Preterm Birth: Associated Neuroendocrine, Medical, and Behavioral ...

4 downloads 0 Views 193KB Size Report
and behavioral factors associated with preterm delivery were, re- spectively, previous .... term; 84 of these had idiopathic preterm birth without complications and.
0021-972X/01/$03.00/0 The Journal of Clinical Endocrinology & Metabolism Copyright © 2001 by The Endocrine Society

Vol. 86, No. 6 Printed in U.S.A.

Preterm Birth: Associated Neuroendocrine, Medical, and Behavioral Risk Factors* KRISTINE ERICKSON, POUL THORSEN, GEORGE CHROUSOS, DIMITRI E. GRIGORIADIS, ON KHONGSALY, JAMES MCGREGOR, JAY SCHULKIN

AND

Department of Psychology, American University (K.E.), Washington, D.C. 20016; Division of Child Development, Disability, and Health, Centers for Disease Control and Prevention (P.T.), Atlanta, Georgia 30333; Danish Epidemiology Sciences Centre, Aarhus University, Aarhus, Denmark; Pediatric and Reproductive Endocrinology Branch, National Institute of Child Health and Human Development, National Institutes of Health (G.C.), Bethesda, Maryland 20892; Neurocrine Biosciences, Inc. (D.E.G., O.K.), San Diego, California 92121; Department of Obstetrics and Gynecology, University of Colorado (J.M.), Denver, Colorado 80217; and Department of Physiology and Biophysics, Georgetown University (J.S.), Washington, D.C. 20007 ABSTRACT Increased CRH secretion by the placenta of pregnant women has been associated with preterm birth. Certain indices of risk, both medical and psychosocial in nature, have been linked to preterm delivery. Levels of total, bound, and free CRH, CRH-binding protein (CRH-BP), and cortisol were measured prospectively in a large sample of pregnant Danish women who delivered preterm and term infants. Measures of maternal serum hormones were taken at 7–23 and 27–37 weeks gestation and, for those who delivered at term, at 37– 43 weeks gestation. At 7–23 weeks gestation, maternal levels of total CRH (P ⫽ 0.01), bound CRH (P ⫽ 0.03), and CRH-BP (P ⫽ 0.01) were higher in the preterm than in the term group. At 27–37 weeks ges-

P

RETERM DELIVERY (PTD), i.e. initiation of delivery before 259 days gestation, represents a significant clinical challenge associated with high perinatal morbidity and mortality rates (1–3). The reported incidence of PTD in Western societies ranges from 5–10% of all pregnancies (4 – 6). Maturation of the fetus is promoted by proper, progressive activation of the fetal hypothalamic-pituitary-adrenal axis in normal pregnancy (7–9). CRH is synthesized and secreted by Received November 17, 2000. Revision received January 30, 2001. Rerevision received March 1, 2001. Accepted March 21, 2001. Address all correspondence and requests for reprints to: Jay Schulkin, Ph.D., Department of Physiology and Biophysics, Basic Science Building, Georgetown University, Washington, D.C. 20007. E-mail: [email protected]. * This work was supported by the Danish Health Insurance Foundation; Institute of Clinical Research; University of Odense; Foundation of Reproductive Biology; Odense University Hospital; National Fund for Research and Development; Danish Ministry of Health; Medical Scientifical Foundation for Copenhagen, Faroe Islands, and Greenland; Merchant Sven Hansen and Mrs. Ina Hansens Foundation. Director Jacob Madsen and Mrs. Olga Madsens Foundation; Nordic Union of Obstetricians and Gynaecologists Foundation; and Odense University Hospital. The activities of The Danish Epidemiology Science Center are financed by a grant from the Danish National Research Foundation. The study was also supported by a grant from March of Dimes Birth Defects Foundation (20-FY98 –700), New York, and supported under a cooperative agreement from the Centers for Disease Control and Prevention (Atlanta, GA) through the Association of Teachers of Preventive Medicine, Washington, D.C.

tation, levels of total CRH (P ⬍ 0.0001), bound CRH (P ⬍ 0.0001), free CRH (P ⬍ 0.0001), and cortisol (P ⬍ 0.0001) were all higher in the preterm than the term group, whereas levels of CRH-BP (P ⬍ 0.0001) were lower in the preterm than in the term group. The best medical and behavioral factors associated with preterm delivery were, respectively, previous preterm delivery (P ⬍ 0.0001) and engagement in certain risk-taking behaviors (P ⫽ 0.008). The positive relations between preterm delivery and various adverse medical and socioeconomic variables with increases in placental secretion of CRH suggest that the latter may participate in the pathophysiology of preterm delivery. (J Clin Endocrinol Metab 86: 2544 –2552, 2001)

the placenta, and plasma CRH levels progressively increase in both maternal and fetal circulations (8 –12). During the second and third trimesters of a normal pregnancy, CRH is easily detected in maternal blood (13, 14), and during the third trimester, the increasing levels of CRH have been suggested to have a key role in determining the timing of term labor (15). CRH-binding protein (CRH-BP) may act as a regulator of circulating CRH by limiting the concentrations of free, biologically active CRH (16, 17). CRH-BP circulates in the blood of both nonpregnant and pregnant women, and in the latter it may protect against precipitous rises in free CRH levels and, possibly, the occurrence of labor and delivery before term (17–20). By the 34th week of gestation and as the time of parturition approaches, the levels of total CRH continue to increase, whereas the concentrations of CRH-BP decrease, allowing for a rise in free CRH concentrations and, hence, further stimulation of labor processes (17, 18). Total plasma CRH concentrations are elevated during the second trimester in pregnant women who experience PTD (21). Similarly, bacterial infectious diseases (5, 22), preeclampsia (23), growth-restricted fetal development (24), and psychosocial stress (25, 26) all have been associated with elevated levels of total plasma CRH and preterm labor and delivery. Increased levels of CRH in women who deliver preterm may, thus, be an indicator of potential maternal-fetal

2544

NEUROENDOCRINE, MEDICAL, AND BEHAVIORAL RISK FACTORS OF PRETERM BIRTH

2545

distress or increased metabolic and physiological demands (14, 18, 27, 28). The present study reports levels of total, bound, and free CRH, CRH-BP, and cortisol in women who prospectively went on to have preterm or term delivery. Levels of free CRH and CRH-BP were hypothesized to be better predictors for PTD than total CRH levels, and higher levels of free CRH and lower levels of CRH-BP were expected to be associated with preterm, rather than term, delivery. Various medical and psychosocial variables previously linked to preterm delivery were also evaluated to identify whether medical, socioeconomic, or lifestyle conditions may predict high levels of CRH and PTD.

the institutional review board for research on human subjects, Centers for Disease Control and Prevention (Atlanta, GA). Data collection methods were approved by the Danish Data Surveillance Authority. All participants received a pelvic examination at enrollment. Samples taken from the cervical os and the vaginal vault were tested for bacterial vaginosis and sexually transmitted diseases. Samples of venous blood were obtained for determination of total, bound, and free CRH, CRH-BP, and cortisol levels. All participating women enrolled in the study by 23 weeks and 3 days gestational age (mean gestational age, 16 weeks and 4 days). Results of ultrasonographic measurements of the biparietal diameter and the femur length of the fetus at the 18th week gestation confirmed gestational age and the estimated date of delivery for 97.5% of the participants. Serum samples were taken in early pregnancy (before 24 gestational weeks; mean gestational age, 16 weeks 4 days) and between 27–37 gestational weeks.

