Over the last few decades, the contribution of the intrauterine environment to the development of chronic and non-communicable diseases has been highlighted.^1-3 Epidemiological studies have demonstrated that a poor intrauterine environment is associated with an increased risk of hypertension, chronic kidney disease, and/or diabetes. Since early 1980s, it has been hypothesized that nutrient deprivation during distinct periods of prenatal organ development programs the offspring for cardiovascular diseaselater in life.^1-3 Low birth weight (LBW), reflecting a poor intrauterine environment, is associated with diminished nephron endowment and other pathophysiological changes that may later lead to development of high blood pressure.^1,2,4

LBW has been associated with the development of non-insulin-dependent diabetes and with a central pattern of fat distribution, which is a component of the insulin resistance syndrome in adults.⁵ These observations led to the hypothesis that non-insulin-dependent diabetes and cardiovascular disease could be programmed by events in fetal life that lead to persistent changes in body composition and metabolic function.⁶

Currently, more than 75% of the disease burden is attributable to non-communicable diseases, with cardiovascular disease being the leading cause of death.⁷ The distribution of chronic diseases and related risk factors among the general population of Korea is similar to that of other industrialized nations: 12% of the population has diabetes, 30% is overweight, 20% are obese, 41% has high cholesterol, and 21% has metabolic syndrome.⁷

The prevalence of overweight and obesity is increasing at an alarming rate worldwide.^8-12 The prevalence of overweight (body mass index [BMI] ≥25 and <30 kg/m²) has remained remarkably stable in both men and women, while that of extreme obesity (BMI ≥40 kg/m²) has undergone a nine-fold increase from 0.9% in 1960−1962 to 8.1% in 2013−2014. This currently leaves only the remaining ~30% of the U.S. population as having a healthy weight (BMI between 18.5 and 25 kg/m²).^13,14 Similar findings have been reported in developing countries. Nationally representative surveys showed alarming rates of overweight and obesity. The prevalence of obesity (defined by BMI ≥30 kg/m²) was 18.4%, showing a 2.5-fold rise over a 20-year period. Similar surveys conducted in the last 20 years indicate the prevalence of obesity appears to have stabilized among men since 1995, but continues to increase in women.^13,14

Obesity is a multi-factorial disorder and birth weight is thought to be an important factor in its evolution;^8-12 however, the relationships between birth weight and overweight, obesity, and body fat distribution in adult life are not well understood.^1-3,15,16 Improving our understanding of the factors that lead to the development of overweight and central obesity is a major public health challenge as obesity is an important risk factor for metabolic diseases and cardiovascular disease.^8-12 Previous studies assessing the relationship between birth weight and adult overweight, fat mass (FM), or body fat distribution showed inconsistent results, as many of these studies used single measures of body fat or body fat distribution as outcomes in relation to birth weight.¹⁷ Some studies have shown J-shaped or linear associations between birth weight and BMI in childhood, while other studies have shown no significant relationships between birth weight and BMI.¹⁸ Also, other researchers have criticized the use of only the BMI, which is only an indirect measure of body fatness, as the outcome measure because it also includes lean and bone mass^17,18 Moreover, the generalizability of some of these studies is limited as they have been based on localized geographical populations, and have included people from specific ethnicities and/or professional groups.^19-21 To date, no study has assessed a cluster of anthropometric and body composition measurements in a nationally representative sample of the adult population.

Hence, this population-based cross-sectional nationally representative study aimed to investigate the relationships between birth weight and adult anthropometric and detailed body composition measurements (including adult body fat distribution and most indicators of adult body fatness) in men and women.

Study subjects

The detailed methodology of the Australian Diabetes, Obesity and Lifestyle Study (AusDiab) study had been discussed previously.^22,23 The study was approved and obtained via the Ethics Committee (No. 3/2002) of the International Diabetes Institute. All responders gave written informed consent to participate in the survey upon arrival at thetesting site. The AusDiab survey is a longitudinal study in which data were collected from a stratified sample of Australians aged 25 years or over, residing in 42 randomly selected urban and non-urban areas (Census Collector Districts) of the six states of Australia and the Northern Territory.²²

Methods and measurements

Questions addressing birth weight were included in the second round of the AusDiab study, which began in July 2004 and is still ongoing. Participants were asked to report their birth weight, assess the level of accuracy of their estimate, and identify the source of their birth weight data. Finally, participants were asked if they had any additional comments. The birth weight data were linked to the anthropometric findings and results of the baseline AusDiab survey.

