Chronic HPA activity in mothers with preterm delivery: A pilot nested case-control study

Abstract

BACKGROUND:

Chronic hypothalamic-pituitary-adrenal (HPA) axis activity role in the pathogenesis of preterm birth (PTB) remains unclear due to inconsistent measures with limited ability to monitor long-term cortisol concentrations. We explored this relationship using the novel method of assessing cortisol in hair, which is a valid and reliable measure of chronic HPA axis activity.

METHODS:

137 participants (40 PTB cases and 97 controls from a birth cohort of pregnant women in Peru) were interviewed and invited to provide a 9-cm hair sample from the posterior vertex position of the scalp (mean = 13 weeks gestation). Hair cortisol concentration (HCC) was determined using luminescence immunoassay and values were natural-log transformed. PTB cases were defined as women who delivered before 37 gestational weeks. Case-control differences were assessed using multivariable linear and logistic regressions.

RESULTS:

Overall, combined pre-conception and first-trimester HCC was 13% lower among cases as compared with controls (p-value = 0.01). Compared with controls, maternal HCC among PTB cases were 14% (p = 0.11), 10% (p = 0.22) and 14% (p = 0.08) lower for 3–6 months pre-conception, 0–3 months pre-conception, and first trimester, respectively. After adjusting for putative confounders, a 1-unit increase in HCC was associated with 55% reduced odds of PTB (aOR = 0.45; 95% CI: 0.17–1.17). For a 1-unit increase in HCC in the scalp-intermediate and scalp-distal segments (representing HCC concentrations in 0–3 months pre-conception and first trimester), the corresponding odds for PTB were 0.53 (95% CI: 0.19–1.48) and 0.39 (95% CI: 0.13–1.13), respectively.

CONCLUSIONS:

Women who deliver preterm, as compared with those who deliver at term, have lower preconception and first trimester HCC. Our findings suggest that HPA axis activation, integral to the adaptive stress-response system, may be chronically dysregulated in women at increased risk of PTB.

Keywords

HPA axis activity hair cortisol cortisol HCC preterm birth stress

1 Introduction

Preterm birth (PTB), a persistent clinical and public health problem, is a major cause of neonatal mortality and perinatal and pediatric morbidity including respiratory distress syndrome, necrotizing enterocolitis, and intraventricular hemorrhage and developmental delays [1 –3]. Compared with term infants, preterm infants are at an increased risk for immediate and long-term morbidity [4, 5] including increased risk of later-life cardiovascular-related conditions [6 –8]. Globally, approximately 15 million infants are delivered preterm and about 1 million die during the first year of life as a result of their prematurity [9]. In the United States, PTB affects more than 1 in 10 births. While PTB rates in the US have declined between 2007 and 2014, racial and ethnic disparities persist [1]; and although changes in medical care and policy shifts have led a slightly reduced incidence, public health interventions aimed at preventing PTB have been met with limited success and have been ineffective, in part, due to the heterogeneity of the condition [2, 10]. Available evidence shows the potential etiologic roles of maternal acute and chronic stressors on risk of PTB. Of note, exposure to various stressful life events, such as intimate partner abuse and history of childhood abuse, has been repeatedly associated with increased risk of PTB [11, 12]. However, a major challenge to a complete understanding of the potential causal roles of stress on PTB has been the absence of reliable, non-invasive and non-resource-demanding biomarker of stress and stress response. Cortisol, a glucocorticoid hormone released by the HPA axis, plays crucial roles in stress responses including, maintaining homeostatic conditions and functioning of the immune and neurologic systems. Recently, a new method has been developed for assessing cortisol in hair reflecting HPA axis activity [13, 14]. Despite the well-established association between stress and PTB, the role of the HPA axis activity in the pathogenesis of PTB remains unclear. An emerging literature has shown that hypocortisolism, a relative or absolute state of low cortisol, has played a key role. Although the underlying etiology has not yet been fully elucidated, hypocortisolism has been reported to reflect chronic HPA axis activity in responses to exposures to psychological trauma including childhood abuse [15], we conducted the present study to explore the extent to which hypocortisolism reflecting HPA axis activity in preconception and early in pregnancy is associated with PTB in a prospective cohort study of pregnant women attending prenatal clinics in Lima, Peru.