Materials and Methods Study population

Study groups

All women receiving prenatal care at the Department of Obstetrics and Gynecology at Odense University Hospital, Denmark, between November 1992 and February 1994 (n ⫽ 3596) were invited to participate in the study (Fig. 1). Women who enrolled (n ⫽ 3174; 88.3%) did so at their first antenatal hospital visit, before 24 full gestational weeks, and 2927 (81.4%) completed all test and questionnaire requirements. The inclusion criteria were age above 18 yr and ability to understand Danish. The criteria for exclusion were insufficient responses to the questionnaires, placental previa (diagnosed after 30 full gestational weeks), history of severe fetal congenital malformations in previous pregnancy, and uterine cervix insufficiency treated with cervical circlage. All participants gave written informed consent that met the requirements of the scientific ethics committee for Vejle and Funen counties, Denmark, and

FIG. 1. Study flow chart and group inclusion in the preterm and term groups.

The designation of women into the various preterm and term groups was as follows (see Fig. 1). Of the cohort of 2927 women who agreed to participate in and followed through with the study, 170 delivered preterm; 84 of these had idiopathic preterm birth without complications and fulfilled the test and questionnaire requirements. During labor, pregnant women having idiopathic preterm births were asked to provide the same venous blood samples as they had at enrollment.

Seven to 23 weeks gestational age groups The 84 women described above who gave birth preterm were included in this analysis. Of the initial women from the sample invited to participate who gave birth at term, 224 were selected as gestational controls. They were matched at enrollment to within 10 days of the

2546

ERICKSON ET AL.

gestational duration of the preterm cases and agreed to participate and to have specimens taken at 7–23 weeks gestation.

Twenty-seven to 37 weeks gestational age groups As glucocorticoids influence the levels of CRH in pregnant women (29), we excluded betamethasone-treated subjects from the 27–37 gestational week analyses. The resulting preterm sample size for our analyses was 59. We matched 166 controls to the preterm group by gestational age.

Delivery groups The same 59 women with preterm delivery from the case-control study who did not receive betamethasone treatment were enrolled as the preterm delivery cohort. Three hundred women who delivered at term (38 – 43 weeks gestation) served as the delivery comparison group. One hundred and twenty-five of these women were from the earlier casecontrol study; they agreed to be members of the delivery control group when approached at the time of delivery and had specimens taken at that time. The other 175 participants with normal term deliveries, who had delivery specimens taken, were selected by computerized randomization of the remaining study participants. All questionnaire and laboratory data were available for 296 of the delivery control group members.

Procedures Maternal venous blood samples were obtained at each of the examinations (first sample before 24 weeks gestation, second at 27–37 gestational weeks, third at 38 – 43 gestational weeks for those who delivered at term). All samples were obtained in dry sterile tubes (7 mL), cooled at room temperature, centrifuged, and aliquoted within 2 h from their collection. Serum was stored at ⫺80 C until assayed. Vaginal/cervical samples were obtained after the vault of the vagina had been exposed to a sterile nonlubricated vaginal speculum. Sampling techniques for individual microorganisms have previously been described in detail (30). Assessments of total, bound, and free CRH, CRH-BP, and cortisol levels were performed using previously reported procedures (31). Briefly, serum samples from the subjects described above were received frozen (⬃1–2 mL) and thawed on the day of assay. Each sample was then aliquoted into smaller volumes and refrozen at ⫺80 C for any future use. For CRH-BP 16-well nicrotiter plates (wells A1 to H12) were coated overnight at 4 C with anti-CRH-BP monoclonal antibody (100 ␮L/well; 5 ␮g/mL in 0.05 m NaHCO3 buffer, pH 9.5). The plate was blocked using 100 ␮L casein blocking buffer solution (1 ⫻ TBS with 1% casein, pH 7.5) for 1 h. The plate was again washed three times with the wash buffer. Fifty microliters of sample and standard dilutions were added to wells and allowed to incubate with the CRH-BP antibody at 22 C for 1 h. The plate was washed four times in wash buffer. To each well, 50 ␮L rabbit anti-CRH-BP Ab code 5144 (1:5000 in assay buffer) were added. This was incubated at 22 C for 1 h, and the plate was washed four times with wash buffer. Fifty microliters of goat antirabbit horseradish peroxidase-linked secondary AB (GAR*hrp) at 1:5000 dilution in assay buffer (1 ⫻ TBST with 1% BSA, pH 7.5) were then added to each well and incubated at 22 C for 1 h. The plate was washed four times with wash buffer and developed with 50 ␮L TMB microwell peroxidase substrate solution (Kirkegaard & Perry Laboratories, Gaithersburg, MD) for 3–5 min. The reaction was stopped using 50 ␮L 0.2 n sulfuric acid stop solution (0.2 n H2SO4), and the plate was read in a spectrophotometer (Beckman Coulter, Inc., Palo Alto, CA) at 450 nm. The assay process for bound and free CRH was the same as that for CRH-BP. For bound CRH determination in each assay plate, wells A3 to H12 were coated overnight with anti-CRH-BP monoclonal antibody (100 ␮L/well; 5 ␮g/mL in 0.05 m NaHCO3 buffer pH 9.5), and wells A1–2 to H1–2, used to generate a standard CRH curve, were coated with an anti-CRH antibody as described for the CRH enzyme-linked immunosorbent assay (ELISA). The next morning plates were washed once with buffer (1 ⫻ PBS/0.05% Tween 20) and blocked for 1 h at room temperature with 100 ␮L/well 1% casein TBS (50 mm Tris-HCl and 100 mm NaCl buffer, pH 7.5). Then the blocking buffer was decanted, and 50-␮L samples or standards (diluted in TTBS assay buffer containing 1% BSA) were added to the appropriate wells and allowed to bind to either