At baseline, all participants except those who were (1) chairbound, (2) pregnant, or (3) too unsteady on their feet underwent anthropometric measurements while wearing light clothing and no footwear. The methods for obtaining these measurements have been described previously.^22,23 Briefly, height was measured to the nearest 0.5 cm using a stadiometer, and weight was measured using a mechanical beam balance, and was recorded to the nearest 0.1 kg. The BMI was calculated as weight (kg)/height (m)². The BMI groups were classified according to World Health Organization criteria²⁴ as follows: normal <25.0 kg/m², overweight 25.0–29.9 kg/m² or obese ≥30.0 kg/m². Waist and hip circumferences (HCs) were measured using a W606PM Lufkin steel measuring tape. For each of waist and HC, two measurements to the nearest 0.5 cm were recorded. If the variation between the measurements was greater than 2 cm, a third measurement was taken. The mean of the two closest measurements was calculated. The waist-to-hip ratio (WHR) was obtained by dividing the mean waist circumference (WC) by the mean HC.

The body can be divided into a FM component and fat-free mass (FFM) component for assessment purposes. The lean body mass (LBM) represents the weight of muscles, bones, ligaments, tendons, and internal organs. The FFM consists of minerals, protein, glycogen, and water, and therefore encompasses total intracellular and extracellular body water. The total body water (TBW) is the amount of water retained in the body. The TBW content of the adult human is approximately 60% of the body weight and is broadly divided into the intracellular and the extracellular fluid compartments. Generally, men tend to have higher water weight than woman due to a greater amount of muscle.

All subjects underwent bioimpedance measurements except those who (1) were chairbound, (2) were pregnant, (3) had a colostomy/ileostomy, (4) did not have a height measurement, or (5) weighed >150 kg. The scale for the bioimpedance machine (Tanita TBF 105 Body Fat Analyzer; Tanita Corp., Tokyo, Japan) was placed on a firm, flat surface and measurements completed while the participants wore light clothing with no shoes, socks, or hosiery. If the body fat percentage was greater than 70% or impedance <100, the process was repeated. If the second reading was within five percentage points of the first reading, the data from the second reading was recorded. If the second reading was not within five percentage points, the process was repeated until two consecutive readings within five percentage points were obtained. Once obtained, all data from the latter of these two readings were recorded. We analyzed the body fat further by using the recommendations of Wells and Victora²⁵, who suggested that whole-body adiposity is best assessed by calculating the indexes of FM to FFM (FM/FFM) and FM/FFM².

An interviewer-administered questionnaire was used to determine smoking, alcohol consumption, leisure-time physical activity and television viewing. The assessment of socioeconomic status was based on education, occupation, and income. We considered adult height, adult weight, WC, HC, LBM and TBW to be “low” if they were below the sex-specific 10th percentiles, and WC, waist-to-hip ratio, and body fat parentage (fat%) to be “high” if they were above the sex-specific 90th percentiles. The World Health Organization criteria²⁴ define obesity as follows: (1) a BMI of ≥30 kg/m² for both men and women; (2) a WC ≥1.02 m (102 cm) in men and in women of ≥0.88 m (88 cm), also termed abdominal obesity; and (3) a WHR above 0.90 for males and above 0.85 for females. In addition, FM was considered “high” when the result was >90th percentile, which was >48.6 kg for females, and >34.3 kg for males. The fat% was considered “high” when the result was >90th percentile, which was >54.7 kg for females; and >34.3 kg. for males. The FM/FFM ratio was considered “high” when the result was >90th percentile, which was >120.8 for females, and for male >52.2. The FM/FFM² was considered “high” when it was calculated to be >90th percentile, which was >3.12 for females, and >0.80 for males. The height was considered “low” when the result was <10th percentile, which was ≤155 cm for females, and ≤168 cm for males. The weight was considered “low” when the result was <10th percentile, which was ≤54 kg for females; and ≤69.6 kg for males. The HC was considered “low” when the result was <10th percentile, which was ≤92.8 cm for females, and ≤95.9 cm for males. Lean mass was considered low when the results was <10th percentile, ≤35.3 kg for females, and ≤56.0 kg for males. Water mass was considered “low” when the result was <10th percentile, which was ≤25.8 kg for females, and ≤41.0 kg for males. The birth weights, if recorded in pounds and ounces, were converted to kilograms for the statistical analyses. LBW was defined as a birth weight <2.5 kg. Birth weight was also divided equally into quintiles for further categorical analyses.

