0021-972X/01/$03.00/0 The Journal of Clinical Endocrinology & Metabolism Copyright © 2001 by The Endocrine Society
Vol. 86, No. 2 Printed in U.S.A.
COMMENT Cord Blood Leptin and Insulin-Like Growth Factor Levels are Independent Predictors of Fetal Growth HELEN CHRISTOU, JEAN M. CONNORS, MARY ZIOTOPOULOU, VASSILIA HATZIDAKIS, ELIZABETH PAPATHANASSOGLOU, STEVEN A. RINGER, AND CHRISTOS S. MANTZOROS Division of Newborn Medicine, Children’s Hospital (H.C.) and Brigham and Women’s Hospital (H.C., S.A.R.), Hematology Division (J.M.C.), Department of Medicine, Brigham and Women’s Hospital and Division of Endocrinology (M.Z., V.H., E.P., C.S.M.), Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts 02215 ABSTRACT The insulin-like growth factor (IGF) system is the dominant endocrine regulator of fetal growth, whereas insulin has a permissive role. Although a role for leptin in fetal growth has been suggested recently, the mechanism by which leptin may be related to fetal growth is not known; but leptin may interact with the IGF system in utero as it does in the extrauterine life. In the context of a hospital-based case control study, we collected anthropometric and demographic data and measured serum leptin, IGF-I, IGF-II, insulin, cortisol, and IGF binding protein 3 concentrations in 142 cord blood samples from full-term deliveries. Cord leptin, IGF-I, and insulin levels correlated positively with birth weight (r ⫽ 0.46, r ⫽ 0.41, and r ⫽ 0.21, respectively, P ⬍ 0.01)
by univariate analysis and were significantly higher in large-forgestational-age (LGA) infants, compared with appropriate-for-gestational-age (AGA) infants. Cord leptin concentrations correlated with insulin levels (r ⫽ 0.36, P⬍0.01) but not with IGF-I levels (r ⫽ 0.20). Multiple linear and logistic regression analysis demonstrated an independent positive relationship of both leptin and IGF-I with birth weight and AGA/LGA status. The positive association of leptin levels with birth weight and AGA/LGA status cannot be attributed to IGF-I. This suggests the existence of alternative mechanisms underlying leptin’s associations with fetal growth that should be further explored. (J Clin Endocrinol Metab 86: 935–938, 2001)
R
EGULATION OF FETAL growth is complex and multifactorial. Diverse factors, including intrinsic fetal conditions as well as maternal and environmental factors, lead to aberrant intrauterine growth. The mechanisms controlling intrauterine growth are incompletely understood, but an interaction of maternal, placental, and fetal endocrine factors is likely to govern partitioning of nutrients and rate of fetal cell proliferation and maturation. The main established endocrine regulators of fetal growth include insulin and the insulin-like growth factor (IGF) system (1). Whereas IGF-I cord levels correlate well with birth weight (2, 3), insulin’s role is thought to be permissive in fetal growth (1), and IGF-II levels correlate poorly with birth weight (4). In addition, studies in null mutant mice have demonstrated that, whereas IGF-I is the dominant growthpromoting factor during the rapid phase of somatic growth in late gestation (5, 6), IGF-II is important in fetal growth during early gestation only (7). Leptin, a 16-kDa hormone secreted by adipocytes and the placenta, has recently been positively associated with intrauterine growth (8 –11). This hormone is detectable in cord
blood; however, the mechanism by which it may be related to fetal growth is not known. Insulin may regulate (increase) leptin levels in humans, and leptin is directly associated with GH and with circulating concentrations of IGF-I (12–14) in extrauterine life. It is therefore reasonable to speculate that either insulin or IGF-I may mediate leptin’s effect on fetal growth. Because the role of leptin and the IGF system in determining fetal growth has not yet been evaluated jointly, we sought to determine whether leptin’s role in fetal growth is mediated by the IGF system or whether leptin plays an independent role in determining fetal growth. Subjects and Methods This hospital-based 2:1 case control study was approved by the Institutional Review Board of Brigham