JCE & M • 2001 Vol. 86 • No. 6

the anti-CRH-BP monoclonal antibody or the assay buffer, respectively, for 1 h at room temperature. Before assay plates were washed four times with wash buffer and exposed to a rabbit anti-CRH antibody (1:2000 in assay buffer) for 1 h at room temperature, the sample’s supernatants remaining after 1 h anti-CRH-BP monoclonal antibody capture of the bound CRH/CRH-BP complex were transferred and stored on a separate plate to await free CRH determination using the CRH ELISA. The assay plates then were washed four times with wash buffer. Incubations and development of the color reaction were as described for the CRH-BP ELISA. The quantity of CRH captured by the complex of anti-CRH-BP monoclonal antibody and anti-CRH antibody was compared with the amount of CRH captured by an anti-CRH antibody used for the standard curve. In the case of free CRH determination, plates were coated as described for the bound CRH/CRH-BP ELISA, except for the addition of 5 ␮g/mL of a protein G affinity-purified sheep anti-CRH antibody. Each plate (wells A1–2 to H1–2) has its own standard CRH curve. The plates were blocked as described for the CRH-BP ELISA and then exposed to samples (the supernatants from the CRH/CRH-BP ELISA) and standards diluted in assay buffer and incubated for 1 h at room temperature. The plates were then washed four times with wash buffer and exposed to a rabbit anti-CRH antibody (100 ␮L/well; 1:2000 dilution in TTBS buffer). Incubations and development of the color reaction were the same as described for the CRH-BP ELISA. Levels of cortisol were determined using a commercial kit (Milenia, Diagnostic Products, Los Angeles, CA). Briefly, in a 96-well microtiter plate 25 ␮L sample or calibration standards were combined with 100 ␮L ligand-labeled cortisol and 100 ␮L cortisol antiserum. The mixture was incubated for 1 h, then 25 ␮L cortisol enzyme-labeled antiligand were added to all wells to a final volume of 250 ␮L. The samples were incubated for an additional 30 min, and the plate was washed four times with 300 ␮L buffered wash solution containing buffered saline, surfactants, and a preservative. Finally, 200 ␮L TMB/substrate solution (3,3⬘,5,5⬘-tetramethylbenzidine) were added to all wells and allowed to develop in the dark for 30 min. After the incubation, 50 ␮L stop solution (acidic solution supplied in kit) were added to all wells, and the plate was read in a spectrophotometer (Beckman Coulter, Inc.) at 450 nm within 15 min. All plasma sample values were calculated using a fourparameter logistic equation based on a standard curve obtained simultaneously. Each plasma sample from patients was assessed in two or three independent assays for measurement of total, bound, and free CRH, CRH-BP, and plasma cortisol. Although there was variability between human subjects, the values in any one patient were reproducible.

Survey and behavioral measurements Participants completed three questionnaires, the first (I) just before inclusion, the second (II) at 30 weeks gestation, and the third (III) at birth. Questionnaire I contained items regarding previous and present medical history, questionnaire II contained requests for social and demographic information, and questionnaire III concerned present urogenital and obstetric problems. Women who delivered before 30 weeks gestation were asked to complete questionnaires II and III after delivery. All questionnaires were evaluated while the participant was present, and a midwife assisted with information on questionnaire III. Questionnaire II included items about potentially stressful life situations, such as serious illness or death in the family or financial problems. Participants gave detailed information about life events, home duties, and demands at work. Also, participants provided information about behaviors and habits, including smoking, risk-taking behavior (defined as seldom or no use of a seat belt while driving), and selfreported psychosocial stressors. Classification by socio-economic status was based on detailed job descriptions and educational backgrounds for each participant and her partner according to principles set by the Institute for Social Research (Copenhagen, Denmark) (32). The highest value for either the participant or her partner was set as the label for women with a partner.

Statistical analyses The Wilcoxon signed rank sum test was used for comparisons of hormone levels between preterm and term delivery groups. Univariate comparisons of proportions of risk factors were carried out using the ␹2

NEUROENDOCRINE, MEDICAL, AND BEHAVIORAL RISK FACTORS OF PRETERM BIRTH test. Univariate and multivariate odds ratios and confidence intervals were calculated for risk factors. Differences in hormone levels of risk factor groups were calculated using Student’s t tests. All data were analyzed using the software package SPSS version 10 (SPSS, Inc., Chicago, IL).

Results

Demographic characteristics of the subjects included in the study, including maternal age and socioeconomic status, are described in Table 1. TABLE 1. Demographic characteristics of the preterm delivery sample (n ⫽ 84) and entire control sample (n ⫽ 400) Characteristic

Maternal age (yr) ⬍20 20 –25 26 –29 30 –35 ⬎35 Socioeconomic status group 1 (highest) 2 3 4 5 (lowest) Marital status Not single Single Medical variables Multiparous Nulliparous Multigravid Primigravid

Preterm delivery [n (%)]

Term delivery [n (%)]

1 (1.2) 28 (33.3) 35 (41.7) 14 (16.7) 6 (7.1)

3 (0.8) 83 (20.7) 158 (39.5) 135 (33.8) 21 (5.2)

4 (4.8) 9 (10.7) 19 (22.6) 29 (34.5) 15 (17.9)

34 (8.5) 78 (19.5) 109 (27.2) 99 (24.8) 48 (12.0)

79 (94.0) 5 (6.0)

378 (94.5) 22 (5.5)

39 (46.4) 45 (53.6) 54 (64.3) 30 (35.7)

190 (47.5) 210 (52.5) 247 (61.7) 153 (38.3)

