Although this study did not report on muscle mass, fat distribution, or muscle-fat ratios, it reported a weak but significant association between whole-body bone mineral content and birth weight category (after adjustment for height and sex), which was not observed in our study. Two other studies have suggested that muscle mass is reduced in low-birth-weight individuals. One of these studies, which was based on a timed overnight urine collection for creatinine excretion (30), considered weight and not height in the comparison; the other study, which was based on anthropometric measures of the thigh adjusted for weight but not height (8). The results of the present study, which are based on DXA estimates of muscle mass, suggest differences between the low- and high-birth-weight groups after adjustment for both weight and height. This could help explain another observation (31), which is the reported positive relation between muscular strength and birth weight, after adjustment for weight and height. Unfortunately, no estimates of muscle mass were made in this last study.
Anthropometric adjustments and predictions of body composition and fat distribution measured by DXA
Body fat
This study suggests that the use of BMI (or weight plus height) to compare body composition between the high- and low-birth-weight groups can be misleading. At the same BMI, or same weight and height, percentage body fat was greater in the low-birth-weight group (4% of body weight) than in the high-birth-weight group. Therefore, previous studies that have attempted to control for differences in body composition using BMI may have done so inadequately. This could help explain (together with differences in fat distribution, in the muscle-to-fat ratio, or in both) the apparent paradox of a lower risk of cardiovascular disease and diabetes in high-birth-weight individuals with the same and even greater BMI than in low-birth-weight individuals (11, 14). An analogous situation concerns ethnic differences in body composition. At the same adult BMI, Asians often have 2–6% more body weight as fat than do whites (15–18), who generally have a lower risk of cardiovascular disease and diabetes than do Asians. A study that compared anthropometric estimates of body composition in neonates born in England (mean birth weight: 3.49 kg) and India (mean birth weight: 2.70 kg; 2.76 kg in the low-birth-weight group in the present study) suggests that ethnic differences are present early in life (19). The extent to which these ethnic differences are due to genetic or environmental factors operating in early life is uncertain. The failure of skinfold thicknesses to demonstrate significant differences in percentage body fat between birth-weight groups in our study may have been due to greater technical error in skinfold-thickness measurements relative to DXA, which showed significant differences between birth weight groups.
Fat distribution
Although DXA measurements showed significant differences between the low- and the high-birth-weight groups in the distribution of body fat between central and peripheral body segments, no significant differences were obtained when the ratio of waist-to-hip circumferences (see below) or trunk-to-limb skinfold thickness were used. This last observation may have been due in part to the poorer precision and accuracy of skinfold-thickness measurements obtained at single subcutaneous sites compared with the reference DXA measurements of total fat, both internal and subcutaneous fat, over entire body segments. However, data from this study challenge previous assumptions about the use of skinfold-thickness ratios (or waist-to-hip ratio) for comparing fat distribution between body segments in low- and high-birth-weight groups in an unbiased manner. The same skinfold-thickness ratios (or the same waist-to-hip ratios) and same BMIs were found to be associated with greater trunk and limb fat distribution measured by DXA (ie, anthropometric indexes of fat distribution may have underestimated the central fat distribution in the low-birth-weight group compared with the high-birth-weight group). For skinfold-thickness ratios, this may have been because of differences between the low- and high-birth-weight groups in the distribution of fat between internal and external fat and between different parts of the same body segment. For the waist-to-hip circumference ratio, it is possible that different musculoskeletal to fat proportions produce the same circumferences. The significantly greater bitrochanteric breadth observed in the high- than in the low-birth-weight group, after adjustment for height or weight plus height, is consistent with this proposal.
Previous attempts to assess fat distribution on the basis of the waist-to-hip ratio or the ratio of trunk to limb skinfold thicknesses have produced inconsistent results, with some studies reporting significantly more central fat distribution in individuals with low birth weight (32–36) and others reporting no such relations (37–40). One of the reasons for the lack of consistency is that different skinfold-thickness ratios [subscapular to triceps (32–35, 40), subscapular + suprailiac to triceps + thigh (35), and subscapular + midaxillary to triceps + medial calf (34)] or different circumference ratios [eg, waist to hip (36–40) or waist-to-thigh (8)] have been used for this purpose. Another reason is that some studies have adjusted for weight and not height (36, 40), others have adjusted for height and not weight (8), whereas most studies have adjusted for BMI (32, 33, 35, 38, 39), which controls for neither weight or height (Table 2), and a few studies adjusted for none of these (41–43).
