| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Hospital for Children and Adolescents (E.K., L.D., S.A.) and Department of Obstetrics and Gynecology (M.S., R.K., S.A.), Helsinki University Central Hospital, 00029 HUS, Helsinki, Finland; National Public Health Institute (E.K., H.Y., T.F., J.E.), 00300 Helsinki, Finland; and MRC Environmental Epidemiology Unit (C.H.D.F., C.O., D.J.P.B., R.I.G.H., D.I.W.P.), Southampton General Hospital, SO16 6YD Southampton, United Kingdom
Address all correspondence and requests for reprints to: Eero Kajantie, M.D., The Hospital for Children and Adolescents, Helsinki University Central Hospital, PL 280, 00029 HUS, Finland. E-mail: eero.kajantie{at}hus.fi.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
0.05) and inversely with percentage body fat (r = -0.13, P = 0.01) and waist circumference (r = -0.11, P = 0.03). IGFBP-1 was inversely correlated with adult body mass index (BMI) (r = -0.46, P < 0.0001), fasting glucose and insulin concentrations, and blood pressure. There were correlations between the adult level of IGFBP-1 and birth weight (r = 0.11, P = 0.03) and ponderal index (weight/length3) at birth (r = 0.13, P = 0.01), but IGF-I was not related to birth measurements. There were interactive effects between childhood height or BMI and adult BMI on IGF-I and IGFBP-1 in adulthood. Tall height and high BMI at 7 yr were associated with low IGF-I (P = 0.03 for height and P = 0.003 for BMI) and high IGFBP-1 (P = 0.02 and P = 0.06) in adulthood but only in those subjects whose current BMI was below median. On further analysis these interactive effects were particularly strong for height in childhood and adult lean BMI (lean body mass/height2). Among men and women of below-average lean BMI, tall height at 7 yr was associated with low adult IGF-I (P = 0.007) and high IGFBP-1 (P = 0.0004) concentrations [interaction (7-yr height x adult lean BMI); P = 0.008 for IGF-I and 0.001 for IGFBP-1]. There is no evidence that reduced fetal growth programs IGF-I concentrations in old age. An association between small size at birth and low IGFBP-1 concentrations may in part reflect fetal programming effects on insulin resistance. Given the anabolic effects of the GH-IGF-I axis, subjects with tall height in childhood but low adult lean body mass may be at risk of late-life GH-IGF-I axis dysfunction. Prospective studies should address whether this group is susceptible to type 2 diabetes, coronary heart disease, and osteoporosis.
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The IGF system encompasses two IGFs, IGF-I and IGF-II, and at least six IGF-binding proteins (IGFBPs). A large body of evidence, ranging from studies in genetically engineered mice (5, 6) and human IGF-I gene mutations (7) to clinical studies in infants (8, 9, 10) and children (11), links the IGF system with both fetal and childhood growth. In adult life, IGF-I is an important anabolic hormone, required for the maintenance of bone integrity and lean body mass (12). Both pre- and postnatally IGF-I secretion is regulated by nutritional status, mainly through the stimulatory effect of insulin (13). After late infancy, the main regulator is pituitary GH (13). Circulating IGFBP-1 in nonpregnant adults is mostly inhibitory on IGF-I action (14). Although the regulation of IGFBP-1 is complex, insulin has a major inhibitory effect (15).
IGF-I and IGFBP-1 have been implicated in the pathogenesis of cardiovascular disease. Hypopituitarism is associated with increased cardiovascular mortality (16) and low IGF-I concentrations, possibly reflecting relative insufficiency of the GH-IGF-I axis and predates impaired glucose tolerance (17) and is associated with coronary heart disease (18, 19, 20). That the relationship with coronary heart disease may be causal is supported by a recent study in middle-aged men and women showing low IGF-I to predict ischemic heart disease over a follow-up of 16 yr (21). Low IGFBP-1 is associated with impaired glucose tolerance, elevated blood pressure, and obesity (19, 20, 22). However, these relationships are not entirely uniform. High IGF-I has been shown to predict coronary artery disease progression after myocardial infarction (23), and high IGFBP-1 is associated with increased cardiovascular and overall mortality in elderly men (24).