and Women’s Hospital and was conducted between July and September, 1998. Cord blood samples were collected from 140 consecutively enrolled vaginal and cesarean full-term deliveries (142 subjects) and were stored at 4 C for a maximum of 12 h. They were then centrifuged, and serum was aliquoted and stored at ⫺70 C until assayed. Demographic and anthropometric data were collected from the maternal and infant medical records. Infants were categorized as appropriate for gestational age (AGA, birth weight between 10th and 90th percentiles) and large for gestational age (LGA, birth weight more than 90th percentile). Infants who were small for gestational age were not included in our study. Leptin was assayed by RIA using a commercially available kit from Linco Research, Inc., St. Louis, MO. All other hormone determinations were done by radioimmunometric assays using commercially available kits from DSL, Webster, TX. The sensitivity of the assays was 0.5 ng/mL
Received May 24, 2000. Revision received October 24, 2000. Accepted October 27, 2000. Address all correspondence and requests for reprints to: Christos S. Mantzoros, M.D., Division of Endocrinology, Beth Israel Deaconess Medical Center, 99 Brookline Avenue Boston, Massachusetts 02215. Email:
[email protected].
935
936
for leptin, 2.06 ng/mL for IGF-I, 12 ng/mL for IGF-II, 0.5 ng/mL for IGF binding protein (IGFBP)3, 1.3 IU/mL for insulin, and 0.3 g/dL for cortisol. All the samples were analyzed in one batch. The intraassay coefficients of variation were 5– 8% for leptin, 3.9 –7% for IGF-I, 3.4 – 6.5% for IGF-II, 1.8 –3.9% for IGFBP3, 4.5– 8.3% for insulin, and 5.3–11% for cortisol. Statistical methods used were the t test for comparison of anthropometric data and hormone levels between the two subject groups, Spearman’s test for univariate correlations, and multivariate linear and logistic regression analysis. All data analysis was done using the SPSS, Inc. (Chicago, IL) 8.0 software for Windows.
Results
Full-term infants were included in the study (mean gestational age, 39.5 ⫾ 0.08 weeks). Mean growth parameters were on the 50th percentile, as expected. Birth weight was 3482 ⫾ 39.9 g, length was 50.3 ⫾ 0.22 cm, head circumference was 34.6 ⫾ 0.12 cm, and body mass index (BMI) was 27.4 ⫾ 0.28 kg/m2. Forty-five percent of study subjects were male; and, because of our study design, 33% of the study subjects were LGA. Mean maternal age was 30.7 ⫾ 0.49 yr, and cesarean delivery occurred in 17.6% of study subjects. Conditions known to influence intrauterine growth, such as smoking, diabetes, and pregnancy-induced hypertension (PIH), were present in a small number of cases. Specifically, there were 4 cases of maternal diabetes (all in the LGA group), 6 cases of PIH (2 in the AGA and 4 in the LGA group), and 3 mothers were smokers (2 in the AGA and 1 in the LGA group). The distribution of vaginal deliveries and cesarean sections was not significantly different between the LGA group (11 infants) and the AGA group (15 infants). The majority of mothers received lactated Ringer’s solution during labor; however, a small number of mothers did receive glucose infusions during labor (3 mothers of LGA infants and 2 mothers of AGA infants, not significantly different). The mean 1- and 5-min APGAR scores were not statistically different between the 2 groups. Comparison of anthropometric data and cord hormone levels among AGA and LGA infants is shown (see Table 1). As expected, birth weight, length, and BMI were significantly higher in LGA, compared with AGA, infants. Mean leptin concentrations were significantly higher in the LGA group of infants, compared with the AGA group (13.97 ⫾ 1.08 vs. 8.11 ⫾ 0.62 ng/mL, respectively, P ⬍ 0.001). Significantly higher mean IGF-I, IGF-II, and insulin levels were also found in the LGA group of infants, compared with the AGA group, TABLE 1. Mean values and standard error of the mean (SEM) of study anthropometric indices and hormones in AGA and LGA infants
BMI (kg/m2) Length (cm) Weight (g) Leptin (ng/mL) IGF-I (ng/mL) IGF-II (ng/mL) IGFBP3 (ng/mL) Insulin, IU/mL Cortisol, g/dL a b
JCE & M • 2001 Vol. 86 • No. 2
CHRISTOU ET AL.
