Early Origins of the Metabolic Syndrome: Role of Small Size at Birth, Early Postnatal Weight Gain, and Adult IGF-I

Abstract

Background:

The relationship between low birth weight and increased risk for metabolic syndrome (MetS) in later life has been frequently described, but mechanisms underlying this association remain unknown.

Methods:

In 280 young adults of the PROGRAM study, aged 18–24 yr, we investigated associations of birth weight, gain in weight for length during early life, and adult IGF-I sd score (SDS), with number of MetS components (ordinal regression analyses), prevalence of MetS components and MetS (logistic regression analyses), and other metabolic parameters (linear regression analyses). Revised criteria of the National Cholesterol Educational Program (Adult Treatment Panel III) were used to determine components of MetS. The other metabolic parameters were C-reactive protein, insulin sensitivity, trunk fat mass, total cholesterol, and low-density lipoprotein cholesterol.

Results:

More gain in weight for length SDS in the first 3 months of life was significantly associated with an increased number of MetS components [odds ratio (OR) = 1.34], prevalence of low high-density lipoprotein cholesterol (OR = 1.49), prevalence of MetS (OR = 2.51), increased C-reactive protein levels, and lower insulin sensitivity (P = 0.007) at the age of 21 yr. Low birth weight SDS was associated with lower insulin sensitivity (P = 0.036), but low birth weight SDS and adult IGF-I SDS were not significantly associated with any of the MetS components or MetS prevalence at 21 yr.

Conclusion:

Our study demonstrates that higher gain in weight for length in the first 3 months of life is associated with a higher prevalence of MetS at 21 yr, whereas low birth weight and low adult IGF-I are not.

The relationship between low birth weight and increased risk for metabolic syndrome (MetS) in later life has been frequently described (1–3). Adults born small for gestational age (SGA) have lower insulin sensitivity and higher abdominal fat mass, and lipoprotein levels are more often disturbed, all contributing to a higher prevalence of MetS (4). The majority of children born SGA show catch-up growth within 2 yr after birth, resulting in a normal stature in childhood and adulthood (5). Because accelerated gain in weight for length in early life has been associated with adverse health profile in adulthood (6–9), it might well be that the association between small size at birth and MetS in later life can be ascribed to accelerated early weight gain.

Another factor contributing to the relationship between SGA birth and later risk for MetS might be IGF-I. Decreased serum levels of IGF-I have been reported in adults born SGA (10). Furthermore, low IGF-I levels have been associated with each of the components of MetS [waist circumference, triglycerides (TG), high-density lipoprotein (HDL) cholesterol (HDL-c), blood pressure, and fasting glucose] (11, 12).

Our aim was to unravel mechanisms involved in the association between small size at birth and components of MetS in early adulthood. We hypothesized that, in contrast to lower weight at birth, accelerated early weight gain for length, and/or lower IGF-I levels are associated with an increased risk for MetS. We therefore investigated associations of birth weight, first-year gain in weight for length, and serum IGF-I levels with MetS components according to revised criteria of the National Cholesterol Educational Program (NCEP) Adult Treatment Panel III (11). MetS criteria were defined for use in clinical practice; however, additional metabolic parameters are also relevant to study with regard to cardiovascular disease (CVD) risk, such as insulin sensitivity determined by frequently sampled iv glucose tolerance test and trunk fat mass determined by dual-energy x-ray absorptiometry scan.