It has been proposed that programming of the IGF axis in utero or childhood could contribute to increased cardiovascular risk in adulthood. The evidence is limited to studies of children and young adults. Low birth weight has been associated with reduced serum IGFBP-1 in girls with precocious pubarche (25), elevated urinary IGF-I excretion in children with catch-up growth (26), and elevated serum IGF-I concentrations in prepubertal children (27, 28) and young adult women (29). No data have been reported for older adults. We set out to study how serum IGF-I and IGFBP-1 concentrations are related to cardiovascular risk factors and fetal and childhood growth in a well-characterized cohort of men and women aged 65–75 yr.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The original study cohort comprised 7086 men and women born as singletons during 1924–1933 at Helsinki University Central Hospital, who went to school in the city of Helsinki and were resident in Finland in 1971. They have detailed birth records (2, 3), which include birth weight, length, placental weight, head circumference, and gestational age at birth, and school health cards with an average of 10 (SD 4) measurements of height and weight between the ages of 7 and 15 yr. The study protocol was approved by the Ethics Committee of the National Public Health Institute, and written informed consent was obtained from all subjects.
Design
From the original study cohort, 421 subjects born at term (37 wk gestation or more) attended a clinic after an overnight fast between 0830 and 1000 h (4). Of these, 27 were on medication for type 2 diabetes and were excluded from the study because their medication could have altered their IGF-I and IGFBP-1 concentrations. To assess parameters of cardiovascular risk, we chose to perform a 75-g oral glucose tolerance test and measure body mass index (BMI), waist circumference, body fat content, and blood pressure as well as serum cholesterol [total, low-density lipoprotein (LDL) and high-density lipoprotein (HDL)], triglyceride, and fibrinogen concentrations. Anthropometric and blood pressure measurements were performed as described (4). Body fat mass was determined based on bioelectric impedance analysis (Omron BF 300 body fat monitor, Omron, Tokyo, Japan), and lean mass was calculated as total body weight - fat mass. To express the lean and fat masses in units adjusted for height, we calculated lean BMI as lean mass (kilograms)/[height (meters)]2 and fat BMI as fat mass (kilograms)/[height (meters)]2. The oral glucose tolerance test included glucose and insulin measurements from samples drawn at baseline and at 30 and 120 min. Measurements of plasma glucose concentration and serum insulin, proinsulin and 32–33 split proinsulin, total cholesterol, HDL cholesterol, LDL cholesterol, and triglyceride concentrations were performed by standard methods (4). Serum fibrinogen concentration was measured using an ACL autoanalyzer (Advanced Instrumental Laboratories, Milan, Italy). IGF-I concentration was measured by an IGF-I ELISA kit (DSL-10–5600; Diagnostic Systems Laboratories, Inc., Webster, TX). IGFBP-1 concentrations were measured as described (30).
Statistical analysis
IGF-I, IGFBP-1, glucose, insulin, proinsulin, HDL cholesterol, and triglyceride concentrations were log transformed to normality. Multiple linear regression and partial correlation analysis were used to assess associations between variables. Regression and correlation analyses were adjusted for sex, age, and adult BMI (unless lean or fat BMI was already included in the model). Height, weight, and BMI for each birthday between 7 and 15 yr were derived by first converting each measurement to a Z-score using the method of Royston (31). Successive Z-scores were interpolated with a piecewise linear function to obtain Z-scores separately for both sexes for each birthday. The Z-scores were back transformed to obtain corresponding height, weight, and BMI at these ages. A Z-score was not assigned if the child had not been measured within 2 yr of that age. Because findings were similar for all ages, we chose to present the 7-yr data to minimize confounding effects of individual variation in the timing of puberty.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The subjects’ birth and childhood data and adult clinical characteristics are summarized in Table 1
. Compared with present-day standards, the mean birth weight corresponds to -0.5 SD at 40 wk gestation in both sexes (32), mean height at 7 yr to -0.6 SD in both sexes (33) and BMI at 7 yr to -0.3 SD in boys and -0.1 SD in girls (34). Fourteen subjects (3.6%) were born small for gestational age, defined as birth weight less than -2 SD (32). These subjects had similar IGF-I and IGFBP-1 concentrations, compared with the rest of the study population. Mean IGF-I concentrations were higher in men, whereas IGFBP-1 concentrations were higher in women. There was a negative correlation between IGF-I and IGFBP-1 concentrations (r = -0.33; P < 0.0001). Twenty percent of men and 13% of women had undiagnosed or diet-treated diabetes mellitus (DM), and 25% of men and 30% of women had impaired glucose tolerance (IGT), defined by 1998 World Health Organization criteria (35). Compared with subjects with normal glucose metabolism, those with IGT or DM had similar IGF-I concentrations (geometric mean 102.9 µg/liter vs. normal subjects 98.1 µg/liter; P = 0.2), lower IGFBP-1 concentrations (55.8 µg/liter vs. 72.8 µg/liter; P = 0.004), higher lean BMI (17.5 kg/m2 vs. 16.7 kg/m2; P < 0.0001) and higher fat BMI (10.7 vs. 9.6 kg/m2; P < 0.0001). There were no significant differences between subjects with IGT and DM.