P ⬍ 0.05. P ⬍ 0.001.
AGA (n ⫽ 96) Mean ⫾ SEM
LGA (n ⫽ 46) Mean ⫾ SEM
26.80 ⫾ 0.37 49.51 ⫾ 0.27 3225 ⫾ 31.30 8.11 ⫾ 0.62 43.39 ⫾ 3.38 316.47 ⫾ 8.57 1749 ⫾ 51.37 3.09 ⫾ 0.16 14.76 ⫾ 0.80
28.54 ⫾ 0.35a 52.02 ⫾ 0.25b 4007 ⫾ 41.70b 13.97 ⫾ 1.08b 73.81 ⫾ 5.87b 372.81 ⫾ 22.09a 1868.13 ⫾ 150.21 3.88 ⫾ 0.26a 14.60 ⫾ 1.11
whereas IGFBP3 and cortisol levels were not significantly different between the two groups (Table 1). We then examined univariate correlations among the study hormones in AGA infants (Table 2) and found statistically significant correlations between leptin and insulin levels (r ⫽ 0.36, P ⬍ 0.01), IGF-I and insulin (r ⫽ 0.35, P ⬍ 0.01), and IGF-I and IGF-II (r ⫽ 0.48, P ⬍ 0.01). There was also a statistically significant negative correlation between cord blood cortisol and IGF-I levels (r ⫽ ⫺0.24, P ⬍ 0.05). There was no statistically significant correlation between leptin and IGF-I or IGF-II levels. Given that even nonsignificant associations can introduce confounding results and that aggregate confounding from several factors can be substantial, we then performed multivariate regression analysis adjusting for potential confounding variables. Table 3A presents a summary of the univariate and multivariate regression analysis for birth weight in relation to the study hormones, adjusted for successively introduced covariates, as detailed in Table 3A. Log leptin levels positively correlated with birth weight, even after adjusting for gender, birth weight category (AGA or LGA), maternal smoking, diabetes, and PIH and all other study hormones, including IGF-I (r3 ⫽ 0.18, P ⫽ 0.012). IGF-I also correlated strongly with birth weight after adjustment for the same confounding variables (r3 ⫽ 0.21, P ⫽ 0.001). IGF-II, IGFBP3, and insulin’s correlation with birth weight, on the other hand, did not persist after adjustment for the other covariates. Table 3B shows multiple logistic regressionderived odds ratios (ORs) for leptin and IGF-1 by birth weight category adjusted for successively introduced covariates, as detailed in Table 3B. After adjustment for gender, IGF-I, IGF-II, IGFBP3, insulin, cortisol, maternal diabetes, smoking, and hypertension, the OR for leptin is 25.1 (P ⫽ 0.01); whereas, after adjustment for the same variables and leptin, the OR for IGF-I is 2.87 (P ⫽ 0.03). Taken together, these data indicate that these two hormones are the only independent predictors of birth weight and AGA/LGA status. Discussion
The role of leptin among the complex network of factors controlling fetal growth is incompletely understood. Leptin is positively associated with birth weight, but the mechanism(s) underlying this association remains unknown. This association may reflect either a simple relationship with adipose tissue mass or an active role for leptin in fetal growth. Several hormonal axes could influence or mediate leptin’s ability to modulate intrauterine growth. Our study demonstrates that the positive correlation between cord leptin levels TABLE 2. Spearman correlation coefficients between leptin and the other measured hormones among AGA infants
IGF-I IGF-II IGFBP3 Insulin Cortisol a b
Leptin
IGF-I
IGF-II
IGFBP3
Insulin
0.20 0.07 0.12 0.36a 0.12
0.48a 0.09 0.35a ⫺0.24b
0.16 0.15 ⫺0.03
0.02 ⫺0.07
0.13
Correlation is significant at the 0.05 level (two-tailed). Correlation is significant at the 0.01 level (two-tailed).