2547

Subject variables associated with preterm delivery

Table 2 describes various medical, socioeconomic, behavioral, and other possible risk factors that might be linked to preterm delivery. According to univariate analyses, previous preterm delivery was the best predictor of preterm delivery in this sample [␹2 ⫽ 22.06; P ⬍ 0.0001; odds ratio (OR), 5.5; confidence interval (CI), 2.4 –12.2]. Fifteen percent of those who delivered preterm had a history of preterm delivery. The experience of significant contractions during the index pregnancy was a second medical variable associated with preterm delivery in this sample (␹2 ⫽ 4.23; P ⫽ 0.04; OR, 1.7; CI, 1.0 –2.8). In addition, the rate of women in the preterm group who reported having a serious medical disease was higher (8.3%) than that in the term group (3.3%; ␹2 ⫽ 4.54; P ⫽ 0.03; OR, 2.7; CI, 1.1–7.0). Nulliparity was not associated with either term or preterm delivery. Multiple logistic regression indicated that previous preterm delivery (OR, 7.6; CI, 2.9 – 20.0) was still the best predictor of preterm delivery, whereas serious medical disease and significant contractions were less strongly associated. However, multiple logistic regression indicated that nulliparity was strongly associated with preterm delivery. Several socioeconomic and work-related variables were associated with preterm delivery. A low level of education (i.e. finished school before ninth grade with no further education) was more prevalent in the preterm than in the term delivery group (␹2 ⫽ 6.98; P ⫽ 0.008; OR, 2.9; CI, 1.3– 6.5). Compared with the women in the term delivery group, more than twice as many women in the preterm group had a low level of education. Members of the preterm delivery group were also more likely to be public assistance recipients (␹2 ⫽

TABLE 2. Percentage of women in term and preterm delivery groups with selected medical, demographic, and other possible risk variables Preterm delivery (%)

Medical variables Previous preterm delivery Serious medical disease Nulliparity Significant contractions Socio-economic and employment-related variables Low level of educationa Public assistance recipient 42⫹ workhours/week 6⫹ workhours standing/day 6⫹ workhours walking/day Unemployed Behavioral and other variables Risk-taking behaviorb High sexual activity Smokingc Self-reported social stress Biological variables Total CRH Bound CRH CRH-binding protein

Univariate analyses

Multivariate analysis

Yes

No

P value

Crude OR (CI)

P value

Adj OR (CI)

15.5 8.3 53.6 34.5

3.2 3.3 52.5 23.8

⬍0.0001 0.03 0.86 0.04

5.451 (2.426 –12.245) 2.706 (1.046 –7.004) 1.044 (0.651–1.673) 1.693 (1.021–2.806)

⬍0.0001 0.09 0.02 0.16

7.560 (2.864 –19.96) 2.538 (0.861–7.49) 2.015 (1.117–3.634) 1.542 (0.839 –2.834)

11.9 35.7 44.0 59.5 58.3 31.0

4.5 21.8 33.5 45.8 41.5 23.0

0.008 0.007 0.07 0.02 0.005 0.12

2.868 (1.273– 6.460) 1.999 (1.206 –3.314) 1.563 (0.969 –2.521) 1.744 (1.081–2.813) 1.973 (1.225–3.180) 1.501 (0.894 –2.519)

0.03 0.24 0.93 0.79 0.36 0.46

2.930 (1.105–7.772) 1.634 (0.718 –3.721) 1.040 (0.447–2.420) 0.894 (0.387–2.062) 1.499 (0.629 –3.573) 0.659 (0.219 –1.977)

14.3 24.4 3.6 13.1

4.5 16.7 1.5 8.3

0.001 0.10 0.20 0.16

3.539 (1.633–7.659) 1.608 (0.910 –2.841) 2.450 (0.60 –10.001) 1.676 (0.810 –3.467)

0.02 0.39 0.41 0.43

2.828 (1.159 – 6.900) 1.329 (0.698 –2.530) 1.971 (0.396 –9.801) 1.417 (0.596 –3.367)

23.8 32.1 35.7

11.3 15.8 24.5

0.004 0.001 0.041

2.465 (1.366 – 4.448) 2.534 (1.490 ⫾ 4.309) 1.712 (1.037–2.826)

0.01 0.003 0.005

2.358 (1.214 – 4.582) 2.499 (1.375– 4.541) 2.182 (1.258 –3.786)

OR, Odds ratio; CI, confidence interval; Adj OR, logistic regression. a Did not complete school past ninth grade. b Defined as no or rare use of seat belt. c 15⫹ cigarettes/day at enrollment.

2548

ERICKSON ET AL.

7.39; P ⫽ 0.007; OR, 2.0; CI, 1.2–3.3). Unemployment was not significantly higher in the preterm than in the term group. Significantly more women in the preterm delivery group worked at jobs requiring 6 or more hours of standing per day (␹2 ⫽ 5.28; P ⫽ 0.02; OR, 1.7; CI 1.1–2.8). Also, more women in the preterm group spent 6 or more work hours walking per day (␹2 ⫽ 7.97; P ⫽ 0.005; OR, 2.0; CI, 1.2–3.2). Working more than 42 h/week was more prevalent in the preterm delivery group (44.0%) than in the term delivery group (33.5%). Multivariate analysis indicated that the only socioeconomic variable strongly associated with preterm delivery was a low level of education; the other variables (public assistance, unemployment, and standing or walking at work) were less strongly associated. Of the behavioral variables, risk-taking behavior (defined as no or seldom use of safety belt while driving a vehicle) was more prevalent in the preterm than the term delivery group (␹2 ⫽ 11.43; P ⫽ 0.001; OR, 3.5; CI, 1.6 –7.7). Other variables

FIG. 2. Levels of hormones in pregnant women who delivered preterm and at term. A, Free CRH levels; B, bound CRH levels; C, total CRH levels; D, CRH-BP levels. Seven to 23 week measurements include all participants; 27–37 week measurements include those women who did not receive betamethasone treatment. Bars represent the SEM. *, P ⬍ 0.05; **, P ⬍ 0.0001.

JCE & M • 2001 Vol. 86 • No. 6

in this category, such as smoking and self-reported psychosocial stressors, were not more prevalent in the preterm than in the term delivery group. Multivariate analysis indicated that risk-taking behavior was still significantly associated with preterm delivery (OR, 2.8; CI, 1.2– 6.9). Neuroendocrine levels at 7–23 weeks gestation (gestational age-matched comparisons)

All preterm and term subjects were used in the analyses. Figure 2 illustrates total, bound, and free CRH and CRH-BP levels during pregnancy. Maternal levels of total CRH (P ⫽ 0.01) and bound CRH (P ⫽ 0.03) were significantly different, with women in the preterm delivery group having higher total and bound CRH levels than those in the term group. Maternal levels of free CRH during early pregnancy, however, were not significantly different between the preterm and term groups (P ⫽ 0.08). In addition, levels of CRH-BP

NEUROENDOCRINE, MEDICAL, AND BEHAVIORAL RISK FACTORS OF PRETERM BIRTH

were significantly higher (P ⫽ 0.01; see Fig. 2D) in the preterm than the term group. Figure 3 illustrates plasma cortisol levels in the concurrent sample during various stages of pregnancy. Levels of cortisol in the term and preterm groups were not significantly different during early pregnancy. Neuroendocrine levels during 7–23 weeks gestation and risk factors