2 Methods

2.1 Study population and analytical population

This nested case-control study was embedded in an ongoing prospective cohort study, the Pregnancy Outcomes, Maternal, and Infant Study (PrOMIS). In this cohort, participants were recruited from women attending prenatal clinics at the Instituto Nacional Materno Perinatal (INMP) in Lima, Peru, from October 2014 to November 2015. The INMP is the primary reference establishment for maternal and perinatal care operated by the Ministry of Health of the Peruvian government. Women who were 18 years of age or older, were able to speak and read in Spanish, and initiated prenatal care before 16 weeks gestational age were invited to participate. Pregnant women were excluded if they had intellectual disabilities, twins, fetal malformation, or a history of chronic hypertension, diabetes mellitus, sepsis, or renal failure. The institutional review boards of the Harvard T.H. Chan School of Public Health and the INMP approved this study and written informed consent was obtained from all participants. After obtaining written informed consent, in private settings, structured questionnaires was administered to collect information on participants’ sociodemographic, hair characteristics, and medical and reproductive histories. Recruited women were then followed from early pregnancy to time of delivery. After delivery, maternal and infant medical records were abstracted for information on the course and outcome of pregnancy. Participants were asked to provide two 9 cm hair samples at enrollment (less than 16 weeks gestational age).

2.2 Hair sample collection

Trained research staff collected 9 cm hair samples from the posterior vertex region of the scalp as close to the scalp as possible. Hair samples were cut into three 3 cm segments with each 3cm segment corresponding to 3 months based on a rate of hair growth of approximately 1 cm/month [16]. For the current analysis, we used the hair samples collected at enrollment in first trimester. Hair samples collected at delivery were not included in this analysis given the temporal overlap between second hair samples and outcome of interest (preterm). The 3cm segment closest to the scalp (scalp-proximal) reflected cortisol levels during first trimester, and the next two 3cm segments (scalp-intermediate and scalp-distal) reflected cortisol levels during 0–3 months pre-conception and 3–6 months pre-conception, respectively. Collected hair samples were then wrapped in aluminum foil, and stored in manila envelopes away from light and at room temperature using desiccants.

2.3 Preterm birth (PTB)

We defined PTB according to the American College of Obstetricians and Gynecologists (ACOG) guidelines [17]. Gestational age was determined using the last menstrual period (LMP) and confirmed by ultrasound examination, conducted prior to 20 weeks of gestation. Using detailed information collected from medical records, we categorized PTB cases according to the three pathophysiological groups previously described (i.e. spontaneous preterm delivery, preterm premature rupture of membranes, and medically induced PTB) [18, 19]. Spontaneous preterm delivery cases were comprised of women whose medical records indicated a physician diagnosis of spontaneous labor onset (with intact fetal membranes) and delivery prior to the completion of 37 weeks’ gestation. Preterm premature rupture of membranes cases were comprised of women whose medical records indicated a physician diagnosis of rupture of fetal membranes (prior to the onset of labor) and delivery prior to the completion of 37 weeks’ gestation. Women who delivered prior to 37 completed weeks of gestation as a result of medical intervention were not eligible for this study.

2.4 Analytical population

The analytical population was selected from pregnant women who enrolled during the second wave of data collection in the PrOMIS study. Among the 2,068 PrOMIS participants, 647 women provided 9cm hair samples. Of the 647 participants, 20 participants were excluded due to missing or incomplete medical records, leaving 627 women. Among them, we identified and sampled all 44 women who delivered before 37 completed weeks of gestation as preterm cases (mean±standard deviation (SD) = 33.3±3.5). We then randomly sampled 100 women who delivered at term (≥37 weeks of gestation; mean±SD = 39.0±1.0) as term controls. The 100 participants selected for this analysis did not differ when compared to all participants (N = 2,068) or to those who provided hair samples, delivered at term, but were not sampled. Hair samples for a total of 144 participants (44 cases and 100 controls) were selected and submitted for laboratory analyses. For cases, 4 women had missing HCC data. For controls, 3 women had missing HCC data. Therefore, our final analytic sample consisted of 40 cases and 97 controls.

2.5 Laboratory analysis

Prior to lab shipment, selected hair samples were randomized (according to case control status) to mitigate bias resulting from batch effects. Hair samples were then shipped to Professor Kirschbaum’s laboratory in Dresden, Germany for analysis. Hair cortisol extraction procedures were similar to those described in detail elsewhere [20, 21]. Hair samples were first washed in 2.5 mL of isopropanol twice for three minutes each time, and then dried for 12 hours in an extractor hood. After drying, 7.5mg of whole non-pulverized hair was weighed out and then minced into small pieces. Hair samples were then incubated for 18 hours in 1.8ml high-grade methanol at room temperature, after which 1.6ml of clear supernatant was transferred to a vial and the methanol evaporated off at 55°C using a steady stream of nitrogen for thirty minutes. The residue from each segment was then resuspended in 225μl of distilled water, and 20μl of internal standard (cortisol-d4) was added. From the approximately 245μl resuspension per segment, 50μl were analyzed using the Cortisol Saliva Luminescence Immunoassay, IBL International ®, with one immunoassay batch defined as one 96-well plate. Resuspensions from all hair segments from the same woman were analyzed together in the same immunoassay batch, thereby reducing the influence of variability across batches. Cortisol units were reported in picograms per milligram (pg/mg). The average intra-assay coefficient of variation (CV) per immunoassay batch was 11.9% and the inter-assay was 19.4%. All laboratory procedures and analyses were performed without knowledge of case or control status.