Subjects and Methods

Subjects

The PROGRAM study cohort consists of 323 healthy participants with an age between 18 and 24 yr. Participants were recruited from hospitals in The Netherlands, where they had been registered because of being small at birth (SGA with a birth length < −2 sd (n = 102)] (13) or showing short stature [with an adult height < −2 sd after being born SGA (n = 42 of 102) or appropriate for gestational age (n = 40)] (14). In addition, healthy subjects (neither small at birth nor having short stature) from schools with different educational levels were randomly asked to participate. This design was purposely chosen to increase the contrast within the study population regarding birth size and adult stature. All participants fulfilled the same inclusion criteria: 1) age 18–24 yr, 2) born singleton, 3) born at term (≥36 wk of gestational age), 4) Caucasian, 5) uncomplicated neonatal period without signs of severe asphyxia (defined as an Apgar score below 3 after 5 min) and without sepsis or long-term complications of respiratory ventilation such as bronchopulmonary dysplasia, 6) maximum duration of respiratory ventilation and/or oxygen supply in the neonatal period of 2 wk. Subjects were excluded if they had been suffering from any serious complication or condition (including necrotizing enterocolitis, intraventricular hemorrhage with a degree of 3 or more, spastic hemiplegia, or quadriplegia) or from any disease or had received any treatment known to interfere with growth (e.g. GH deficiency, severe chronic illness, emotional deprivation, GH treatment, or treatment with glucocorticosteroids or radiotherapy) or if they had endocrine or metabolic disorders, chromosomal defects, syndromes, or serious dysmorphic symptoms suggestive for a yet unknown syndrome.

Birth data were taken from records of hospitals, community health services, and general practitioners. Information regarding socioeconomic status (SES) was obtained using questionnaires. Education level of the participant was used as socioeconomic indicator to determine SES (15). In The Netherlands, periodical measurements of weight and length are performed for each child. Thus, weight and length at 3, 6, 9, and 12 months after birth had been measured prospectively at primary healthcare centers or hospitals. These data were collected from the records of the healthcare centers and hospitals during the study period March 2006 to September 2007.

The Medical Ethics Committee of Erasmus Medical Center, Rotterdam The Netherlands, approved the study. Written informed consent was obtained from all participants. Of 323 subjects, data on each component of the metabolic syndrome were available for 280 subjects (of whom n = 87 were born SGA, and n = 70 had short stature). Of these 280 subjects, data on the first 3 months growth were available for 184 subjects.

Measurements

All participants visited the Erasmus Medical Center in Rotterdam and were reimbursed for travel expenses. Before the visit, participants fasted for at least 12 h and abstained from smoking and drinking alcohol for at least 16 h. All anthropometric measurements were performed twice; the mean value was used for analyses.

Height was measured to the nearest 0.1 cm by a Harpenden stadiometer and weight to the nearest 0.1 kg by a scale (Servo Balance KA-20-150S). Lean body mass and fat mass were measured on one dual-energy x-ray absorptiometry machine (Lunar Prodigy; GE Healthcare, Chalfont St. Giles, England, UK) (16). Insulin sensitivity index (capacity of insulin to promote glucose disposal) was determined using the Bergman minimal model (MinMod Millenium version 6.01, MinMOD Inc., Los Angeles, CA), which calculated the paired glucose and insulin data obtained by frequent measurements during an iv glucose tolerance test (17–19) with tolbutamide (20). Blood pressure was measured after 10 min at rest, in the sitting position, using the nondominant arm with an automatic device (Accutorr Plus; Datascope Corp., Montvale, NJ) three times with 5 min in between, and the mean value was taken to reflect resting blood pressure.

Assays

All fasting blood samples were drawn between 0800 and 1300 h, centrifuged after clotting, and were kept frozen until assayed (−80 C). Briefly, plasma glucose levels were determined on a VITROS analyzer 750 (Ortho-Clinical Diagnostics, Johnson & Johnson Co., Beerse, Belgium). Plasma insulin levels were measured using an immunoradiometric assay (Medgenix Diagnostics, Fleunes, Belgium). TG was measured using an automated enzymatic method with the GPO-PAP reagent kit (Roche Diagnostics, Mannheim, Germany). HDL-c level was measured using a homogeneous enzymatic colorimetric assay (Roche Diagnostics). Low-density lipoprotein (LDL)-c was calculated using the Friedewald formula: LDL-c = total cholesterol − HDL-c − 0.45 × TG. For high-sensitivity C-reactive protein (CRP), an in-house high-sensitivity ELISA with polyclonal rat CRP antibodies for catching and tagging (Dako, Glostrup, Denmark) was used. Serum IGF-I and IGF-binding protein 3 (IGFBP-3) levels were measured in one laboratory using an automated chemiluminescence immunometric assay (Immulite-1000 systems; Siemens Healthcare Diagnostics, Tarrytown, NY). Serum levels were expressed as sd score (SDS) to adjust for age and sex using reference data from a healthy Dutch population (21). The assays have been previously described in detail (22).

