| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
School of Medicine, Health Policy, and Practice (P.M.L.S., A.C., A.J.M.), University of East Anglia, Norwich NR4 7TJ, United Kingdom; and Twin Research and Genetic Epidemiology Unit (R.S., J.B.R., T.D.S., A.J.M.), Kings College, London SE1 7EH, United Kingdom
Address all correspondence and requests for reprints to: Paula Skidmore, School of Medicine, Health Policy, and Practice, University of East Anglia, Norwich NR4 7JT, United Kingdom. E-mail: p.skidmore{at}uea.ac.uk.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Setting, Design, and Participants: We conducted a cross-sectional study in a cohort of 1194 female twins aged 18–74 yr. The relationship between birth weight and insulin resistance was analyzed using a regression method allowing for a simultaneous estimation of within- and between-pair influences. The approach allows the influence of individual fetal nutrition on adult insulin resistance to be distinguished from effects that are mediated by confounding factors in the maternal environment.
Main Outcome Measures: Insulin resistance was measured by the homeostasis model assessment.
Results: Individual level regression analyses showed no significant relationship between birth weight and insulin resistance. There was a significant positive relationship between insulin resistance and current body mass index (BMI) (a 26% increase in insulin resistance per SD increase in BMI; confidence interval, 22.6–29.5%). This significant relationship was accounted for in equal parts by individual-specific effects and by confounding factors in the shared environment of the twins. The relationship with birth weight became significant only after adjustment for BMI and was mediated only through between-pair differences.
Conclusions: These results suggest that insulin resistance is influenced more by current body size than birth weight and that postnatal growth is potentially more important than fetal growth in the subsequent development of insulin resistance.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Investigating the relationship between birth weight and variables measured in adult life in twins allows the influences of birth weight, maternal environment, and genetic factors to be separated analytically. The current study aimed to investigate the relationship between birth weight, current body mass index (BMI), and insulin resistance in a population-based cohort of adult female twins. Studying the strength of the association in the group as a whole provides a measure that is comparable to the results of studies of unmatched singleton cohorts. However, additional information is obtained by modeling the relationship within and between twin pairs. Any observed relationship can be attributed to factors that are specific to the environment of the individual members of each pair and to confounding by factors in the shared environment of the twins. Furthermore, comparing the strength of the relationships in monozygotic (MZ) and dizygotic (DZ) pairs also provided a measure of the extent to which the relationship might be mediated by genetic factors.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The study subjects comprised a sample of twins enrolled on the Twins UK Registry. This is a nationwide registry of twin volunteers who have been recruited through successive media campaigns (4). The register was set up to examine women’s health, and the majority of twins recruited to the register are female. Twins enrolled onto the registry are not selected for disease-specific studies and have been shown to be representative of the UK population with respect to their anthropometric characteristics and the frequency of disease-related traits and lifestyle exposures (5, 6). Zygosity was determined by questionnaire and confirmed by multiplex DNA fingerprinting (PE Applied Biosystems, Foster City, CA). Ethical approval for this study has been obtained from St. Thomas’ Hospital Research Ethics committee, and informed consent was obtained from all subjects.
The participants in the present study were part of a sample invited to attend for clinical assessment at St Thomas’ Hospital between 1996 and 2000. All were female and between the ages of 18 and 80 yr. The sample selected for interview favored the inclusion of DZ twins for genetic linkage studies, but this selection was otherwise unrelated to other characteristics of the twins. Evaluation included measurement of fasting insulin and glucose data and the completion of a nurse-led interview. Birth weight was ascertained from the participants’ recall. BMI was calculated from height and weight measurements using the Quetelet index. Medication history was coded according to the British National Formula Number 40 (2000).
Laboratory methods
Serum samples were stored at –40 C until analysis. Fasting insulin was measured by immunoassay (Abbott Laboratories Ltd., Maidenhead, UK), and fasting glucose was measured on an Ektachem 700 multichannel analyzer using an enzymatic colorimetric slide assay (Johnson and Johnson Clinical Diagnostic Systems, Amersham, UK). Insulin resistance was calculated using the homeostasis model assessment (HOMA) (7). Subjects with insulin or glucose levels falling outside the calculable levels for HOMA were excluded from the analyses (n = 88).
Because body fat levels may increase after the menopause (8), we also adjusted for menopausal status. All subjects with either type 1 or type 2 diabetes (those diagnosed by a doctor, who were taking any form of diabetic medication, or those with fasting glucose levels of 
7.0 mmol/liter) were excluded (n = 35). The twin of any participant excluded for any reason was also excluded from these analyses. All statistical analyses were performed using STATA version 9.
Statistical methods
The insulin resistance data were log transformed to achieve a normal distribution. Birth weight and BMI measures were standardized giving a mean of 0 and SD of 1. As birth weight and later body size are themselves likely to be correlated, we examined a sequence of models including birth weight alone (model 1), BMI alone (model 2), and birth weight and BMI (model 3). This approach follows the sequence of models recommended by Lucas et al. (9) for analysis of the association between birth weight and adult characteristics. Model 3, which examines the association between early size adjusted for later size, is statistically equivalent to assessing how insulin resistance in the adult is influenced by change in size after birth. Effectively, this provides a measure how insulin resistance is influenced by centile crossing. We also tested for a birth weight and BMI interaction using a model including birth weight, BMI, and a birth weight and BMI interaction term.