Statistical analyses

All analyses were performed using the Stata software program (Stata Corp., College Station, TX, USA). The birth/weights, if recorded in pounds and ounces, were converted to kilograms for statistical analyses. LBW was defined as birth weight <2.5 kg. Birth weight was also divided equally into quintiles for further categorical analyses.

Apart from the FM, FM/FFM and FM/FFM² variables, the remaining variables were approximately normally distributed. Student t-test was used to assess the differences in anthropometric and body composition measurements among those who reported their birth weight and those who did not. We examined age-adjusted anthropometric measurements by sex-specific quintiles, and the multivariate-adjusted means of fat% and FM/FFM² by were examined by sex-specific categories. In multivariate analyses, we adjusted for age, BMI, physical activity, smoking status, alcohol intake and socioeconomic status (based on education, income, and dwelling type). We used linear regression to assess the strength of the associations between birth weight (per 1 kg increase) and adult height, weight, WC, HC, LBM, TBW, WHR, BMI, and body fat indices, in females and males separately. We used logistic regression to calculate the age-adjusted odds ratios for: (1) obesity according to BMI, WC, and WHR, (2) “low” height, weight, HC, LBM, and TBW, and (3) “high” FM, fat%, FM/FFM, and FM/FFM² among people with LBW relative to those with normal birth weight (NBW). No significant interaction terms between various covariates were identified in our models.

Of the 7,157 respondents to our questionnaire, 4,502 reported information related to their birth weight. Their birth weights ranged from 0.4 to 7.0 kg with a mean and standard deviation (SD) of 3.37±0.7 kg, as shown in Fig. 1. The average values of 3.35±0.6 vs. 3.37±0.7 kg were nearly equal for those who obtained their birth weight from family members and for those who obtained it from medical records, respectively, after adjustment for age and sex (P=0.36). The mean birth weight of females (3.28 kg; SD, 0.6 kg) was lower than that of males, (3.5 kg; SD, 0.7 kg). The prevalence of LBW (<2.5 kg) was 10% and 6% in females and males, respectively.

Table 1 presents the comparison of relevant characteristics between participants who reported their birth weight and those who did not. Those who did not report their birth weight were older, shorter and were more likely to have higher WHR and lower LBM. Table 2 presents the participants’ anthropometric data by quintiles of birth weight for females and males. Females in the lowest birth weight quintile had the lowest mean height, weight, LBM, FM, fat%, FFM, and TBW, but no differences were apparent with respect to BMI or WC. Males in the lowest birth weight quintile had lower mean height, weight, BMI, LBM, TBW, WC, HC, FM, fat%, and FFM than those of higher birth weight quintiles; however, WHR did not differ by birth weight quintile. The differences according to birth weight quintiles persisted even after adjustment for BMI, physical activity, smoking, alcohol intake and socioeconomic status. Both females and males with LBW had the highest fat% after adjustment for age, BMI, physical activity, smoking status, alcohol intake and socioeconomic status. This relationship was even more pronounced when the relationship between birth weight and FM/FFM² was examined.

