Maternal corticosteroids and depression during gestation and decreased fetal heart rate variability

Abstract

Objective

Maternal depression during gestation is an adverse factor in fetal brain development that manifests in later childhood behavioral problems. Fetal heart rate variability (FHRV) mediated by parasympathetic input is a marker of gestational nervous system development. Biological mediators of adverse effects of maternal depression may involve the mother’s corticosteroids; however, links between depression, corticosteroids, and early nervous system development remain inconclusive.

Methods

Heart rate was recorded in 23 fetuses by transabdominal Doppler at 28–33 weeks gestation. The SD of interbeat intervals over 20 min assessed FHRV. Maternal depression ratings and hair concentrations of cortisol and cortisone were assayed. An auditory sensory gating paradigm assessed newborn development of cerebral inhibition. Parents rated their infant’s temperament characteristics on the Infant Behavior Questionnaire-Revised Short Form (IBQ-R).

Results

Maternal depression was associated with lower FHRV, especially for male fetuses, β = −0.633, P = 0.045. Maternal depression was associated with lower cortisol to total corticosteroids ratios, β = −0.519, P = 0.033. Lower cortisol ratios were associated with decreased FHRV, β = 0.485, P = 0.019. Decreased FHRV was associated with increased newborn sensory gating deficits, β = −0.992, P = 0.035, indicating poorer development of cerebral inhibition. Higher FHRV was related to increased infant IBQ-R self-regulatory behaviors, r = 0.454, P = 0.029.

Conclusion

Maternal depression is associated via corticosteroids with decreased development of nervous system control of fetal heart rate. Decreased FHRV indicates developmental alterations in gestation that correlate with altered brain function and subsequent regulatory challenges in early infancy.

Introduction

Maternal stress and depression during gestation affect fetal brain development, with long-lasting effects on offspring behavior and temperament [1–3]. Both inflammation and cortisol secretion in response to maternal stress and depression have been proposed as mechanisms, with evidence for increased effects of inflammation in males and cortisol in females during early gestation [4]. By the beginning of the third trimester, fetal heart rate (FHR) and variability (FHRV) are markers of the developing central nervous system’s increasing control of the heart. FHRV, in particular, has been characterized as indicating parasympathetic nervous system influence. Limited research has evaluated associations between maternal prenatal stress, cortisol secretion, and FHRV. Decreased FHRV is associated with poorer neonatal and later infant neurodevelopmental outcomes [5–8].

This study aimed to characterize FHRV in relation to maternal depression [9], specifically at the beginning of the third trimester, when parasympathetic influence on the heart is rapidly increasing. In early second trimester, during the initial period of brain development, the male fetus is more protected from maternal cortisol due to higher levels of 11-β-hydroxysteroid dehydrogenase-2 (11BHSD-2) in the placenta and fetal tissues [10]. 11BHSD-2 irreversibly metabolizes cortisol to cortisone, which does not activate cortisol receptors. 11BHSD-1, which is a reversible enzyme, is not present in the fetus at this stage. However, beginning at 26 weeks and until term, 11BHSD-1 activity increases in the fetus, which metabolizes cortisone back to cortisol [11]. Rising progesterone levels at the same time decrease the expression of 11BHSD-2 [12]. 11BHSD-2 is methylated in placentas of stressed mothers, which further decreases its expression [13]. The ratio of cortisol to cortisol + cortisone can be used to reflect overall 11BHSD activity and represent the level of corticosteroids reaching the fetus.

We hypothesized that women and men would have different responses to maternal prenatal depression, but while sex differences in FHRV have been reported in some studies, variations in methodologies have resulted in conflicting outcomes [14]. Second, we hypothesized that cortisol levels in the mother that are associated with depression, and its metabolite cortisone produced by the placenta, would have roles in the effects of maternal depression on FHRV. Third, we hypothesized that the developmental effects of maternal depression assessed by FHRV would manifest after birth in the development of cerebral inhibition in newborns as well as temperament characteristics in infancy.