Linear regression analysis was first undertaken treating the twins as individuals, allowing a direct comparison with findings in singleton populations: E(Yij) = β0 + βcXij, where Yij and Xij, respectively, represent insulin resistance (Y) value and birth weight (X) of twin j from pair i. βc represents expected change in insulin resistance per SD increase in birth weight in individuals. The regression analysis took into account the correlated structure (clustering by family) of the data. Second, following the approach described in detail by Carlin et al. (10), the effect of birth weight of each individual twin on insulin resistance was examined in a model parameterized with birth weight included as 1) a variable representing the mean birth weight of the pair from which the twin is derived and 2) a variable representing the individual twin’s difference from the pair mean. This approach provides a simultaneous estimation of within and between-pair influences of birth weight on insulin resistance: E(Yij) = β0 + βw(Xij – Xi) + βBXi, where Xi is the mean value of X for twin pair i. The within-pair coefficient βw gives the expected change in Y for a one-unit change in the difference between the individual X and the twin-pair average X value. The between-pair coefficient βB gives the expected change in Y for a one-unit change in the twin-pair average X, while holding the individual deviation from the average constant.
The within-pair effect βw represents an association that is free of confounding due to factors that are common to the twin pair (for example, shared intrauterine exposures, such as maternal smoking, or common factors in the shared family environment, such as social class). The between-pair effect βB reflects further variation in Y that can be explained by variation in the twin-pair mean of X. Variation due to confounding resulting from the maternal environment would be expected to be detected in βB but not βw (10).
The regression analyses were carried out in MZ and then DZ twins separately, with adjustment for age and menopausal status, and where no significant differences existed between zygosity groups, the data were pooled.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
). Mean birth weight was 2.45 kg, slightly lower than in comparable singleton populations, and was lower in MZ (2.30 kg) twins than in DZ twins (2.49 kg). All other anthropometric measurements were similar to comparable singleton populations (11, 12). There was no significant difference in insulin resistance between MZ and DZ twins.
|
shows scatter plots of insulin resistance and birth weight and BMI.
|
)The results from the individual-level regression analysis indicated that were was no significant relationship between birth weight and insulin resistance (model 1).
|
Within- and between-pair analysis (Table 3)
The analysis of model 1 confirms that there is no significant relationship between birth weight and insulin resistance, as indicated by the lack of significance for βw. In model 2, the relationship between adult BMI and insulin resistance is mediated through between- and within-pair effects, indicating that both individual-specific factors plus shared environmental factors exert an influence. When adult BMI is included in the birth weight and insulin resistance model (model 3), the relationship between birth weight and insulin resistance becomes significant between pairs only, and not within pairs. As in the analysis of the data overall, this provides evidence to support an influence of change in size.
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Previous research has highlighted that those who are small at birth but who exhibit excessive weight gain over the life course are at highest risk of developing insulin resistance and the metabolic syndrome (13, 14). We do not have multiple measures of growth in this cohort and cannot determine which period of growth is most influential on insulin resistance in later life. However, it has been shown that those who were born small are more likely to experience rapid weight gain in the first 2 yr of life (15), and evidence from a recent systematic review indicates that rapid weight gain in infancy is strongly related to overweight and obesity in adulthood (16). Evidence from animal studies (17) shows that catch-up growth as a result of being born small has lifelong effects on health including an influence on the age of maturity, lifespan, fertility, adult body size, the ability to maintain adequate energy reserves, the number of cells in organs, and the number of adult muscle fibers and on glucose tolerance and insulin regulation.
Limitations of the study include the reliance on recalled birth weight rather than more accurate prospective measures. However, previous research has shown that although there are some disadvantages in the use of recalled birth weight (18, 19, 20), mainly due to underestimation of birth weight in smaller babies, the majority of women (between 72 and 88%) can precisely recall the birth weight of their children. The use of recalled birth weight may therefore lead to underestimation of the strength of any associations found in the individual-level regressions. However, because there is no reason why there should be discordance in recall error within pairs, this should not have any real influence on the relationship within each twin pair.
Insulin resistance was measured indirectly by calculating HOMA. Although HOMA provides an indirect measurement of insulin resistance, HOMA scores do correlate highly with insulin resistance scores derived from euglycemic clamp tests (21). The intrauterine growth in twins may be different from that in singletons (22) as indicated by their lower birth weights. However, we have previously reported that the twins enrolled in the UK registry are representative of the singleton population for a wide range of anthropometric and disease-related characteristics (4, 5, 6).
In our data, a 1-SD decrease in birth weight was associated with a 3.3% increase in insulin resistance, an effect size that is comparable to those found in studies of singletons (3) including the Newcastle Thousand Families Study (23), the British Women’s Heart and Health Study (24), and the European Youth Heart Study (13). In these, the sizes of associations were all under 6%. Previous twin studies have also found similar effect sizes (25).
This is the largest twin study to date to investigate the relationships between birth weight and insulin resistance. It is the first to use the twin model to compare within- and between-pair differences simultaneously to separate maternal from individual-specific effects. Our data show that adult size is significantly related to insulin resistance in adult life and that these associations are mediated equally through both individual and shared influences. Any significant effects with birth weight are mediated purely through shared environmental factors.
|
Footnotes |
|---|
Disclosure Statement: The authors have nothing to disclose.
First Published Online November 20, 2007
Abbreviations: BMI, Body mass index; CI, confidence interval; DZ, dizygotic; HOMA, homeostasis model assessment; MZ, monozygotic.
Received June 26, 2007.
Accepted November 9, 2007.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  P. M. Skidmore, A. Cassidy, R. Swaminathan, J B. Richards, M. Mangino, T. D Spector, and A. J MacGregor An obesogenic postnatal environment is more important than the fetal environment for the development of adult adiposity: a study of female twins Am. J. Clinical Nutrition, August 1, 2009; 90(2): 401 - 406. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