|
Table 2 shows correlations between IGF-I and IGFBP-1 concentrations and clinical measurements. IGF-I concentrations showed no correlation with current height or BMI but were negatively correlated with waist circumference (P = 0.03) and percentage body fat (P = 0.01). They were positively correlated with insulin and proinsulin variables, strongest with intact proinsulin and blood pressure and fibrinogen concentrations. IGFBP-1 concentrations showed strong negative correlations with adult total, lean, and fat BMI but no correlation with body fat percentage. Strong negative correlations were observed among IGFBP-1 concentration and glucose and insulin concentrations; total, intact, and 32–33 split proinsulin concentrations; and blood pressure. IGF-I and IGFBP-1 did not, however, correlate with serum total cholesterol, HDL or LDL cholesterol, or triglyceride concentrations.
|
IGF-I concentrations were not related to birth weight (Fig. 1) or any of the other birth measurements. IGFBP-1 concentrations showed a weak positive correlation with birth weight (r = 0.11; P = 0.03; Fig. 1) and ponderal index at birth (r = 0.13; P = 0.01), which became nonsignificant when not adjusted for current BMI. The strength of these associations was weakened after adjustment for fasting insulin concentration (birth weight: r = 0.08; P = 0.1; ponderal index at birth: r = 0.12; P = 0.02). IGF-I concentration was not correlated with height or BMI at 7 yr. There was a positive correlation between IGFBP-1 concentration and BMI at 7 (r = 0.15; P = 0.004) but no relationship with height. There were no interactive effects on IGF-I or IGFBP-1 between birth measurements and measurements at 7 yr or between birth and adult measurements.
|
Height and BMI at the age of 7 yr were positively correlated (r = 0.21; P < 0.0001). Height at 7 yr was correlated with adult height (r = 0.70; P < 0.0001) but not with adult BMI or lean or fat BMI. BMI at 7 yr was correlated with adult total (r = 0.30; P < 0.0001), lean (r = 0.34; P < 0.0001), and fat (r = 0.27; P < 0.0001) BMI.
There were interactive effects on IGF-I and IGFBP-1 between childhood BMI or height and adult BMI (Table 3). In subjects with an adult BMI below the median, high BMI or tall height at 7 yr was associated with low IGF-I and high IGFBP-1 concentrations. No associations were observed in subjects with above-median adult BMI. Because of the differing associations of IGF-I and IGFBP-1 with fat and lean body composition, we examined this further by dividing BMI into fat and lean components. The trends of IGF-I and IGFBP-1 with 7-yr height and BMI were accentuated in subjects of below-median lean BMI (Table 4A). No interactive effects were seen with adult fat BMI (Table 4B)6, percentage body fat, waist circumference, or height. There was also no any interactive effect between sex and any birth or childhood measurement.
|
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
IGF-I is a major endocrine and paracrine regulator of tissue growth and metabolism, with at least six binding proteins controlling its bioavailability. IGFBP-1 has both inhibitory and stimulatory effects on IGF-I, depending on the degree of phosphorylation. In nonpregnant adults, virtually all of the circulating IGFBP-1 is in the highly phosphorylated isoforms, which have a high affinity for, and are more inhibitory on, IGF-I (14). The strong associations we observed between low IGFBP-1 concentration and cardiovascular risk factors are in keeping with the known inhibition of IGFBP-1 synthesis by insulin (15) and clinical studies linking low IGFBP-1 with impaired glucose tolerance (19, 22), elevated blood pressure (22), and obesity (19, 22). The relationships between IGF-I concentration and cardiovascular risk factors were less straightforward to interpret, perhaps reflecting the dual regulation of IGF-I by insulin and GH. The association of high IGF-I with insulin resistance and blood pressure is consistent with the corresponding inverse associations with IGFBP-1 and is likely to reflect the differential regulation of IGF-I and IGFBP-1 synthesis by insulin. The strong association of IGF-I with fibrinogen may be related to the link between high circulating IGF-I and coronary artery disease progression (23). In contrast, high percentage body fat and waist circumference, well-established risk factors of atherosclerotic disease (36), were associated with low IGF-I concentrations. This could be a consequence of low GH activity, consistent with the fact that adult patients with frank GH deficiency have increased body fat (12), especially intra-abdominal fat (37) and higher mortality from cardiovascular disease (16). However, we cannot be sure because assessment of GH secretion is complex and highly impractical in population studies like ours. Nevertheless, these differing relationships of IGF-I with cardiovascular risk factors may in part explain why some studies associate cardiovascular disease with low (18, 19, 20, 21) and some with high (23) IGF-I.