Statistical analysis

SDS for birth length, birth weight, and first-year weight and length were calculated to correct for gestational age and sex (13). SDS for adult height and weight were calculated to correct for sex and age (14). Revised criteria of the NCEP Adult Treatment Panel III were used to determine components of MetS (11, 23). MetS was defined as having three or more of the following risk factors: abdominal obesity (waist circumference for men >102 cm and for women >88 cm), TG at least 1.7 mmol/liter, HDL-c for men at least 1.03 and for women at least 1.3 mmol/liter, blood pressure at least 130/85 mm Hg, and fasting glucose of at least 5.6 mmol/liter.

Ordinal regression analyses were performed to determine associations of birth weight SDS, weight gain during the four 3-month periods in the first year of life, and IGF-I SDS, with the number of MetS components per individual. All regression analyses were adjusted for age, gender, gestational age, and SES. We additionally adjusted for smoking and alcohol use (after removing SES), which did not change our results (data not shown). Weight gain analyses were additionally adjusted for gain in length SDS in the same period, and IGF-I analyses were additionally adjusted for IGFBP-3. We additionally tested whether adult height SDS was a significant confounder in the associations studied, because IGF-I is related to adult height, and subjects with short stature were oversampled in the study population.

Next, logistic regression analyses were performed to investigate associations between birth weight SDS, gain in weight for length in the first 3 months of life, and IGF-I SDS with prevalence of each of the components of MetS and prevalence of MetS (three or more of the components).

Finally, to identify associations of birth weight SDS, gain in weight for length in the first 3 months of life, and IGF-I SDS with other metabolic parameters, we performed linear regression analyses with the dependent variables CRP, insulin sensitivity, trunk fat mass, total cholesterol, and LDL-c.

Statistical package SPSS version 17.0 (SPSS Inc., Chicago, IL) was used for analyses. Results were regarded statistically significant if P < 0.05.

Results

Clinical characteristics, components of the MetS, and additional metabolic parameters are shown in Table 1. MetS, according to the revised NCEP criteria, was present in 5.4% of all participants.

Table 1.

Clinical characteristics and components of the MetS in the study group

	n	Mean (sd)
Clinical characteristics
Age	280	20.9 (1.6)
Gender (male/female)	280	112/168
Gestational age	280	39.2 (1.7)
Birth length SDS	280	−1.48 (1.5)
Birth weight SDS	280	−1.18 (1.4)
3-month length SDS	183	−1.13 (1.25)
6-month length SDS	182	−1.03 (1.20)
9-month length SDS	181	−1.00 (1.27)
12-month length SDS	179	−1.02 (1.21)
3-month weight SDS	184	−0.96 (1.17)
6-month weight SDS	183	−0.91 (1.09)
9-month weight SDS	181	−0.88 (1.13)
12-month weight SDS	175	−0.79 (1.08)
Adult height SDS	280	−1.05 (1.37)
Adult weight SDS	280	−0.60 (1.42)
IGF-I SDS	280	−0.32 (0.83)
IGFBP-3 SDS	280	−1.03 (0.68)
MetS components
Waist circumference (cm)	280	77.2 (9.98)
% higha	280	8.6
TG (mmol/liter)	280	1.03 (0.51)
% higha	280	10.4
HDL-c (mmol/liter)	280	1.37 (0.37)
% lowa	280	34.6
Systolic BP (mm Hg)	280	119.6 (11.5)
% higha	280	17.1
Diastolic BP (mm Hg)	280	74.1 (7.62)
% higha	280	8.2
Glucose (mmol/liter)	280	4.91 (0.47)
% higha	280	7.9
MetS (%)a	280	5.4
Additional metabolic parameters
CRP (mg/liter)	273	3.79 (6.6)
Insulin sensitivity (μU/ml)	130	6.82 (4.4)
Trunk fat (kg)	273	7.65 (4.7)
Total cholesterol (mmol/liter)	279	4.54 (0.9)
LDL-c (mmol/liter)	279	2.70 (0.8)

BP, Blood pressure.