When applying the traditional definition of LBW (<2.5 kg), as shown in Table 3, females with LBW were shorter, had higher WC, WHR, FM, fat%, and FFM, and lower LBM and TBW than females with NBW (≥2.5 kg); however, there were no significant differences with respect to current body weight, HC or BMI. Males with LBW were shorter, had lower LBM, weight, BMI, LBM, TBW, WC, HC, FM, fat%, and FFM than those with NBW, although there no significant differences with respect to WC and WHR. Overall, compared to people with birth weight ≥2.5 kg, the indices of body fatness we assessed were significantly greater in LBW females than in males.

We examined the predicted associations of each kilogram increase in birth weight with body size and composition. Among females, for each kg of birth weight, there was a predicted increase (95% confidence interval [CI]) of 2.0 cm (1.7–2.4 cm) in height, 1.70 kg (0.8–2.6 kg) in weight, 1.67 kg (1.4–2.0 kg) in LBM, 1.23 kg (1.0–1.4 kg) in TBW, with P<0.001 for all. However, there was a predicted decrease of 0.01 unit (–0.01 to –0.002) in WHR (P=0.003), 0.72% (–1.34 to –0.1) in fat% (P=0.023), and 0.14% (–0.21% to –0.07%) in FM/FFM² (P<0.001). Among males, for each kg of birth weight, there was an increase of 2.3 cm (1.9–2.8 cm) in height, 3.8 kg (2.8–4.7 kg) in weight, 1.5 cm (0.8–2.3 cm) in WC, 1.36 cm (0.8–1.9 cm) in HC, 2.3 kg (1.8–2.7 kg) in LBM, and 1.7 kg (1.4–2.0 kg) in TBW, with P<0.001 for all.

Table 4 shows that the obesity markers WC, WHR, and BMI were 47%, 61%, and 45% greater, respectively, among females with LBW compared to normal birth weight females, together with 36%, 58%, 64%, and 56% increases in the risks of high (>90th percentile) FM, fat%, FM/FFM and FM/FFM2, respectively. Both females and males with LBW had 1.7–2.8 times the risk for having lower height, weight, HC, lean mass, and TBW relative to those with normal birth weight.

This is the first study to examine the associations of birth weight with measures of adult body size and composition in a nationally representative sample of the general population. The results indicated significant associations of birth weight with adult body habitus among the general population. Lower birth weights were associated with lower height, weight, LBM and TBW in both females and males. LBW was also associated with greater central fat and total body fat.

Using the conventional definition of LBW (<2.5 kg), highlighted the sex differences: all measures of LBM were lower in LBW participants of both sexes, including the prevalence of central obesity; in addition, body fatness as assessed using several different indices, including BMI, FM, fat%, FFM, FM/FFM, and FM/FFM², was significantly increased in LBW females and males compared with those of higher birth weights.

Fetal development is one of the critical periods with regard to adult obesity. The association of birth weight with LBM is consistent with the theory that restricted intrauterine nutrition limits cell division and cell growth and modifies fetal organ structure,²⁶ and the development of muscle mass, in particular; it is also thought these adverse effects may not be fully reversed by subsequent improvements in nutrition.^27-29 It is possible that factors influencing birth weight at term may also influence storage of fat in later life.^16,30-33 However, these changes or modification may be qualitative, or quantitative and differ by sex.

Several studies have suggested an inverse association between birth weight (or famine exposure during early gestation) and various measures of abdominal obesity in childhood or adult life.^34,35 Our results, like those of Ravelli et al.,³⁶ confirmed this phenomenon was more pronounced in adult females with LBW than in males. This relationship was independent of current body size and persisted after adjustment for all major confounders.

Additional studies of newborns have suggested that preterm small-for-gestational age infants store excess calories as fat, and their protein reserves in the form of muscle mass remain low.^37-39 Previous research during a wartime famine in the Netherlands showed that under-nutrition in early pregnancy resulted in increased rates of obesity in males at 19 years of age, whereas under-nutrition in the third trimester or early postnatal life resulted in a reduced likelihood of obesity.^40,41 However, subsequent follow-up of the Dutch men at 50 years of age showed that this effect on adiposity did not persist.³⁶ These findings might explain the sex differences observed in our study, where males with LBW did not exhibit significantly higher adiposity or central obesity. The basis for restriction of the phenomenon to females or its accentuation in females is poorly understood, but probably reflects a survival or reproductive advantage.^40,41 One study found that body-fat percentage and leptin concentrations were positively associated with birth weight.⁴² Other studies have shown that leptin levels were higher in women than in men, and women had greater body fat and FM than men. It has been suggested that this discrepancy in body composition may be due to the increased deposition of subcutaneous fat in female newborns, such that sex differences in body composition are already present in newborns.⁴³