Methods

Subjects

Women were enrolled from a public safety-net prenatal clinic at 14–16 weeks gestation. Women were asked to participate in a prospective study of stress in pregnancy on their child’s development. Gestational age was established by first ultrasound to corroborate the reported last menstrual period. Exclusions were fetal anomaly and maternal medical morbidity. The Colorado Multiple Institution Review Board approved the study. All mothers, and fathers if available, provided informed consent. Maternal sociodemographics and health, including psychiatric diagnoses, substance use, BMI, and prenatal vitamin use, were assessed. Self-ratings on the Center for Epidemiological Studies of Depression (CESD) Scale were acquired in the same session as fetal heart rate recording. The CESD is the most sensitive instrument for depression for women from underserved communities [15]. Maternal CESD ratings ≥16 are consistent with a diagnosis of major depressive disorder.

Fetal heart rate recording

Fetal heart rate was collected via a single wide-array Doppler transducer (Wallach Fetal2EMR Fetal Monitor, CooperSurgical, Inc.) positioned on the mother’s abdomen while she rested in a semi-recumbent position. The Fetal2EMR has an accuracy rate of ±1 bpm. Following a 10-min quiet period, the fetal heart rate was recorded for 5 min before presenting a 10-s vibroacoustic stimulation. Poststimulation recording then continued for an additional 15 min. Heart rate data collected from the monitor’s output port were digitized and analyzed offline. Heart rate records were visually inspected for signal loss. Heart rate values exceeding a maximum fluctuation of 5 bpm from the average of the previous five values were rejected. Heart rate values were converted to interbeat intervals for analysis. Mean interbeat intervals and SDs were computed in 1-min epochs and averaged across the entire recording. FHRV was indexed by the SD of interbeat intervals.

Hair cortisol and cortisone assays

Maternal hair (2.5 cm) cut from the posterior vertex close to the scalp integrates cortisol and cortisone levels from the past 10–12 weeks of gestation [16] (see Detailed Methods, Supplemental Digital Content 1, http://links.lww.com/WNR/A646 which describes the mass spectroscopy assay). Correlation between maternal hair cortisol levels and the area under the curve for salivary cortisol, sampled 30 min before waking, before lunch, and 10 h after waking for 3 days in the second trimester is r² = 0.34, P = −0.04 [17]. Hair samples were also obtained from neonates and consist of hair formed during the third trimester. The ratio of cortisol to cortisol + cortisone can be used to represent the fraction of active cortisol in the context of cortisone, as it reflects total 11-BHSD activity.

Electrophysiological recording of inhibition of cerebral P50 auditory evoked potentials

Newborns were recorded during active sleep at 44 weeks gestational age. Two identical auditory stimuli 500 ms apart elicit P50_S1 and P50_S2 evoked responses. Vertex electroencephalographic activity is recorded from 100 ms before to 200 ms after S1 or S2. A minimum of 75 artifact-free recordings are digitally band filtered (10–50 Hz, −6 dB) to isolate the P50 component. Lower P50_S2μV amplitudes adjusted for P50_S1μV indicate increased inhibition. Decreased inhibition of P50_S2μV has been found previously in newborns whose mothers were depressed during gestation [18] (see Detailed Methods, Supplemental Digital Content 1, http://links.lww.com/WNR/A646, which describes the electrophysiological recording of auditory P50 inhibitory sensory gating).

Infant behavior

Mothers completed the Infant Behavior Questionnaire-Revised Short Form (IBQ-R) when the infant was 3 months of age. A validated Spanish version was offered to primary Spanish-speaking women. The 91-items are used to rate 14 aspects of child behavior, which the IBQ-R developers clustered into three indices by factor analysis. Surgency/extraversion summarizes the infant’s level of activity and positive affect. Negative affectivity reflects infant negative emotionality, including fearfulness, frustration, and sadness. Orienting/regulation encompasses emerging capacities for infant regulation (e.g. duration of attentional orienting, recovery from distress, and soothability). Inter-parent agreement is adequate for orienting/regulation (Cronbach’s α = 0.71–0.75) across a range of socioeconomic groups [19].

Statistical analyses

Generalized linear models and multiple regression analyses were used (see Detailed Statistical Analyses, Supplemental Digital Content 1, http://links.lww.com/WNR/A646 which contains the complete statistical models). Gestational age at examinations and fetal sex were covariates for all analyses. Other potential covariates were screened for differences between pregnancies with male and female fetuses and association with FHRV. None were found (see Detailed Methods, Supplemental Digital Content 1, http://links.lww.com/WNR/A646 which contains a Table of demographic and other differences).