For each risk factor that was significantly linked to preterm delivery in the univariate analyses, new analyses were performed on hormone levels of women who delivered preterm and at term. This was to determine whether a relation exists between risk factors and levels of the hormones under investigation. Because some sample sizes are small and statistical power is low for some risk factor subgroups, Table 3 describes the mean levels of total, bound, and free CRH, CRH-BP, and cortisol found in women with the risk factors who had preterm and term deliveries. The results indicated that for the three medical variables more prominent in those who delivered preterm than term, significantly different levels of total, bound, and free CRH, CRH-BP, and cortisol were not detected between women who delivered preterm and those who delivered at term with previous preterm delivery or serious medical disease. For subjects who experienced significant contractions during the index pregnancy, total CRH (P ⫽ 0.048) and bound CRH (P ⫽ 0.007) were significantly higher in women who delivered preterm than those who delivered at term. Some socioeconomic and employment-related variables were linked to CRH levels. Women receiving public assis-

FIG. 3. Cortisol levels in pregnant women who delivered preterm and at term. Seven to 23 week measurements include all participants; 27–37 week measurements include those women who did not receive betamethasone treatment. Bars represent the SEM. *, P ⬍ 0.001.

2549

tance who delivered preterm had higher bound CRH (P ⫽ 0.024) than those who delivered at term. Similarly, women who stood at work for 6 or more hours a day and who delivered preterm had higher levels of bound CRH (P ⫽ 0.05) than those with the same risk factor who delivered at term. Women who walked 6 or more h a day at work and delivered preterm also tended to have higher bound CRH levels than those with the same risk factor who delivered at term (P ⫽ 0.07). The one behavioral variable linked to preterm delivery, risk-taking behavior, was not linked to CRH. Levels of total, bound, and free CRH, CRH-BP, and cortisol in those engaging in risk-taking behavior who delivered preterm or term were not significantly different. Neuroendocrine levels at 27–37 weeks gestation (gestational age-matched comparisons)

As some of the women were treated with betamethasone, analyses were performed that included only the women who did not receive this medication (n ⫽ 59) and a control group matched to the preterm group by gestational age (n ⫽ 166). Maternal levels of total (P ⬍ 0.0001), bound (P ⬍ 0.0001), and free CRH (P ⬍ 0.0001) levels were higher in the preterm group than in the gestational age-matched term group (Fig. 2). Conversely, levels of CRH-BP were lower in the preterm group than in the gestational age-matched term group (P ⬍ 0.0001). Furthermore, cortisol levels were higher in the preterm than in the gestational age-matched term group (P ⫽ 0.001; Fig. 3). Neuroendocrine levels at 27-37 weeks gestation and risk factors

Women who did not receive betamethasone treatment were selected for the 27–37 gestational week analyses of risk factors and levels of total, bound, and free CRH, CRH-BP, and cortisol levels. Because of the small sample sizes in some of the comparisons and the resulting possibility for type II errors, Table 4 provides mean levels of hormones within each risk factor in gestational age-matched women who delivered preterm and at term as well as P values. Gestational age-matched total, bound, and free CRH levels were generally higher in those who delivered preterm than in those who delivered at term. However, some of the differences within risk factors were not statistically significant. Specifically, bound CRH levels were not significantly different in the presence of a serious medical disease between women who delivered at term and preterm. Free CRH levels were not significantly different between women who delivered preterm and at term in the presence of previous preterm delivery and a serious medical disease. When CRH-BP levels between preterm and term deliveries were significantly different, binding protein was higher in term than in preterm deliveries. For previous preterm delivery, serious medical disease, significant contractions, and low level of education, binding protein levels were not significantly different between term and preterm deliveries. As with CRH levels, mean cortisol levels were higher in the preterm than term delivery groups in the presence of each risk factor, although differences reached statistical signifi-

2550

JCE & M • 2001 Vol. 86 • No. 6

ERICKSON ET AL.

TABLE 3. Mean maternal levels during 7th–23rd weeks gestation of total, free, and bound CRH; CRH-binding protein; and cortisol in women with risk factors who delivered preterm and term Preterm delivery

Risk factor

Total CRH (ng/L)

Free CRH (ng/L)

Bound CRH (ng/L)

CRH-BP (ng/mL)

Cortisol (nmol/L)

Previous preterm delivery

No (n ⫽ 13) Yes (n ⫽ 13) P

86.1 (169.5) 173.1 (265.4) NS

18.6 (45.2) 13.4 (37.4) NS

67.6 (125.1) 159.8 (243.2) NS

107.9 (61.9) 102.1 (52.7) NS

492.95 (306.36) 434.13 (154.70) NS

Serious medical disease

No (n ⫽ 13) Yes (n ⫽ 7) P

52.4 (75.8) 77.5 (89.5) NS

3.2 (11.1) 29.4 (59.4) NS

49.3 (70.9) 48.2 (82.7) NS

90.3 (39.5) 121.1 (53.5) NS

454.57 (237.30) 382.92 (91.43) NS

Significant contractions

No (n ⫽ 95) Yes (n ⫽ 29) P

75.5 (166.8) 153.1 (229.2) 0.048

18.2 (97.5) 21.6 (51.7) NS

57.3 (105.2) 131.6 (184.9) 0.007

130.8 (267.6) 125.3 (79.7) NS

423.12 (169.29) 384.14 (109.62) NS

Low level of education

No (n ⫽ 18) Yes (n ⫽ 10) P

63.4 (112.0) 82.2 (130.4) NS

0.2 (0.4) 0.1 (0.0) NS

63.3 (112.0) 82.2 (130.4) NS

261.1 (601.2) 144.2 (96.5) NS

383.53 (120.73) 368.96 (131.80) NS

Public assistance recipient

No (n ⫽ 86) Yes (n ⫽ 30) P

117.7 (376.5) 159.8 (217.5) NS

53.3 (262.5) 27.1 (78.2) NS

64.5 (125.2) 132.8 (178.3) 0.024

134.1 (283.9) 151.5 (210.6) NS

433.41 (176.14) 450.08 (185.68) NS

6⫹ h standing/day at work

No (n ⫽ 181) Yes (n ⫽ 50) P

101.4 (287.4) 146.7 (232.3) NS

35.5 (190.4) 36.9 (107.1) NS

66.0 (131.2) 109.9 (166.9) 0.05

122.3 (203.5) 148.7 (176.4) NS

424.58 (165.24) 464.01 (212.64) NS

6⫹ h walking/day at work

No (n ⫽ 164) Yes (n ⫽ 49) P

103.5 (296.0) 142.8 (225.1) NS

35.7 (195.6) 34.0 (107.4) NS

67.9 (130.0) 108.9 (163.2) 0.07

121.8 (212.5) 151.7 (178.0) NS

420.18 (167.80) 457.03 (219.48) NS

Risk-taking behavior

No (n ⫽ 18) Yes (n ⫽ 12) P

169.3 (449.5) 129.1 (157.2) NS

82.02 (329.1) 0.1 (0.0) NS

87.4 (137.4) 129.1 (157.2) NS

105.3 (102.7) 126.0 (80.5) NS

434.65 (252.01) 467.60 (178.29) NS

The values are the means and SD from each individual patient sample measured in two or three independent assays. Although there is an inherent degree of variability in the levels of CRH and CRH-BP between human subjects, the values for any given patient on all measures were reproducible.