2.6 Covariates

Trained interviewers collected information on maternal age (years), educational attainment (≤12 vs. >12 years), marital status (married or living with a partner vs. single), employment during pregnancy (yes vs. no), difficulty accessing basic foods (yes vs. no), difficulty paying for medical care (yes vs. no), smoking status during pregnancy (yes vs. no), and alcohol consumption during pregnancy (yes vs. no). Early pregnancy body mass index (BMI) (kg/m²) was calculated from the participants’ weight measured to the nearest 0.1kg and height measured with a telescopic height instrument to the nearest 0.1cm. Medical and reproductive history questions assessed parity (nulliparous vs. multiparous); gestational age at enrollment and delivery using last menstrual period (in weeks); infant sex (male vs. female), and seasons of delivery (December – February, March-May, June – August, or September – November). Measures of maternal perceived stress and symptoms of depression were assessed at enrollment using the 14-item Perceived Stress Scale [22], and the 9-item Patient Health Questionnaire [23, 24], respectively. Self-reported hair characteristics were assessed at enrollment and included: natural hair color (black vs. brown); hair structure (straight vs. curly); hair washing frequency (<6 times per week vs. 6–7 times per week); shampoo and conditioner use (shampoo only vs. shampoo and conditioner); chemical hair treatment use (tint, dye, or perm vs. none); and hair cutting frequency (≤3 months vs. >3 months).

2.7 Statistical analysis

We first explored the distribution of maternal socio-demographic characteristics, medical and reproductive histories according to preterm and term delivery status. We then used the Chi-square tests, for categorical variables, and Student’s t tests, for continuous variable, to assess differences in the distributions of covariates for cases and controls. Since the distribution of the HCC were skewed, the values were transformed to the natural logarithms for statistical analysis. For cases and controls, we reported the mean and standard deviations (SD) of HCC measures for the three hair segments, respectively. Between cases and controls, we calculated the unadjusted percentage differences and adjusted mean differences in HCC. In the adjusted models, confounders specified a priori, which were associated with HCC, were associated with preterm independent of HCC, and were not in the pathway from HCC to preterm, included maternal age (continuous), difficulty paying for medical care (yes vs. no), parity (nulliparous vs. multiparous) and gestational age at hair collection. To summarize the HCC across the entire 9 cm hair segment into one measure, we also reported the mean HCC across three 3 cm segments, corresponding to the mean HCC across pre-conception and first trimester. We fitted logistic regression models to derive odds ratios (ORs) and 95% confidence intervals (95% CIs) for preterm birth in relation to one-unit increase in HCC. In addition, we presented the HCC levels stratified by infant sex. Using the R gamm4 package [25], we fitted the generalized additive mixed models (GAMM) to examine HCC concentrations in relation to the estimated mean gestational age, after accounting for maternal age, difficulty paying for medical care, and parity. The GAMM incorporated repeated measures (i.e., three 3cm hair segments for each woman) of HCC across pre-conception and first trimester, and regressed these levels on a smooth function of gestational age. The GAMM allowed the inspection of potentially non-linear associations between HCC and gestational age, after accounting for within-subject correlation in the repeated measures. We plotted the predicted HCC (lines) from GAMM to present HCC trend over time (i.e., from the pre-conception to first trimester time frame). We conducted sensitivity analyses that incorporated the temporal variations of the three hair segments for each woman. We fitted two-stage mixed-effects models [26] to account for the longitudinal features of the HCC trajectory for each woman. In the model, the longitudinally time-varying exposure was first modeled as a function of time (e.g., using random slopes and intercepts) for each woman. At Stage 2, the best unbiased linear predictor estimates of subject-specific intercepts and slopes from the Stage 1 were simultaneously included as continuous variables in the logistic regression models, along with aforementioned confounders [26]. All analyses were performed using the R statistical software platform [27]. The level of statistical significance was set at p-value <0.05, and all tests were two-sided.