The key role of the IGF system in regulating growth is illustrated by mice with knockout of the IGF-I gene (5) or IGFBP-1 gene overexpression (6) as well as human IGF-I gene deletion (7). All these examples are characterized by severe pre- and postnatal growth retardation. In addition, numerous studies show low circulating IGF-I and high IGFBP-1 concentrations in fetuses with intrauterine growth restriction (8, 9, 10) and in children with decreased rates of growth (11). Studies on monozygotic and dizygotic adult twin pairs have shown that the proportion of variance attributable to genetic effects is 38% for the circulating IGF-I concentration, whereas no significant heritability is found for IGFBP-1 (38). Recently preliminary evidence has suggested that a polymorphism in the promoter region of the IGF-I gene is associated with low birth weight, (39) low adult IGF-I concentration, and an increased risk of type 2 diabetes and myocardial infarction (40).
We found no relationship between IGF-I concentration and birth measurements and, thus, could not demonstrate any effect of fetal programming on circulating IGF-I in elderly people. This is at variance with studies in children who have linked low birth weight with elevated serum IGF-I (27, 28) or elevated urinary IGF-I excretion after catch-up growth (26). In young adults, some studies show a similar association (29), whereas others do not (41). Possible explanations include age-induced decline in the activity of the GH-IGF-I axis (42) or an existing association being obscured by the relatively high morbidity or selective survival of the study population. It is moreover of note that in a population study like ours, with only a small number of subjects born small for gestational age, possible programming effects of severe intrauterine growth restriction cannot be excluded.
The weak association we observed between low birth weight or thinness at birth and low IGFBP-1 concentration is consistent with previous observations in young adults (41) and girls with precocious pubarche (25). This was attenuated after adjustment for insulin concentrations, suggesting that the effect may in part be mediated through insulin, which shows inverse associations with birth size in this cohort (4). Other possible mechanisms include conditions associated with chronic intrauterine hypoxia such as preeclampsia, in which IGFBP-1 secreted by the fetal liver is a key mediator of reduced fetal growth (43).
We found that low IGF-I and high IGFBP-1 concentrations were strongly associated with high 7-yr height or BMI but only in subjects of below median adult lean BMI. This was an unexpected finding, and we can only speculate about the possible mechanisms involved. Low adult lean BMI may reflect failure to gain lean body mass during childhood and puberty or subsequent excessive loss of lean body mass. It may therefore identify individuals with relative GH/IGF-I deficiency, or at least a low GH/IGF-I effect, either now or in the past. One possible scenario is that the group with tall height or high BMI in childhood and low lean BMI and IGF-I in adulthood could be a result of an advanced tempo of childhood growth, early puberty (44), and subsequent early cessation of growth with interrupted lean body mass acquisition. In general, only about 5% of lean mass is acquired after 20 yr (45). However, this scenario remains speculative because we do not have sufficient measurements to determine the timing of puberty in this cohort. An alternative explanation for low IGF-I in this group includes subsequent undernutrition, which in animal experiments has been shown to cause hepatic GH resistance (46). Finally, discordance between tall height in childhood and low adult lean body mass may indicate loss of lean body mass in adult life. This could result from disease or lifestyle, again factors that the present study was not designed to assess.
Although the exact mechanisms behind our observations remain to be elucidated, their importance is due to their putative health consequences. Low serum IGF-I is associated with subsequent development of type 2 diabetes (17), possibly in part because of reduced survival of pancreatic ß-cells (47). This is of interest because previous observations in this cohort show that, although impaired glucose tolerance is associated with low childhood BMI (4), diabetes is predicted by tall height and high BMI in childhood (48). Correspondingly, coronary heart disease, which is similarly associated with low circulating IGF-I (18, 19, 20, 21), is in this cohort associated with high childhood BMI in men (2) and tall height in women (3) who were small at birth. Further features of a low-IGF state include reduced bone mineral density (49). This is again consistent with epidemiological observations in our cohort that show an increase in the incidence of hip fracture in subjects with tall height at 7 yr and reduced height increment thereafter (50).