^aBased on revised criteria of the NCEP.

Table 2 shows results of ordinal regression analyses. Gain in weight SDS in the first 3 months of life was associated with an increased number of MetS components at the age of 21 yr [odds ratio (OR) = 1.34, 95% confidence interval (CI) = 1.01–1.78], adjusted for age, gender, gestational age, SES, and gain in length in the same period. Thus, per one SDS increase in weight gain, the chance of having a higher number of MetS components increases by 34%. There were no significant associations of birth weight SDS, weight gain during the other 3-month periods in the first year of life, or serum IGF-I SDS at 21 yr with number of MetS components. Subjects with short stature were oversampled in the study population. We, therefore, additionally adjusted for adult height SDS. After adjustment for adult height SDS, results were similar, and adult height SDS itself was not a significant determinant of the number of MetS components.

Table 2.

Associations of birth size, first-year gain in weight for length SDS, and adult IGF-I SDS with the number of MetS components per individual

	Number of components MetS
	OR	95% CI
Birth weight SDSa	1.09	0.91–1.30
Gain in weight SDS
0–3 monthsb	1.34	1.01–1.78
3–6 monthsb	1.10	0.64–1.89
6–9 monthsb	0.80	0.37–1.73
9–12 monthsb	0.47	0.19–1.13
IGF-I SDSc	1.06	0.78–1.45

^aAdjusted for age, gender, gestational age, and SES.

^bAdjusted for age, gender, gestational age, SES, and gain in length SDS in the same period.

^cAdjusted for age, gender, gestational age, SES, and IGFBP-3 SDS.

Table 3 shows associations of birth length SDS and birth weight SDS, gain in weight adjusted for length from birth up to 3 months of age, and IGF-I SDS at 21 yr with prevalence of each of the MetS components and prevalence of MetS (having three or more components). There were no associations of low birth weight SDS with higher prevalence of any of the MetS components.

Table 3.

Associations of birth size, early weight gain, and IGF-I, with MetS components and prevalence of MetS

	High waist circumference		High TG		Low HDL-c		High BP		High glucose		MetS
	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI
Model 1
Birth weight SDS	0.760	0.439–1.316	0.933	0.590–1.477	1.091	0.810–1.470	0.995	0.702–1.410	1.269	0.753–2.139	0.492	0.224–1.080
Birth length SDS	1.124	0.714–1.767	0.972	0.643–1.469	0.945	0.719–1.242	1.097	0.790–1.524	0.743	0.445–1.242	0.795	0.600–1.948
Model 2
Δ weight 0–3 months	1.618	0.921–2.841	1.336	0.794–2.248	1.485	1.060–2.080	1.022	0.715–1.459	0.911	0.517–1.607	2.509	1.200–5.247
Δ length 0–3 months	0.998	0.533–1.867	1.022	0.569–1.835	0.843	0.572–1.241	0.952	0.615–1.473	1.366	0.675–2.764	0.987	0.432–2.255
Model 3
Adult IGF-I SDS	0.656	0.373–1.151	0.664	0.395–1.116	1.069	0.752–1.520	1.476	0.930–2.341	1.820	0.861–3.849	0.857	0.409–1.795

For model 1, birth weight SDS and birth length SDS are entered simultaneously. For model 2, gain in weight from birth until 3 months of age (Δ weight 0–3 months) and gain in length SDS from birth until 3 months of age (Δ length 0–3 months) are entered simultaneously. All analyses are additionally adjusted for gender, age, SES, and gestational age. IGF-I SDS analyses are additionally adjusted for IGFBP-3. BP, Blood pressure.