3 Results

The socio-demographic, medical and reproductive histories for cases and controls are shown in Table 1. Women who delivered preterm birth were older and were more likely to have difficulty paying for basic foods and medical care, as compared with controls who delivered at term. They were also more likely to use chemical hair treatment, and shampoo or conditioner when washing their hair. No statistically significant difference was seen for PHQ-9 and PSS scores between cases and controls. Maternal adjusted mean scalp HCC were 14% (p = 0.11), 10% (p = 0.22) and 14% (p = 0.08) lower for 3–6 months pre-conception, 0–3 months pre-conception, and first trimester, respectively, among preterm cases as compared with controls (Table 2). On average, for the entire pre-conception to first trimester period the adjusted mean HCC were 13% lower among women who delivered preterm as compared with those who delivered at term (p-value = 0.001). Notably, for mothers who delivered female newborns, the adjusted mean HCC differences between cases and controls were large (–0.17, –0.07, and –0.18 for 3–6 months pre-conception, 0–3 months pre-conception, and first trimester, respectively). Case-control differences in mean HCC concentrations for mothers who delivered male newborns were less pronounced (–0.09, –0.08, and –0.05 for 3–6 months pre-conception, 0–3 months pre-conception, and first trimester, respectively). Results from logistic regression analyses are presented in Table 3. Using the scalp-distal segments that represented HCC concentrations in 3–6 months pre-conception as the exposure, 1-unit increase in HCC was associated with 55% decreased odds of preterm birth (aOR = 0.45; 95% CI: 0.17–1.17) after controlling for other covariates. The odds ratios for preterm birth were 0.53 (95% CI: 0.19–1.48) and 0.39 (95% CI: 0.13–1.13), respectively, for 1-unit increase in HCC in the scalp-intermediate and scalp-distal segments representing HCC concentrations in 0–3 months pre-conception and first trimester. On average, a 1-unit increase in HCC in pre-conception and first trimester was associated with 60% (OR = 0.40; 95% CI: 0.13–1.19) reduced odds of preterm birth. When stratified by newborn sex, the associations between HCC and preterm birth were more evident in women who delivered female newborns as compared with those who delivered male newborns (Table 3 and Fig. 1) When accounting for the longitudinal features of the HCC trajectory for each woman, we observed similar magnitudes of associations from two-stage mixed-effects models (Supplement Tables 1-3) as compared to the estimates from the logistic regressions using the average HCC across pre-conception and first-trimester, confirming the robustness of our findings. During pre-conception and first-trimester, the predicted HCC values from GAMM were consistently lower among preterm cases as compared with controls.

Table 1
Characteristics of study population according to preterm case and control status, Lima, Peru

Characteristics Term controls (N = 97) Preterm cases (N = 40) P-value

n % n %

Age (years) ^* 26.4 (5.8) 29.5 (7.8) 0.01

Age (years)

<35 86 88.7 27 67.5 0.007

≥35 11 11.3 13 32.5

≤12 years of education 37 38.1 11 27.5 0.43

Married/living with a partner 77 79.4 31 77.5 0.99

Employed 46 47.4 12 30.0 0.09

Difficulty paying for basics 39 40.2 25 62.5 0.03

Difficulty paying for medical care 26 26.8 26 65.0 <0.001

Early pregnancy measured BMI (kg/m²)