We conclude that, in elderly people, small size at birth is weakly related to low serum IGFBP-1 concentration but not associated with changes in IGF-I. Both IGF-I and IGFBP-1 are related to childhood growth in subjects who have low lean body mass as adults. This finding suggests that subjects with low adult lean body mass and tall height at 7 yr may be at risk of late-life relative GH-IGF-I insufficiency. Whether these subjects are also more susceptible for type 2 diabetes, coronary heart disease, and osteoporosis remains the question that needs to be addressed in prospective studies.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Abbreviations: BMI, Body mass index; DM, diabetes mellitus; HDL, high-density lipoprotein; IGFBP, IGF-binding protein; IGT, impaired glucose tolerance; LDL, low-density lipoprotein.
Received August 30, 2002.
Accepted November 26, 2002.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  F. Schreiner, S. Stutte, P. Bartmann, B. Gohlke, and J. Woelfle Association of the Growth Hormone Receptor d3-Variant and Catch-up Growth of Preterm Infants with Birth Weight of Less Than 1500 Grams J. Clin. Endocrinol. Metab., November 1, 2007; 92(11): 4489 - 4493. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. S Schernhammer, S. S Tworoger, A H. Eliassen, S. A Missmer, J. M Holly, M. N Pollak, and S. E Hankinson Body shape throughout life and correlations with IGFs and GH Endocr. Relat. Cancer, September 1, 2007; 14(3): 721 - 732. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Troisi, N. Potischman, and R. N. Hoover Exploring the Underlying Hormonal Mechanisms of Prenatal Risk Factors for Breast Cancer: A Review and Commentary Cancer Epidemiol. Biomarkers Prev., September 1, 2007; 16(9): 1700 - 1712. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. M. Martin, J. M. P. Holly, G. Davey Smith, and D. Gunnell Associations of Adiposity from Childhood into Adulthood with Insulin Resistance and the Insulin-Like Growth Factor System: 65-Year Follow-Up of the Boyd Orr Cohort J. Clin. Endocrinol. Metab., September 1, 2006; 91(9): 3287 - 3295. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Sandhu, G. D. Smith, J. Holly, T. J. Cole, and Y. Ben-Shlomo Timing of Puberty Determines Serum Insulin-Like Growth Factor-I in Late Adulthood J. Clin. Endocrinol. Metab., August 1, 2006; 91(8): 3150 - 3157. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. Rogers, C. Metcalfe, D. Gunnell, P. Emmett, D. Dunger, J. Holly, and and the Avon Longitudinal Study of Parents and Chi Insulin-Like Growth Factor-I and Growth in Height, Leg Length, and Trunk Length between Ages 5 and 10 Years J. Clin. Endocrinol. Metab., July 1, 2006; 91(7): 2514 - 2519. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. Bray, D. Gunnell, J. M. P. Holly, N. Middleton, G. D. Smith, and R. M. Martin Associations of Childhood and Adulthood Height and the Components of Height with Insulin-Like Growth Factor Levels in Adulthood: A 65-Year Follow-Up of the Boyd Orr Cohort J. Clin. Endocrinol. Metab., April 1, 2006; 91(4): 1382 - 1389. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E E Frezza, M S Wachtel, and M Chiriva-Internati Influence of obesity on the risk of developing colon cancer Gut, February 1, 2006; 55(2): 285 - 291. [Full Text] [PDF]
|
|
|
|
 |
|
 A. Afghani, M. L. Cruz, and M. I. Goran Impaired Glucose Tolerance and Bone Mineral Content in Overweight Latino Children With a Family History of Type 2 Diabetes Diabetes Care, February 1, 2005; 28(2): 372 - 378. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A L Fowden and A J Forhead Endocrine mechanisms of intrauterine programming Reproduction, May 1, 2004; 127(5): 515 - 526. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. S. Sandhu, J. M. Gibson, A. H. Heald, D. B. Dunger, and N. J. Wareham Association between Insulin-Like Growth Factor-I: Insulin-Like Growth Factor-Binding Protein-1 Ratio and Metabolic and Anthropometric Factors in Men and Women Cancer Epidemiol. Biomarkers Prev., January 1, 2004; 13(1): 166 - 170. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. A. Laughlin, E. Barrett-Connor, M. H. Criqui, and D. Kritz-Silverstein The Prospective Association of Serum Insulin-Like Growth Factor I (IGF-I) and IGF-Binding Protein-1 Levels with All Cause and Cardiovascular Disease Mortality in Older Adults: The Rancho Bernardo Study J. Clin. Endocrinol. Metab., January 1, 2004; 89(1): 114 - 120. [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 |