COMMENTS TABLE 3. Birth weight in relation to study variables: A. Univariate and multivariate regression analysis for birth weight in relation to study variables (leptin and IGF-I values were logarithmically transformed to normalize the respective distributions) B. Multiple logistic regression-derived adjusted ORs for AGA vs. LGA infants (leptin and IGF-I are expressed in standard deviates) A
Leptin IGF-I IGF-II IGFBP3 Insulin B
Leptin IGF-I
r1
P
r2
P
r3
P
0.46 0.41 0.21 0.06 0.21
⬍0.001 ⬍0.001 0.014 NS 0.01
0.18 0.19 0.04 0.00 0.03
⬍0.001 0.001 NS NS NS
0.18 0.21 ⫺0.025 ⫺0.032 ⫺0.030
0.012 0.001 NS NS NS
OR1
P
OR2
P
35.58 3.02
⬍0.001 ⬍0.001
25.1 2.87
0.01 0.003
r1, Univariate standardized regression coefficient; r2, multivariate standardized regression coefficient adjusted for gender and birth weight category (AGA or LGA); r3, multivariate standardized regression coefficient adjusted for gender, birth weight category, maternal smoking, diabetes and pregnancy-induced hypertension and mutually for the variables shown in the table; OR1, adjusted for gender; OR2, adjusted for gender, IGF-II, IGFBP3, insulin, cortisol, maternal diabetes, smoking and pregnancy-induced hypertension and mutually for leptin and IGF-I; NS, not significant.
and birth weight is independent of both insulin and the IGF system, the two endocrine systems currently believed to play a major role in fetal growth. The IGF system mediates the effect of GH in postnatal growth and plays a major role in fetal growth (1). Recent experimental data indicate that low leptin levels are associated with decreased GH in rodents (12); and leptin administration, which normalizes circulating leptin levels, also normalizes GH secretion (14). Leptin is also directly associated with circulating concentrations of IGF-I and IGF-I/IGFBP3 ratios (13). It is thus reasonable to speculate that leptin’s association with fetal growth may be mediated by the IGF system, and specifically by IGF-I. In agreement with previous studies (3, 8 –11), we found positive correlations between birth weight and cord blood levels of insulin, IGF-I, and leptin. In our study, IGF-II cord blood levels did not correlate with birth weight. We also found significantly higher levels of leptin, IGF-I, IGF-II, and insulin in cord blood samples of LGA, compared with AGA, infants. It has been suggested that IGFBP3, the principal carrier for IGFs in adult serum is regulated in a coordinated manner with IGF-I (4). Contrary to a previous report (2), IGFBP3 cord blood levels were not significantly increased in LGA, compared with AGA, infants in our study. Similarly, cord blood cortisol levels were not different between the two groups of infants. We found a statistically significant correlation between insulin and leptin levels in AGA infants, which is in agreement with the previously reported positive relationship between leptin and insulin (14). Although the nature of the relationship of insulin and leptin remains uncertain, some reports have suggested that insulin has a stimulatory effect on leptin secretion in vitro and in vivo (14). In contrast, we found no statistical correlation between cord leptin levels and IGF-I or IGF-II levels. This is consistent with a previous report in which no correlation was found between cord leptin
937
and IGF-I levels in newborn infants (15). These findings suggested the possibility of an independent association between leptin and birth weight. Using multivariate logistic regression analysis, we sought to determine whether the association between leptin levels and fetal growth would persist after adjustment for potential confounders. We found a statistically significant relationship between cord leptin levels and birth weight that is independent of other variables, including insulin and components of the IGF system. The correlation of cord leptin levels with birth weight is consistent with earlier reports (8 –11). The basis of this correlation may be related to the direct relationship between adipose tissue mass and circulating leptin concentrations (16), but the recognition of the placenta as a source of leptin production (17) may suggest an active role for leptin in fetal growth as well. In addition, we previously reported that cord blood leptin levels are independently associated with intrauterine growth after adjusting for adiposity, placental weight, insulin, and maternal