Gain in weight relative to length in the first 3 months of life was significantly associated with a higher prevalence of low HDL-c (OR = 1.49; 95% CI = 1.06–2.08), and prevalence of MetS (OR = 2.51; 95% CI = 1.20–5.25). When we additionally adjusted for adult height SDS in the analyses including birth weight and early weight gain, results were similar. Furthermore, adult height SDS itself was not a significant determinant.

There was no significant association of IGF-I SDS with any of the MetS components, also after adjustment for IGFBP-3 SDS. However, after additional adjustment for adult height SDS, we found a borderline significant positive association of IGF-I SDS with increased fasting glucose levels (OR = 2.013; 95% CI = 0.950–4.264; P = 0.068). In that model, adult height SDS itself was inversely associated with prevalence of increased fasting glucose levels (OR = 0.643; 95% CI = 0.427–0.967; P = 0.034).

In Table 4, results from linear regression analyses are shown with additional metabolic parameters as dependent variables, namely CRP, insulin sensitivity, trunk fat mass, total cholesterol, and LDL-c. Lower birth weight SDS (16.8% lower insulin sensitivity per SDS decrease, P = 0.036) as well as accelerated gain in weight SDS in the first 3 months of life (21% lower insulin sensitivity per SDS increase in weight gain, P = 0.007), showed significant associations with lower insulin sensitivity at 21 yr.

Table 4.

Associations of birth size, early weight gain, and IGF-I, with MetS related parameters

	CRP			Insulin sensitivity			Trunk fat mass			Total cholesterol			LDL-c
	β (%)	P	Adjusted R2	β (%)	P	Adjusted R2	β (%)	P	Adjusted R2	β (%)	P	Adjusted R2	β (%)	P	Adjusted R2
Model 1
Birth weight SDS	−5.32	0.159	0.184	16.8	0.036	0.055	−5.32	0.181	0.114	−0.69	0.607	0.152	−1.61	0.430	0.057
Birth length SDS	2.32	0.625		−4.73	0.502		2.32	0.540		−1.25	0.310		−1.38	0.463
Model 2
Δ weight 0–3 months	30.9	0.009	0.216	−21.0	0.007	0.072	8.11	0.064	0.153	−0.28	0.844	0.150	1.54	0.494	0.049
Δ length 0–3 months	−0.16	0.989		19.2	0.085		−0.08	0.986		2.96	0.085		1.20	0.655
Model 3
Adult IGF-I SDS	−16.6	0.133	0.188	−8.71	0.262	0.047	−2.97	0.538	0.107	1.80	0.269	0.153	3.96	0.115	0.051

For model 1, birth weight SDS and birth length SDS are entered simultaneously. For model 2, gain in weight from birth until 3 months of age (Δ weight 0–3 months) and gain in length SDS from birth until 3 months of age (Δ length 0–3 months) are entered simultaneously. All analyses are additionally adjusted for gender, age, SES, and gestational age. IGF-I SDS analyses are additionally adjusted for IGFBP-3. β, Regression coefficient.

There was no association of birth weight SDS with CRP, trunk fat mass, total cholesterol, or LDL-c.

Accelerated gain in weight SDS in the first 3 months of life was associated with higher levels of CRP (30.9% higher CRP per SDS increase in weight gain, P = 0.009) and was borderline significantly associated with trunk fat mass (8.11% higher trunk fat mass per SDS increase in weight gain, P = 0.064), adjusted for gender, age, SES, and gestational age. IGF-I was not associated with any of the additional metabolic parameters. Additional adjustment for adult height SDS did not change the results.

Discussion

In this study, we investigated mechanisms involved in the reported association of small size at birth with MetS in later life. We studied associations of low birth weight, early life gain in weight for length, and adult serum levels of IGF-I with components of MetS, prevalence of MetS (having three or more of the components), and several other metabolic parameters. Our results imply that gain in weight relative to length SDS in the first 3 months of life is the most important determinant of MetS, because that was associated a higher number of MetS components at the age of 21 yr, in contrast with birth weight SDS, weight gain during the other 3-month periods in the first year of life, and adult IGF-I levels. Furthermore, gain in weight relative to length SDS in the first 3 months of life was associated with increased prevalence of MetS, higher prevalence of low HDL-c, increased levels of CRP, and lower insulin sensitivity in early adulthood.