<25 48 49.5 20 50.0 0.12

25–29.9 40 41.2 12 30.0

>30 8 8.2 8 20.0

Smoking during pregnancy 2 2.1 1 2.5 1.00

Alcohol consumption during pregnancy 3 3.1 3 7.5 0.49

Nulliparous 44 45.4 16 40.0 0.70

Gestational age at hair collection (weeks)^* 13.0 (4.0) 12.1 (3.5) 0.19

Gestational weeks at delivery 39.0 (1.0) 33.3 (3.5) <0.001

Mode of delivery

Vaginal 45 46.4 11 27.5 0.041

C-section 52 53.6 29 72.5

Male infant 55 56.7 23 57.5 1.00

Season of delivery

December – February 1 1.0 2 5.0 0.45

March – May 16 16.5 8 20.0

June – August 44 45.4 18 45.0

September – November 36 37.1 12 30.0

Black hair 55 56.7 26 65.0 0.04

Straight hair 66 68.0 27 67.5 0.08

Hair wash <6 times per week 81 83.6 33 82.5 0.08

Use of shampoo or conditioner when washing

Shampoo only 33 34.0 6 15.0 0.01

Shampoo and conditioner 64 66.0 32 80.0

Chemical hair treatment use (tint, dye, or wave) 37 38.1 10 25.0 0.04

Hair cutting frequency: ≤3 months 41 42.3 13 32.5 0.06

Childhood abuse 29 29.9 13 32.5 0.92

Patient Health Questionanre-9 (PHQ-9)^* 6.4 (4.2) 6.6 (3.8) 0.77

Perceived Stress (PSS)^* 29.0 (4.9) 29.8 (4.4) 0.35

Characteristics	Term controls (N = 97)	Preterm cases (N = 40)	P-value
Age (years) ^*	26.4 (5.8)		29.5	(7.8)	0.01
Age (years)
<35	86	88.7	27	67.5	0.007
≥35	11	11.3	13	32.5
≤12 years of education	37	38.1	11	27.5	0.43
Married/living with a partner	77	79.4	31	77.5	0.99
Employed	46	47.4	12	30.0	0.09
Difficulty paying for basics	39	40.2	25	62.5	0.03
Difficulty paying for medical care	26	26.8	26	65.0	<0.001
Early pregnancy measured BMI (kg/m²)
<25	48	49.5	20	50.0	0.12
25–29.9	40	41.2	12	30.0
>30	8	8.2	8	20.0
Smoking during pregnancy	2	2.1	1	2.5	1.00
Alcohol consumption during pregnancy	3	3.1	3	7.5	0.49
Nulliparous	44	45.4	16	40.0	0.70
Gestational age at hair collection (weeks)^*	13.0 (4.0)		12.1	(3.5)	0.19
Gestational weeks at delivery	39.0 (1.0)		33.3	(3.5)	<0.001
Mode of delivery
Vaginal	45	46.4	11	27.5	0.041
C-section	52	53.6	29	72.5
Male infant	55	56.7	23	57.5	1.00
Season of delivery
December – February	1	1.0	2	5.0	0.45
March – May	16	16.5	8	20.0
June – August	44	45.4	18	45.0
September – November	36	37.1	12	30.0
Black hair	55	56.7	26	65.0	0.04
Straight hair	66	68.0	27	67.5	0.08
Hair wash <6 times per week	81	83.6	33	82.5	0.08
Use of shampoo or conditioner when washing
Shampoo only	33	34.0	6	15.0	0.01
Shampoo and conditioner	64	66.0	32	80.0
Chemical hair treatment use (tint, dye, or wave)	37	38.1	10	25.0	0.04
Hair cutting frequency: ≤3 months	41	42.3	13	32.5	0.06
Childhood abuse	29	29.9	13	32.5	0.92
Patient Health Questionanre-9 (PHQ-9)^*	6.4 (4.2)		6.6	(3.8)	0.77
Perceived Stress (PSS)^*	29.0 (4.9)		29.8	(4.4)	0.35

^*mean (standard deviations) Abbreviations: PHQ-9, Patient Health Questionnaire-9; PSS, Perceived Stress Scale For continuous variable, P-value was calculated using the Student’s t test; for categorical variable, P-value calculated using the chi-square test.

Table 2

Pre-conception and pregnancy hair cortisol concentration (pg/mg) according to preterm birth status

Time window for HCC	Term controls (N = 97)		Preterm cases (N = 40)		Unadjusted % difference	Adjusted mean Log(HCC) difference^*	Adjusted mean Log(HCC) difference^**
	Mean Log(HCC)	SD	Mean Log(HCC)	SD
Hair sample segments
Scalp-distal: (3–6 mo pre-conception)	0.40	0.44	0.32	0.45	–25.00	-0.14(-0.32, 0.03)	-0.14( -0.32, 0.03)
Scalp-intermediate: (0–3 mo pre-conception)	0.56	0.43	0.53	0.39	–5.66	-0.10( -0.27, 0.06)	-0.10 (-0.26, 0.07)
Scalp-proximal: (first trimester)	0.71	0.41	0.63	0.36	–12.70	-0.14 (-0.30, 0.02)	-0.13 (-0.29, 0.03)
Average (pre-conception and first trimester)	0.56	0.44	0.49	0.42	–14.29	-0.13 (-0.23, -0.03)	-0.13 (-0.23, -0.03)

Abbreviations: SD, standard deviations ^*Adjusted for maternal age (continuous), difficulty paying for medical care (yes vs. no), and parity (nulliparous vs. multiparous) ^**Further adjusted for gestational age at hair collection (continuous).

Table 3

Unadjusted and adjusted odds ratios of preterm birth in relation to one-unit increase in In transformed pre-conception and trimester-specific hair cortisol concentration (pg/mg)

Time window for HCC	Unadjusted model	Adjusted model	Adjusted model
	OR (95% CI)	OR (95% CI)^*	OR (95% CI)^**
Hair sample segments
Scalp-distal: (3–6 mo pre-conception)	0.69 (0.30, 1.59)	0.45 (0.17, 1.15)	0.45 (0.17, 1.17)
Scalp-intermediate: (0–3 mo pre-conception)	0.84 (0.35, 2.05)	0.50 (0.18, 1.38)	0.53 (0.19, 1.48)
Scalp-proximal: (first trimester)	0.59 (0.23, 1.51)	0.37 (0.13, 1.05)	0.39 (0.13, 1.13)
Average (pre-conception and first trimester)	0.66 (0.25, 1.72)	0.38 (0.13, 1.12)	0.40 (0.13, 1.19)

Abbreviations: OR, odds ratio; CI, confidence interval ^*Adjusted for maternal age (continuous), difficulty paying for medical care (yes vs. no), and parity (nulliparous vs. multiparous) ^**Further adjusted for gestational age at hair collection (continuous).