leptin levels (18), again indicating a direct role for leptin in fetal growth. Moreover, cord leptin levels are decreased by conditions that lead to low birth weight, such as maternal smoking and prematurity (19). Whether leptin is directly implicated in determining fetal growth, or it simply represents a marker for fetal nutrition and growth, remains to be shown by future interventional studies. It is interesting to note that neither birth weight nor postnatal linear growth were impaired in the two reported cases of human leptin deficiency (20). As the role of leptin in energy homeostasis is further understood, new insights are likely to emerge regarding its role in human fetal growth. Recent data indicate that there is significant diurnal and seasonal variation (14) in leptin levels, with highest levels occurring in the spring and summer. In addition, although relatively little is known about factors that regulate leptin production in fetal life, it is known that insulin and glucocorticoids are hormones that regulate leptin production in adults (14). We have controlled for these potential confounders by performing the study over a short time period and by adjusting for these hormones in the analysis. Although differential misclassification is unlikely, given the blinded laboratory analysis, nondifferential misclassification caused by random laboratory error or the diurnal rhythms of leptin levels is possible. Such error would be expected to bias results toward the null and alter the P values toward nonsignificant values. In addition, although confounding was controlled through standard statistical procedures, residual confounding by other serum hormones or unmeasured factors remains a possibility that should be tested by future studies. We recognize that the design of this study may have led to an underestimation of the true association between hormone levels and fetal growth, to the extent that hormone levels measured at birth do not accurately reflect earlier levels that would be etiologically more relevant. Confirmation of our results by additional studies may be warranted. In conclusion, our data indicate that birth weight is independently associated with leptin concentrations in cord blood of newborns. This suggests that leptin may be involved in as-yet-unknown mechanisms that regulate fetal growth. Additional studies will likely lead to an understanding of the mechanism(s) by which leptin may regulate fetal growth.
938
CHRISTOU ET AL. References
1. Milner RDG, Gluckman PD. 1996 Regulation of intrauterine growth. In: Gluckman PD, Heymann MA, eds. Pediatrics and perinatology. The scientific basis, ed 2. Arnold/Oxford University Press. 2. Giudice LC, DeZegher F, Gargosky SE, et al. 1995 Insulin-like growth factors and their binding proteins in the term and preterm human fetus and neonate with normal and extremes of intrauterine growth. J Clin Endocrinol Metab. 80:1548 –1555. 3. Verhaeghe J, Van Bree R, Van Herck E, Laureys J, Bouillon R, Van Assche FA. 1993 C-peptide, insulin-like growth factors I and II and insulin-like growth factor binding protein-1 in umbilical cord serum: correlations with birth weight. Am J Obstet Gynecol. 168:89 –97. 4. Wiznitzer A, Reece EA, Homko C, Furman B, Mazor M, Levy J. 1998 Insulinlike growth factors, their binding proteins and fetal macrosomia in nondiabetic pregnant women. Am J Perinatol. 15:23–28. 5. Baker J, Liu JP, Robertson EJ, Eftstratiadis A. 1993 Role of insulin-like growth factors in embryonic and postnatal growth. Cell. 75:73– 82. 6. Liu JP, Baker J, Perkins AS, Robertson EJ, Eftstratiadis A. 1993 Mice carrying null mutations of the genes encoding insulin-like growth factor I (IGF-I) and type 1 IGF receptor (IGF1R). Cell. 75:59 –72. 7. DeChiara TM, Eftstratiadis A, Robertson EJ. 1990 A growth-deficiency phenotype in heterozygous mice carrying an insulin-like growth factor II gene disrupted by targeting. Nature. 345:78 – 80. 8. Cinaz P, Sen E, Bodeci A, Ezgu FS, Atalay Y, Koca E. 1999 Plasma leptin levels of large for gestational age and small for gestational age infants. Acta Paediatr. 88:753–755. 9. Shekhawat PS, Garland JS, Shivpuri C, et al. 1998 Neonatal cord blood leptin: its relationship to birth weight, body mass index, maternal diabetes and steroids. Pediatr Res. 43:338 –343. 10. Koistinen HA, Koivisto VA, Anderson S, et al. 1997 Leptin concentration in
11.