cance for only three risk factors: specifically, serious medical disease (P ⫽ 0.048), 6 or more hours of standing daily at work (P ⫽ 0.018), and 6 or more hours of walking daily at work (P ⫽ 0.038). Preterm and term delivery (nongestational age-matched comparisons)

As with the 27–37 gestational week measurements, additional analyses were performed, including only those preterm (n ⫽ 59) and term subjects (n ⫽ 296) who did not receive betamethasone treatment. Maternal levels of free CRH (P ⬍ 0.0001) in the term group were higher than those in the preterm group. Total CRH (P ⬍ 0.0001) levels were also higher in the term than the preterm group. However, bound CRH levels at delivery were not significantly different between the two groups (P ⫽ 0.32). In addition, CRH-BP levels were not significantly different between the two groups (P ⫽ 0.39); this was also true for cortisol levels (P ⫽ 0.83). Discussion

Our results provide further evidence that elevated levels of both CRH and cortisol are linked to preterm labor and delivery (18, 19, 21, 25, 27). CRH, however, is one of several systems involved in this process, and the current results do not elucidate the extent to which maternal CRH levels are associated with and are predictive of preterm labor and delivery (5, 33–35). Definitely, the first trimester CRH levels were not a good predictor of preterm birth; 27–37 week CRH levels, however, were more predictive in our sample. Free

CRH levels at 27–37 weeks were a better predictor than total or bound CRH levels, a finding consistent with the idea that it is the free form of CRH that is acting upon the pregnant uterus. Increased CRH-BP levels during normal pregnancy may prevent premature increases in free CRH levels and, therefore, premature delivery (17, 18, 20). In our sample of women, CRH-BP levels were higher in those from the preterm than in those from the term group in the first trimester. During gestational weeks 27–37, however, the levels of CRH-BP decreased in the preterm group and increased in the term group, such that the term group had significantly higher levels of CRH-BP than the preterm group. Therefore, the corresponding levels of total, bound, and free CRH were higher in the preterm than in the term group during 27–37 weeks gestation. Our results suggest that a number of social and behavioral factors (risk-taking behavior, level of education) might be linked to CRH, cortisol, and preterm delivery (25, 26). Nevertheless, we found that the most salient predictor of preterm labor and delivery was prior preterm deliveries. CRH secretion is increased in a number of adverse conditions in both the placenta (22, 23, 25) and the brain (36 –39). When it is overexpressed in the placenta, women are vulnerable to preterm labor and delivery (7, 40). When CRH is overexpressed in the brain of animals or humans, these individuals are vulnerable to the experience of excessive fear and anxiety (40). Interestingly, the neuropeptide CRH1 receptor antagonist antalarmin, which has been shown to delay parturition in sheep (41), has also been shown to reduce fear

NEUROENDOCRINE, MEDICAL, AND BEHAVIORAL RISK FACTORS OF PRETERM BIRTH

2551

TABLE 4. Mean maternal levels during 27th–37th weeks gestation of total, free, and bound CRH; CRH-binding protein; and cortisol in women with risk factors who delivered preterm and term and did not receive betamethasone treatment Preterm delivery

Risk factor

Total CRH (ng/L)

Free CRH (ng/L)

Bound CRH (ng/L)

CRH-BP (ng/mL)

Cortisol (nmol/L)

Previous preterm delivery

No (n ⫽ 7) Yes (n ⫽ 10) P

1172.6 (832.8) 2444.2 (668.5) 0.003

194.6 (323.4) 594.6 (630.9) NS

978.0 (593.5) 1849.6 (164.9) 0.000

133.1 (40.1) 155.8 (114.8) NS

710.42 (384.16) 773.02 (264.09) NS

Serious medical disease

No (n ⫽ 6) Yes (n ⫽ 5) P

1764.5 (954.1) 2139.2 (771.0) NS

273.5 (331.7) 529.8 (551.7) NS

1491.0 (625.4) 1609.4 (412.2) NS

138.3 (42.1) 89.2 (42.5) NS

708.70 (285.67) 1358.86 (627.54) 0.048

Significant contractions

No (n ⫽ 48) Yes (n ⫽ 18) P

1377.9 (848.4) 2291.1 (812.1) 0.000

201.3 (350.6) 611.1 (637.9) 0.001

1176.6 (601.4) 1680.0 (321.5) 0.001

187.5 (85.9) 156.3 (106.9) NS

764.16 (290.85) 896.73 (456.59) NS

Low level of education

No (n ⫽ 4) Yes (n ⫽ 7) P

1168.8 (384.0) 2671.1 (931.0) 0.014

5.8 (11.5) 845.2 (12.8) 0.047

1163.0 (378.1) 1826.0 (356.2) 0.017

133.0 (46.7) 141.7 (99.5) NS

602.43 (162.86) 914.44 (497.09) NS

Public assistance recipient

No (n ⫽ 44) Yes (n ⫽ 20) P

1344.7 (715.8) 2375.1 (979.8) 0.000

131.5 (248.7) 686.9 (657.9) 0.000

1213.2 (555.7) 1688.2 (485.5) 0.000

187.9 (90.7) 130.7 (73.5) 0.001

826.35 (362.60) 1025.19 (503.55) 0.007

6⫹ h standing/day at work

No (n ⫽ 80) Yes (n ⫽ 32) P

1408.1 (833.6) 2235.6 (901.8) 0.000

195.9 (359.2) 572.6 (617.5) 0.000

1212.2 (588.8) 1663.0 (452.3) 0.000

190.6 (113.5) 123.4 (85.7) 0.003

811.62 (340.63) 1082.38 (852.95) 0.018

6⫹ h walking/day at work

No (n ⫽ 72) Yes (n ⫽ 32) P

1426.5 (779.8) 2246.8 (933.7) 0.000

180.4 (325.6) 601.8 (627.7) 0.000

1246.2 (560.1) 1645.0 (464.2) 0.001

185.2 (108.4) 134.3 (89.0) 0.022

815.45 (352.99) 1058.66 (832.25) 0.038

Risk-taking behavior

No (n ⫽ 8) Yes (n ⫽ 8) P

1655.2 (925.4) 2739.0 (612.4) 0.015

296.6 (496.2) 968.0 (505.9) 0.018

1358.6 (565.8) 1771.0 (263.3) NS

170.9 (73.1) 93.8 (52.1) 0.029

787.17 (2528.89) 1044.28 (347.99) NS

The values are the means and SD from each individual patient sample measured in two or three independent assays. Although there is an inherent degree of variability in the levels of CRH and CRH-BP between human subjects, the values for any given patient on all measures were reproducible.