Muscle and muscle-to-fat ratio
The use of midupper arm muscle area as a surrogate measure of muscle mass suggested less muscle in the low- than high-birth-weight group before and after anthropometric adjustments. These differences were not statistically significant, in contrast with the measurements obtained with DXA, which were significantly different. These significant differences were possibly due to the larger technical error associated with the anthropometric measurements, which in reality arise from 2 independent measurements (arm circumference and triceps skinfold thickness). However, the trends observed in this study are consistent with the results of another study (8), which found that birth weight was related to the cross-sectional area of thigh + bone (after adjustment for height and race), assessed by anthropometric methods in applicants for military service in the United States (age: 17–22 y; mean BMI: 23.2).
The present study had several limitations. The number of subjects was small, and it is perhaps surprising that significant results were obtained. However, the study included only subjects with high or low birth weights, excluding subjects with intermediate birth weight, which probably increased the chances of identifying any differences in body composition. It can be hypothesized that intermediate values occur in the "middle" group, but this needs to be established. The different birth weight–disease relations that have been reported, including U-shaped birth weight–adult BMI relations (10), suggest that investigations across all birth weight categories, not just low and high birth weight categories, should be carried out. In our investigation, only men within a narrow age range (64–72 y) were studied, in an attempt to reduce the confounding effects of sex and age, but this limits generalizations to other age groups. However, this approach precludes extrapolation to women and individuals in a different age range. Another limitation was that muscle mass was not measured directly but was estimated through measurements of FFST in the limbs. Finally, because all of the subjects involved in this study were white, there is a need to examine these relations in subjects with different ethnic backgrounds.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Chomtho, J. C. Wells, J. E Williams, A. Lucas, and M. S Fewtrell Associations between birth weight and later body composition: evidence from the 4-component model Am. J. Clinical Nutrition, October 1, 2008; 88(4): 1040 - 1048. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Meas, S. Deghmoun, P. Armoogum, C. Alberti, and C. Levy-Marchal Consequences of Being Born Small for Gestational Age on Body Composition: An 8-Year Follow-Up Study J. Clin. Endocrinol. Metab., October 1, 2008; 93(10): 3804 - 3809. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. D. Gluckman, M. A. Hanson, C. Cooper, and K. L. Thornburg Effect of In Utero and Early-Life Conditions on Adult Health and Disease N. Engl. J. Med., July 3, 2008; 359(1): 61 - 73. [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Wallace, J. S. Milne, R. P. Aitken, and W. W. Hay Jr. Sensitivity to metabolic signals in late-gestation growth-restricted fetuses from rapidly growing adolescent sheep Am J Physiol Endocrinol Metab, November 1, 2007; 293(5): E1233 - E1241. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. R. Gale, M. K. Javaid, S. M. Robinson, C. M. Law, K. M. Godfrey, and C. Cooper Maternal Size in Pregnancy and Body Composition in Children J. Clin. Endocrinol. Metab., October 1, 2007; 92(10): 3904 - 3911. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Elia, P. Betts, D. M Jackson, and J. Mulligan Fetal programming of body dimensions and percentage body fat measured in prepubertal children with a 4-component model of body composition, dual-energy X-ray absorptiometry, deuterium dilution, densitometry, and skinfold thicknesses Am. J. Clinical Nutrition, September 1, 2007; 86(3): 618 - 624. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. B. Jensen, H. Storgaard, S. Madsbad, E. A. Richter, and A. A. Vaag Altered Skeletal Muscle Fiber Composition and Size Precede Whole-Body Insulin Resistance in Young Men with Low Birth Weight J. Clin. Endocrinol. Metab., April 1, 2007; 92(4): 1530 - 1534. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. J. De Blasio, K. L. Gatford, I. C. McMillen, J. S. Robinson, and J. A. Owens Placental Restriction of Fetal Growth Increases Insulin Action, Growth, and Adiposity in the Young Lamb Endocrinology, March 1, 2007; 148(3): 1350 - 1358. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. D. Taylor and L. Poston Developmental programming of obesity in mammals Exp Physiol, March 1, 2007; 92(2): 287 - 298. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P D Gluckman and M A Hanson Adult disease: echoes of the past Eur. J. Endocrinol., November 1, 2006; 155(suppl_1): S47 - S50. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. S Rogers, A. R Ness, C. D Steer, J. C. Wells, P. M Emmett, J. R Reilly, J. Tobias, and G. D. Smith Associations of size at birth and dual-energy X-ray absorptiometry measures of lean and fat mass at 9 to 10 y of age. Am. J. Clinical Nutrition, October 1, 2006; 84(4): 739 - 747. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