12. 13.
14. 15.
16. 17.
18. 19.
20.
JCE & M • 2001 Vol. 86 • No. 2
cord blood correlates with intrauterine growth. J Clin Endocrinol Metab. 82:3328 –3330. Harigaya A, Nagashima K, Nako Y, Morikawa A. 1997 Relationship between concentration of serum leptin and fetal growth. J Clin Endocrinol Metab. 82:3281–3284. Carro E, Senaris R, Considine RV, Casanueva FF, Dieguez C. 1997 Regulation of in vivo growth hormone secretion by leptin. Endocrinology. 138:2203–2206. Paolisso G, Ammendola S, Del Buono A, et al. 1997 Serum levels of IGF I and IGFBP3 in healthy centenarians: relationship with plasma leptin and lipid concentrations, insulin action and cognitive function. J Clin Endocrinol Metab. 82:2204 –2209. Mantzoros CS. 1999 The role of leptin in human obesity and disease; a review of current evidence. Ann Intern Med. 130:671– 680. Lepercq J, Lahlou N, Timsit J, Girard J, Hauguel-de Mouzon S. 1999 Macrosomia revisited: ponderal index and leptin delineate subtypes of fetal overgrowth. Am J Obstet Gynecol. 181:621– 625. Clapp JF, Kiess W. 1998 Cord blood leptin reflects fetal fat mass. J Soc Gynecol Invest. 5:300 –303. Masuzaki H, Ogawa Y, Sagawa N, et al. 1997 Nonadipose tissue production of leptin: leptin as a novel placenta-derived hormone in humans. Nat Med. 3:1029 –1033. Varvarigou A, Mantzoros CS, Beratis NG. 1999 Cord blood leptin concentrations in relation to intrauterine growth. Clin Endocrinol (Oxf). 50:177–183. Mantzoros CS, Varvarigou A, Kaklamani VG, Beratis NG, Flier JS. 1997 Effect of birth weight and maternal smoking on cord blood leptin concentrations of full-term and preterm newborns. J Clin Endocrinol Metab. 82:2856 –2861. Montague CT, Farooki IS, Whitehead JP, et al. 1997 Congenital leptin deficiency is associated with severe early-onset obesity in humans. Nature. 387: 903–908.
Erratum Corrections have been made in the article “Fracture risk reduction with alendronate in women with osteoporosis: The Fracture Intervention Trial” by Dennis M. Black, Desmond E. Thompson, Douglas C. Bauer, Kris Ensrud, Thomas Musliner, Marc C. Hochberg, Michael C. Nevitt, Shailaja Suryawanshi, and Steven R. Cummings, for the FIT Research Group (The Journal of Clinical Endocrinology & Metabolism 85:4118 – 4124). On page 4120, in the first paragraph of Results, the second sentence “The women (53.5%) with vertebral fracture had a femoral neck T score of ⫺2.5 or less” should read “The women (53.5%) with vertebral fracture had a femoral neck T score of ⫺1.6 or less.” Also in Results, in the fourth paragraph, the third sentence “The reduction in risk was first significant for clinical vertebral fracture (59%) by month 12 (P ⬍ 0.001)” should read “The reduction in risk was first significant for clinical vertebral fracture (59%) by month 12 (P ⫽ 0.03).” In the legend to Figure 1, the following sentence has been added: “A, Clinical vertebral fracture; B, any clinical fracture; C, nonvertebral fracture; D, hip fracture; E, wrist fracture.” The authors regret the errors.