or anxiety-related behaviors in rats and monkeys (38, 42, 43). Human and animal research has suggested a link between psychological/psychosocial stress during pregnancy and preterm delivery (25, 26, 44), with modulation by CRH being the bridge between the two. Our results indicate that physical (e.g. extensive standing or walking at work) and pathological (e.g. serious medical disease) stressors may be additional risk factors for preterm birth. In addition to affecting the timing of labor and delivery, the increased levels of CRH and cortisol during stressful pregnancy may have long-term consequences for both behavior and physiological regulation of the offspring. It is well known from experimental animal studies that maternal stress leads to life-long elevations in levels of both CRH (45, 46) and glucocorticoids (44) as well as to behavioral and metabolic abnormalities (47, 48). Hence, CRH receptor antagonists may be useful not only for retarding labor, but also for preventing the long-term effects of intrauterine stress in the offspring. References 1. Hack M, Fanaroff AA. 1993 Outcomes of extremely immature infants—a perinatal dilemma. N Engl J Med. 329:1649 –1650. 2. Morrison JC. 1990 Preterm birth: a puzzle worth solving. Obstet Gynecol. 76:5.S–12S. 3. Rush RW, Keirse MJ, Howat P, Baum JD, Anderson AB, Turnbull AC. 1976 Contribution of preterm delivery to perinatal mortality. Br Med J. 2:965–968. 4. Berkowitz GS, Papiernik E. 1993 Epidemiology of preterm birth. Epidemiol Rev. 15::414 – 443. 5. Goldenberg RL, Hauth JC, Andrews WW. 2000 Intrauterine infection and preterm delivery. N Engl J Med. 342:1500 –1507.

6. Kramer MS. 1987 Determinants of low birth weight: methodological assessment and meta-analysis. Bull WHO. 65:663–737. 7. Challis JR, Matthews SG, VanMeir C, Ramirez MM. 1995 Current topic: the placental corticotropin-releasing hormone-adrenocorticotropin axis. Placenta. 16:481–502. 8. Frim DM, Emanuel RL, Robinson BG, Smas CM, Alder GK, Majzoub JA. 1988 Characterization and gestational regulation of corticotropin releasing hormone messenger RNA in human placenta. J Clin Invest. 82:287–292. 9. Riley SC, Walton JC, Herlick JM, Challis JR. 1991 The localization and distribution of corticotropin-releasing hormone in the human placenta and fetal membranes throughout gestation. J Clin Endocrinol Metab. 72:1001–1007. 10. Jones SA, Brooks AN, Challis JR. 1989 Steroids modulate corticotropin-releasing hormone production in human fetal membranes and placenta. J Clin Endocrinol Metab. 68:825– 830. 11. Robinson BG, Emanuel RL, Frim DM, Majzoub JA. 1988 Glucocorticoid stimulates expression of corticotropin-releasing hormone gene in human placenta. Proc Natl Acad Sci USA. 85:5244 –5248. 12. Wolfe CD, Patel SP, Campbell EA, Linton EA, Anderson J, Lowry PJ, Jones MT. 1988 Plasma corticotropin-releasing factor (CRF) in normal pregnancy. Br J Obstet Gynaecol. 95:997–1002. 13. Cambell EA, Linton EA, Wolfe CD, Scraggs PR, Jones MT, Lowry PJ. 1987 Plasma corticotropin releasing hormone concentrations during pregnancy and parturition. J Clin Endocrinol Metab. 64:1054 –1059. 14. Warren WB, Patrick SL, Goland RS. 1992 Elevated maternal plasma corticotropin-releasing hormone levels in pregnancies complicated by preterm labor. Am J Obstet Gynecol. 166:1198 –1204. 15. Grammatopoulos DK, Hillhous EW. 1999 Role of corticotropin-releasing hormone in onset of labor. Lancet. 354:1546 –1549. 16. Behan DP, DeSouza EB, Lowry PJ, Potter E, Sawchenko P, Vale WW. 1995 Corticotropin releasing factor (CRF) binding protein: a novel regulator of CRF and related peptides. Front Neuroendocrinol. 16:362–382. 17. Linton EA, Wolfe CD, Behan DP, Lowery PJ. 1988 A specific carrier substance for human corticotropin-releasing factor in late gestational maternal plasma which could mask the ACTH-releasing activity. Clin Endocrinol (Oxf). 28:315–324. 18. Berkowitz GS, Lapinski RH, Lockwood CJ, Florio P, Blackmore-Prince C, Petraglia F. 1996 Corticotropin-releasing factor and its binding protein: maternal serum levels in term and preterm deliveries. Am J Obstet Gynecol. 174:1477–1483.

2552

ERICKSON ET AL.