4 Discussion

Overall, we found that maternal HCC were 14% (p = 0.11), 10% (p = 0.22) and 14% (p = 0.08) lower for 3–6 months pre-conception, 0–3 months pre-conception, and first trimester, among preterm cases as compared with controls. In multivariable-adjusted models, a 1-unit increase in HCC was associated with 55% reduced odds of preterm birth (OR = 0.45; 95% CI: 0.17, 1.17) using the scalp-distal segments that represented HCC concentrations in 3–6 months pre-conception as the exposure although statistical significance was not reached. The corresponding odds for 1-unit increase in HCC in the scalp-intermediate and scalp-distal segments representing HCC concentrations in 0–3 months pre-conception (OR = 0.53; 95% CI: 0.19–1.48) and first trimester (OR = 0.39; 95% CI: 0.13–1.13), respectively. Findings from previous studies of the relationship between maternal cortisol concentrations and PTB have been inconsistent with several studies documenting increased cortisol concentrations among women with PTB compared to those who deliver at term [28 –30]. For instance, in a study of 300 depressed pregnant women around 20 weeks of gestation, Field et al. found that high urine cortisol concentration was significantly associated with preterm birth (p-value <0.01) [28]. Campbell et al. in their study of 218 pregnant women admitted to hospital with a diagnosis of threatened preterm labor found that higher cortisol concentrations at 28–36 weeks gestation might be predictive preterm labor [29]: every 0.1-unit increase in cortisol was associated with more than 1.20-fold increased odds of giving birth within 48 hours. Similarly, Erickson et al. in their study of gestational age-matched 59 PTB cases and 300 term controls found that plasma cortisol concentration of PTB cases were higher compared to controls [30]. Other investigators have found no associations. For instance, in their observational study of 78 women in a private practice in central Texas, Ruiz et al. found no statistically significant association between maternal plasma cortisol and PTB [31]. Compared to non-pregnant women, maternal cortisol concentration increases 2–4-fold over the course of uncomplicated pregnancy. Prior studies are limited by several methodological concerns including the use of serum, plasma, urine or salivary cortisol measurements to quantify cortisol concentration, each reflecting different physiologically relevant time windows. For example, urine commonly reflects cortisol levels in the past 12–24 hours, while plasma, serum, and saliva reflect more immediate cortisol levels [32]. However, HCC measures long-term systemic cortisol concentrations reflecting the past several months of HPA activity [33]. Although HCC is somewhat vulnerable to several environmental factors (e.g. hair treatments), it outweighs previous methods of assessment (in urine, saliva, or plasma) in terms of the stability of the measure over time, invasiveness and resource demands. To date, we are aware of only three published studies that examined the relationship between HCC and PTB. In their study of 117 women from Montreal, Canada (31 PTB cases and 86 controls), Kramer et al. found that maternal HCC obtained at delivery were somewhat lower among women who had spontaneous PTB (171.7 ng/g) compared to term controls (190.6 ng/g), although not statistically significant. However, the authors did not collect sufficient hair to allow for trimester-specific analysis [34]. Recently, in a study of 29 PTB cases and 29 term controls, Karakash et al. [21] reported that each 10 pg/mg increase in hair cortisol level at delivery was associated with a 33% decreased odds of being a PTB case. In other words, low scalp HCC (i.e., low HCC) were associated with increased odds of PTB. Another study from Duffy et al. [14] reported that HCC was lower for mothers with preterm birth. Cortisol differed significantly in the third trimester between mothers delivering term (mean = 23.80 pg/mg) and preterm (mean = 8.59 pg/mg) and trended toward significance in the second trimester (term: mean = 11.26 pg/mg; preterm: mean = 8.71 pg/mg). Duffy et al. [14] suggested that there has been a blunted, flattened pattern of changes in HCC across gestation in women who delivered preterm birth. In our study, we observed that 1-unit increase in log HCC in first trimester was associated with an 61% decreased odds of PTB. In sum, our findings and those reported by previous investigators [14 , 34] suggest a possible suppression of HPA axis activation, as reflected by a sustained reduced synthesis and release of cortisol in hair. On balance, these findings, now from independent study population, suggests that women who are chronically stressed, but who are unable to mount an adequate stress response (as reflected cortisol synthesis and release) may be at increased risk for PTB. Hypocortisolism is an important factor in the pathophysiology of several disease states and clinical syndromes such as chronic fatigue syndrome, burnout, chronic pelvic pain, fibromyalgia, posttraumatic stress disorder, autoimmunity, irritable bowel syndrome, allergies, inflammation, and atypical depression [21 , 35–37]. Allostatic load, the dysregulation of biological systems with prolonged or poorly regulated allostatic systems (e.g. the HPA axis), has been theorized to play a role in hypocortisolism [35, 38]. Allostatic load results when the allostatic systems are overworked or fail to shut-off after exposure to stressful events [35, 38]. As such, blunted chronic cortisol level, such as observed in the PTB group, may be a response to long-standing chronic stress that result in allostatic load and dysfunction of the normal HPA axis [21]. Taken together, available data suggest that hypocortisolism might play an important role in the pathogenesis of PTB. Some limitations of our study merit discussion and consideration. First, the small number of cases (n = 40) of PTB may have hindered our ability to detect statistically significant associations. Second, distal hair segments may not be a reliable estimate of cortisol concentrations. Therefore, estimates based on distal hair segments should be interpreted with caution. Lastly, although we controlled for important confounding factors, we cannot exclude the possibility that the odds ratios reported are unaffected by residual confounding such as antenatal corticosteroids and other obstetric medical problems. Our results suggest that women who deliver preterm, as compared with those who deliver at term, may have a lower preconception and first trimester HCC. The findings suggest that HPA axis activation, integral to the adaptive stress-response system, may be dysregulated in women at increased risk of PTB. Future research that includes larger study population and replication findings will be needed to confirm our observations and determine the role of HPA axis activity among pregnant women and its relation with hypocortisolism and subsequent risk of PTB.