Lower birth weights predispose humans to a variety of chronic diseases and their risk factors (diabetes, hypertension, cardiovascular diseases, chronic kidney disease, metabolic syndrome, chronic pulmonary disorders, and osteoarthritis).^1-3 We have already described most of these phenomena in the AusDiab Birth weight cohort.^15,16,44,45 In view of the correlations of these conditions with measures of central fat, the greater susceptibility of those with lower birth weight to chronic disease could be mediated, in part, through their relative preservation or amplification of central body fat.⁴⁶ The greater predisposition to chronic disease among lower birth weight females is also compatible with the different influences of birth weight on central fat deposition in males and females.^1-3,15,16

Limitations

The present study was limited by not including the gestational age in the analyses, and thus stratification of the participants’ data by those who had been small-for- gestational age, appropriate-for-gestational age and large-for-gestational age at birth in comparison with the general population was not possible. While many small-for gestational age babies tend to catch up during infancy, the study did not have information regarding the rate of growth during first few years of life.

Conclusions

The findings of this study may have implications for the prevention and management of renal disorders in any country where the incidence of LBW is increasing and the affected newborns survive. Advancements in intensive care and general medical care over time have allowed more lower-birth weight infants to survive to adult life. In all populations, a worldwide trend towards higher levels of body fat and BMI potentially compounds the effects of other risk factors such as the expression of glycemic abnormalities associated with lower birth weights. Modest increases in body fat may have a trivial impact on the burden of metabolic and renal diseases when acting in isolation, but may have substantial impact when combined with other risk factors. It would therefore be prudent to adopt policies of intensified whole-of-life surveillance of lower-birth weight people in anticipation of the potential risks. Also, as the earliest known risk factor for renal disease, consideration of LBW among people in more developed countries could be used in risk stratification for early identification of renal disease or its risk factors. The long-term health outcomes for LBW infants are of potential concern and may guide point-of-care decisions for further testing and management selection that sets a platform for risk reduction based on biological platform stratification.

These findings contribute to our understanding of the determinants of adult body habitus, and inferentially, of their influence on adult health profiles. From a public health perspective, these findings indicate LBW may play an important role in the predisposition to dysglycemia and various metabolic abnormalities.

The authors declare no conflict of interest.

We thank the participants, Survey Team, and Steering Committee of the AusDiab Study for their support of and assistance with the birth weight project since 2004. The AusDiab study is co-coordinated by the Baker IDI Heart and Diabetes Institute.

We have no conflicts of interest to declare; however, the Ausdiab study was provided financial support by the Commonwealth Department of Health and Aged Care, Abbott Australasia, Alphapharm, Aventis Pharmaceutical, AstraZeneca, Bristol-Myers Squibb Pharmaceuticals, Eli Lilly (Australia), GlaxoSmithKline, Janssen-Cilag (Australia), Merck Lipha, Merck Sharp & Dohme (Australia), Novartis Pharmaceutical (Australia), Novo Nordisk Pharmaceutical, Pharmacia and Upjohn, Pfizer, Roche Diagnostics, Sanofi Synthelabo (Australia), Servier Laboratories (Australia), BioRad Laboratories, HITECH Pathology, the Australian Kidney Foundation, Diabetes Australia, Diabetes Australia (Northern Territory), Queensland Health, the South Australian Department of Human Services, the Tasmanian Department of Human Services, the Victorian Department of Human Services, and the Health Department of Western Australia.

Study concept and design: ISA; acquisition of data: ISA; analysis and interpretation of data: all authors; drafting of the manuscript: all authors; critical revision of the manuscript: SH; statistical analysis: ISA; obtained funding: ISA; administrative, technical, or material support: all authors; and study supervision: all authors.