To our knowledge, this is the first study investigating the relationship of small size at birth, accelerated weight gain in early life, and adult IGF-I SDS with MetS and several other metabolic parameters in early adulthood. The association of small size at birth with MetS has been frequently described, but often postnatal weight gain and other factors associated with SGA birth, such as serum level of IGF-I, have not been taken into account. Beardsall et al. (24) previously reported that postnatal weight gain from birth, rather than birth weight, was associated with risk factors for metabolic diseases in childhood. Our study shows that more gain in weight than length SDS in the first 3 months of life is related to prevalence of the MetS, and low birth weight SDS is not. Furthermore, low birth weight SDS was not related to any of the components of MetS and other metabolic parameters, except for an association with lower insulin sensitivity. This association is, however, possibly due to catch-up growth after being born small, because increased gain in weight for length from birth to 3 months of age, which often follows small size at birth, was also more strongly associated with decreased insulin sensitivity. We previously showed that subjects born SGA with catch-up growth have lower insulin sensitivity, which was not found in those without catch-up (25).

Increased gain in weight for length from birth to 3 months of age was also associated with one of the components of MetS, namely with prevalence of decreased HDL-c levels. Furthermore, it was associated with increased levels of CRP, which has been previously associated with both MetS and low HDL-c (26). Our findings are in line with previous studies (6, 27) and suggest that an imbalance in neonatal gain in weight compared with length after birth should be avoided to reduce the risk for MetS in later life. Animal studies showed that early-life catch-up in weight is associated with skeletal muscle insulin resistance and adipose tissue insulin hyperresponsiveness accompanied by suppressed thermogenesis (28, 29). The authors hypothesized that this phenomenon exists for the purpose of sparing glucose to catch-up in fat, which might be the link between early-life accelerated weight gain and risk for later MetS (28).

The term MetS refers to a clustering of risk factors, being a pathophysiological condition underlying CVD and type 2 diabetes (30). Our study population consists of healthy young adults, and it was not possible to study hard endpoints such as CVD and type 2 diabetes. Therefore, we used MetS as an outcome variable. One limitation of this approach is the arbitrary character of the MetS definition and the subsequent possibility to miss critical information, which has been debated (30, 31). We, therefore, decided to also study the components of MetS separately, and in addition to investigate CRP, insulin sensitivity, trunk fat mass, total cholesterol, and LDL-c, which are also determinants of CVD (32–35).

Previous studies reported an association between IGF-I and components of the MetS (12, 36, 37). We, however, did not find significant associations of adult IGF-I SDS with any of the components or prevalence of MetS. We also did not find a significant association between IGF-I and insulin sensitivity. It might, however, still be that lower IGF-I levels affect the risk for developing MetS, but we could not prove this because the variation in IGF-I levels was not large enough in our study population of health young adults. Although we oversampled subjects with short stature, who are likely to have lower levels of IGF-I, we did not include subjects with known GH deficiency.

Adults with short stature have a higher a priori chance of having a smaller waist circumference than normal statured subjects. The NCEP criteria of the MetS do not take this into account, because waist circumference is defined as one of the MetS components. Therefore, in the present study we also investigated associations of birth size, early weight gain, and adult IGF-I with MetS after adjustment for adult height SDS. Shorter stature was associated with a higher prevalence of increased fasting glucose levels, when adjusted for IGF-I and IGFBP-3. An explanation of this finding might be the adverse effect of low/subnormal GH levels on glucose homeostasis (38, 39), assuming that the subjects with short stature in our study population have lower levels of GH that those with normal stature.

In conclusion, our study indicates that the reported association of small size at birth with MetS in adulthood is mainly due to an increased gain in weight for length in the first 3 months after birth. Although our results show the importance of balanced weight gain during the early postnatal period, we recognize that there might be other critical windows later in childhood that remained unstudied in the present study. Our findings point to the need to investigate the optimal target of postnatal weight gain after birth in all infants, regardless whether they are born SGA or appropriate for gestational age.