Author contributions

All authors substantially contributed to this work.

Financial support

This research was supported by an award from the National Institutes of Health (NIH), the Eunice Kennedy Shriver Institute of Child Health and Human Development (R01-HD-059835). The NIH had no further role in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; and in the decision to submit the paper for publication.

Conflict of Interest

The authors have no conflict of interest to report.

Footnotes

Acknowledgments

The authors are indebted to the participants of the PrOMIS study for their cooperation. They are also grateful to Ms. Elena Sanchez and the dedicated staff members of Asociacion Civil Proyectos en Salud (PROESA), Peru for their expert technical assistance with this research.

The supplementary material is available in the electronic version of this article: .

References

Centers for Disease Control and Prevention, 2018. Preterm Birth | Maternal and Infant Health | Reproductive Health | CDC [WWW Document]. URL https://www.cdc.gov/reproductivehealth/maternalinfanthealth/pretermbirth.htm (accessed 7.10.19).

Romero

, Dey

, Fisher

. Preterm labor: One syndrome, many causes. Science. 2014;345(6198):760–5.

Goldenberg

, Culhane

, Iams

, Romero

. Epidemiology and causes of preterm birth. Lancet. 2008;371(9606):75–84.

Fawke

, Lum

, Kirkby

, Hennessy

, Marlow

, Rowell

, et al. Lung function and respiratory symptoms at 11 years in children born extremely preterm: The EPICure study. Am J Respir Crit Care Med. 2010;182(2):237–45.

Kajantie

, Osmond

, Barker

, Eriksson

. Preterm birth–a risk factor for type 2 diabetes? The Helsinki birth cohort study. Diabetes Care. 2010;33(12):2623–5.

Smith

, Sterne

, Tynelius

, Lawlor

, Rasmussen

. Birth weight of offspring and subsequent cardiovascular mortality of the parents. Epidemiology. 2005;16(4):563–9.

Robbins

, Hutchings

, Dietz

, Kuklina

, Callaghan

. History of preterm birth and subsequent cardiovascular disease: A systematic review. Am J Obset Gynecol. 2013;210(4):285–97.

Catov

, Wu

, Olsen

, Sutton-Tyrrell

, Li

, Nohr

. Early or recurrent preterm birth and maternal cardiovascular disease risk. Ann Epidemiol. 2010;20(8):604–9.

March of Dimes, PMNCH, Save the Children, WHO: Born too soon: The global action report on preterm birth. In. Edited by Howson CP, KinneyMV, Lawn JE. Geneva; 2012, pp. 1-126.

10.

Frey

, Klebanoff

. The epidemiology, etiology, and costs of preterm birth. Semin Fetal Neonatal Med. 2016;21(2):68–73.

11.

Sanchez

, Alva

, Diez Chang

, Qiu

, Yanez

, Gelaye

, et al. Risk of spontaneous preterm birth in relation to maternal exposure to intimate partner violence during pregnancy in Peru. Matern Child Health J. 2013;17(3):485–92.

12.

Wosu

, Gelaye

, Williams

. Maternal history of childhood sexual abuse and preterm birth: An epidemiologic review. BMC Pregnancy Childbirth. 2015;15:174.

13.

Meyer

, Novak

. Minireview: Hair cortisol: A novel biomarker of hypothalamic-pituitary-adrenocortical activity. Endocrinology. 2012;153(9):4120–7.

14.

Duffy

, Schminkey

, Groer

, Shelton

, Dutra

. Comparison of hair cortisol levels and perceived stress in mothers who deliver at preterm and term. Biol Res Nurs. 2018;20(3):292–9.

15.