19. Hobel CJ, Arora CP, Korst LM. 1999 Corticotropin-releasing hormone and CRH-binding protein: differences between patients at risk for preterm birth and hypertension. Ann NY Acad Sci. 897:54 – 65. 20. Orth DN, Mount CD. 1987 Specific high affinity binding protein for human corticotropin releasing hormone in normal human plasma. Biochem Biophys Res Commun. 143:411– 417. 21. Majzoub JA, McGregor JA, Lockwood CJ, Smith R, Taggart MS, Schulkin J. 1999 A central theory of preterm and term labor: putative role for corticotropin-releasing hormone. Am J Obstet Gynecol. 180:S232—S241. 22. Petraglia F, Aguzzoli P, Florio P, Baumann P, Genazzani AD, DiCarlo C, Romero R. 1995 Maternal plasma and placental immunoreactive corticotropinreleasing factor concentrations in infection-associated term and pre-term delivery. Placenta. 16:157–164. 23. Goland RS, Tropper PJ, Warren WB, Stark RI, Jozak SM, Conwell IM. 1995 Concentrations of corticotropin-releasing hormone in the umbilical-cord blood of pregnancies complicated by pre-eclampsia. Reprod Fertil Dev. 7:1227–1230. 24. Goland RS, Jozak S, Warren WB, Conwell IM, Stark RI, Tropper PJ. 1993 Elevated levels of umbilical cord plasma corticotropin-releasing hormone in growth-retarded fetuses. J Clin Endocrinol Metab. 77:1174 –1179. 25. Hobel CJ, Dunkel-Schetter C, Roesch SC, Castro LC, Arora CP. 1999 Maternal plasma corticotropin-releasing hormone is associated with stress at 20 weeks gestation in pregnancies ending in preterm delivery. Am J Obstet Gynecol. 180:S257—S263. 26. Wadhwa PD, Dunkel-Schetter C, Chicz-DeMet A, Portoo M, Sandman CA. 1996 Prenatal psychosocial factors and the neuroendocrine axis in human pregnancy. Psychosomat Med. 58:432– 446. 27. Bisits A, Madsen G, McLean M, O’Callaghan S, Smith R, Giles W. 1998 Corticotropin-releasing hormone: a biochemical predictor of preterm delivery in a pilot andomized trial of the treatment of preterm labor. Am J Obstet Gynecol. 178:862– 866. 28. Wolfe CDA, Patel SP, Linton EA, et al. 1988 Plasma corticotropin-releasing factor (CRF) in abnormal pregnancy. Br J Obstet Gynaecol. 95:1003–1006. 29. Korebrits C, Yu DH, Ramirez MM, Marinoni E, Bocking AD, Challis JR. 1998 Antenatal glucocorticoid administration increases corticotropin-releasing hormone in maternal plasma. Br J Obstet Gynaecol. 105:556 –561. 30. Thorsen P, Jensen IP, Ebbesen N, Arpi M, Bremmelgaard A, Jeune B, Møller BR. 1998 Few microorganisms associated with bacterial vaginosis may constitute the pathological core. A population based microbiological study among 3596 pregnant women. Am J Obstet Gynecol. 178:580 –587. 31. Behan DP, Khongsaly O, Liu XJ, et al. 1996 Measurement of corticotropinreleasing factor (CRF), CRF-binding protein (CRF-BP), and CRF/CRF-BP complex in human plasma by two-site enzyme-linked immunoabsorbant assay. J Clin Endocrinol Metab. 81:2579 –2586. 32. Enevoldsen B, Michelsen N, Friis-Hasche E, Kamper-Jurgensen F. 1980 Social classification. Svalastoga’s subdivision according to social status rank the social grouping employed by the Institute of Social Research. Ugeskr Laeger. 142:544 –550. 33. Petraglia F, Volpe A, Genazzani AR, River J, Sawchenko PE, Vale W. 1990 Neuroendocrinology of the human placenta. Front Neuroendocrinol. 11:6 –37.

JCE & M • 2001 Vol. 86 • No. 6

34. McGregor JA, Jackson GM, Lachelin GC, Goodwin TM, Artal R, Hastings C, Dullien V. 1995 Salivary estriol as risk assessment for preterm labor: a prospective trial. Am J Obstet Gynecol. 173:1337–1342. 35. Pepe GJ, Albrecht ED. 1995 Actions of placental and fetal adrenal steroid hormones in primate pregnancy. Endocr Rev. 16:608 – 648. 36. Cratty MS, Ward HE, Johnson EA, Azzaro AJ, Birkle DL. 1995 Prenatal stress increases corticotropin-releasing factor content and release in rat amygdala minces. Brain Res. 675:297–302. 37. Koob GF. 1993 The role of corticotropin-releasing hormone in behavioral responses to stress. In: Chadwick K, Marsh J, Ackrill K, eds. Corticotropinreleasing factor. New York: Wiley & Sons; 277–289. 38. Habib KE, Weld KP, Rice KC, et al. 2000 Oral administration of a corticotropin-releasing hormone receptor antagonist significantly attenuates behavioral, neuroendocrine, and autonomic responses to stress in primates. Proc Natl Acad Sci USA. 97:6079 – 6084. 39. Makino S, Gold PW, Schulkin J. 1994 Effects of corticosterone on CRH mRNA and content in the bed nucleus of the stria terminalis: comparison with the effects in the central nucleus of the amygdala and the paraventricular nucleus of the hypothalamus. Brain Res. 657:141–149. 40. Schulkin J. 1999 Corticotropin-releasing hormone signals adversity in both the placenta and the brain: regulation by glucocorticoids and allostatic overload. J Endocrinol. 161:349 –356. 41. Chan EC, Falconer J, Madsen G, Rice KC, Webster EL, Chrousos GP, Smith R. 1998 A corticotropin-releasing hormone type I receptor antagonist delays parturition in sheep. Endocrinology. 139:3357–3360. 42. Deak T, Nguyen KT, Ehrlich AL, et al. 1999 The impact of the nonpeptide corticotropin releasing hormone antagonist antalarmin on behavioral and endocrine responses to stress. Endocrinology. 140:79 – 86. 43. Shepard JD, Barron KW, Myers DA. Corticosterone delivery to the amygdala increases corticotropin releasing factor mRNA in the central amygdaloid nucleus and anxiety-like behavior. Brain Res. 861:288 –295. 44. Takahashi LK, Turner JG, Kalin NH. 1998 Prolonged stress-induced elevation in plasma corticosterone during pregnancy in the rat: implications for prenatal stress studies. Psychoneuroendocrinology. 23:571–581. 45. Coplan JD, Andrews MW, Rosenblum LA, Owens MJ, Friedman S, Gorman JM, Nemeroff CB. 1996 Persistent levels of cerebrospinal fluid concentrations of corticotropin-releasing factor in adult nonhuman primates exposed to earlylife stressors: implications for the pathophysiology of mood and anxiety disorders. Proc Natl Acad Sci USA. 93:1619 –1623. 46. Williams MT, Hennessy MB, Davis HN. 1995 Corticotropin-releasing hormone administered to pregnant rats alters offspring behavior and morphology. Pharmacol Biochem Behav. 52:161–167. 47. Nathanielez PW. 1999 Life in the womb: the origin of health and disease. Ithaca, NY: Promethian Press. 48. Chrousos GP. 2000 The role of stress and the hypothalamic-pituitary-adrenal axis in the pathogenesis of the metabolic syndrome: neuro-endocrine and target tissue-related causes. Intern J Obes. 24:550 –555.