Hinkelmann

, Muhtz

, Dettenborn

, Agorastos

, Wingenfeld

, Spitzer

, et al. Association between childhood trauma and low hair cortisol in depressed patients and healthy control subjects. Biol Psychiatry. 2013;74(9):e15–7.

16.

Kirschbaum

, Tietze

, Skoluda

, Dettenborn

. Hair as a retrospective calendar of cortisol production-increased cortisol incorporation into hair in the third trimester of pregnancy. Psychoneuroendocrinology. 2009;34(1):32–7.

17.

ACOG technical bulletin. Preterm labor. Number 206–June 1995 (Replaces No. 133, October 1989). Int J Gynaecol Obstet. 1995;50(3):303–13.

18.

Williams

, Mittendorf

, Stubblefield

, Lieberman

, Schoenbaum

, Monson

. Cigarettes, coffee, and preterm premature rupture of the membranes. Am J Epidemiol. 1992;135(8):895–903.

19.

Savitz

, Blackmore

, Thorp

. Epidemiologic characteristics of preterm delivery. Etiologic heterogeneity. Am J Obstet Gynecol. 1991;164(2):467–71.

20.

Albar

, Russell

, Koren

, Rieder

, van Umm

. Human hair cortisol analysis: Comparison of the internationally-reported ELISA methods. Clin Invest Med. 2013;36(6):E312–6.

21.

Karakash

, Tschankoshvili

, Weedon

, Schwartz

, Kirschbaum

, Minkoff

. Hypocortisolism and preterm birth. J Neonatal Perinatal Med. 2016;9(4):333–9.

22.

Cohen

, Kamarck

, Mermelstein

. A global measure of perceived stress. J Health Soc Behav. 1983;24(4):385–96.

23.

Kroenke

, Spitzer

, Williams

. The PHQ-9: Validity of a brief depression severity measure. J Gen Intern Med. 2001;16(9):606–13.

24.

Zhong

, Gelaye

, Fann

, Sanchez

, Williams

. Cross-cultural validity of the Spanish version of PHQ-9 among pregnant Peruvian women: A Rasch item response theory analysis. J Affect Disord. 2014;158:148–53.

25.

Wood

. Scheipl F: Gamm4. Generalized additive mixed models using mgcv and lme4. 2013.

26.

Chen

, Ferguson

, Meeker

, McElrath

, Mukherjee

. Statistical methods for modeling repeated measures of maternal environmental exposure biomarkers during pregnancy in association with preterm birth. Environ Health. 2015;14:9.

27.

R Core Team. Others: R foundation for statistical computing. Vienna, Austria. 2013;3(0).

28.

Field

, Hernandez-Reif

, Diego

, Figueiredo

, Schanberg

, Kuhn

. Prenatal cortisol, prematurity and low birthweight. Infant Behav Dev. 2006;29(2):268–75.

29.

Campbell

, Challis

, DaSilva

, Bocking

. A cohort study found that white blood cell count and endocrine markers predicted preterm birth in symptomatic women. J Clin Epidemiol. 2005;58(3):304–10.

30.

Erickson

, Thorsen

, Chrousos

, Grigoriadis

, Khongsaly

, McGregor

, et al. Preterm birth: Associated neuroendocrine, medical, and behavioral risk factors. J Clin Endocrinol Metab. 2001;86(6):2544–52.

31.

Ruiz

, Fullerton

, Dudley

. The interrelationship of maternal stress, endocrine factors and inflammation on gestational length. Obstet Gynecol Surv. 2003;58(6):415–28.

32.

Russell

, Koren

, Rieder

, Van Uum

. Hair cortisol as a biological marker of chronic stress: Current status, future directions and unanswered questions. Psychoneuroendocrinology. 2012;37(5):589–601.

33.

Sauve

, Koren

, Walsh

, Tokmakejian

, Van Uum

. Measurement of cortisol in human hair as a biomarker of systemic exposure. Clin Invest Med. 2007;30(5):E183–91.

34.

Kramer

, Lydon

, Seguin

, Goulet

, Kahn

, McNamara

, et al. Stress pathways to spontaneous preterm birth: The role of stressors, psychological distress, and stress hormones. Am J Epidemiol. 2009;169(11):1319–26.

35.

Fries

, Hesse

, Hellhammer

. A new view on hypocortisolism. Psychoneuroendocrinology. 2005;30(10):1010–6.

36.

Griep

, Boersma

, Lentjes

, Prins

, van der Korst

, de Kloet

. Function of the hypothalamic-pituitary-adrenal axis in patients with fibromyalgia and low back pain. J Rheumatol. 1998;25(7):1374–81.

37.

Edwards

, Heyman

, Swidan

. Hypocortisolism: An evidence-based review. Integrative Medicine. 2011;10(4):26–33.

38.

McEwen

, Stellar

. Stress and the individual. Mechanisms leading to disease. Arch Intern Med. 1993;153(18):2093–101.