Results

Eight of the 23 women (35%) had CESD rating ≥16 at the time of the FHR recording. In comparison, 36% of childbearing age women in a survey of Los Angeles County had CESD ratings ≥16 [20]. Maternal CESD ratings at 28–33 weeks gestation were associated with FHRV, particularly in male fetuses, Fetal sex*CESD Wald χ²df1 = 5.341, P = 0.021. For males, higher CESD ratings were associated with decreased FHRV, β = −0.633, P = 0.045, Fig. 1.

Fig. 1.

Effect of maternal center for epidemiological studies of depression ratings on fetal heart rate variability at 29–33 weeks gestation. Effects in male fetuses are significant, β = −0.633, P = 0.045.

Both maternal cortisol and cortisone affected FHRV. Their effects were summarized in a single variable, the ratio of cortisol to cortisol + cortisone (Supplemental Table, Supplemental Digital Content 1, http://links.lww.com/WNR/A646). CESD ratings were negatively associated with cortisol/cortisol + cortisone ratios in maternal hair, β = −0.519, P = 0.033, Fig. 2. Black women in this cohort have higher maternal cortisol concentrations than other women [21]. Two women self-identified as Black and contributed to a modest effect of race in this analysis, β = 0.547, P = 0.074.

Fig. 2.

Center for epidemiological studies of depression ratings at 28–33 weeks associated with lower ratio of cortisol to cortisol + cortisone in maternal hair, β = −0.519, P = 0.033.

Lower cortisol ratios were associated with decreased FHRV, β = 0.485, P = 0.019 for both sexes. When cortisol ratios were considered together with CESD ratings for male fetuses, the effect of CESD became nonsignificant, β = −0.385, P = 0.116, while the cortisol ratio remained nearly significant, β = 0.408, P = 0.052.

For males, FHRV was associated with inhibition of the P50 auditory evoked potential to repeated stimuli. Increased FHRV was associated with decreased inhibition of the second P50 response, indicated by an increased amplitude in the newborn P50_S2μV adjusted for P50_S1μV, β = −0.992, P = 0.035. CESD was not a significant factor when considered with FHRV.

For both sexes, FHRV was positively associated with Regulation Index IBQ-R ratings, r = 0.454, P = 0.029, Fig. 3. The two Regulation components most affected were falling reactivity, r = 0.458, P = 0.028, and soothability, r = 0.433, P = 0.035.

Fig. 3.

Fetal heart rate variability at 28–33 weeks gestation is associated with ratings on the Regulation Index of the Infant Behavioral Questionnaire-R(SF) at 3 months of age, r = 0.454, P = 0.029.

Discussion

This study demonstrates that maternal depression is an adverse factor for the gestational development of FHRV, particularly in males. However, when cortisol and cortisone were added to the model, the relationship between FHRV and maternal depression was nonsignificant, suggesting the effects of maternal depression on FHRV are mediated by cortisol and cortisone. As a marker of prenatal nervous system development, lower FHRV was associated with poorer postnatal brain function, as assessed by poorer P50 inhibitory sensory gating and poorer IBQ-R Regulation. Figure 4 summarizes these associations.

Fig. 4.

Summary of associations between fetal heart rate variability and maternal depression, cortisol ratios, P50 inhibition, and infant regulation. Dashed lines represent strength of association when maternal depression and cortisol ratios were considered together. *P = 0.052;**P < 0.05; +P = 0.116.

Neonatal hair cortisol levels are derived principally from cortisone [22]. Higher cortisone levels in pregnancies with male fetuses, related to the higher levels of 11BHSD-2 in the placenta, were associated with protection from maternal cortisol earlier in gestation in an earlier study [4]. In the current study, neonatal hair levels of cortisone and cortisol, representing growth occurring during the last 10 weeks in utero, were correlated in males r = 0.925, P < 0.001, but less so in females, r = 0.437, P = 0.28. These higher male cortisone levels, reflected in lower cortisol to cortisol + cortisone ratios, are now a source of fetal cortisol. At the beginning of the third trimester, for both male and female fetuses, these corticosteroids appear to be impacting the development of FHRV.

The results of this study should be interpreted in the context of the small sample size. Additional research with a larger, more diverse sample is needed.

Decreased FHRV has been associated with both maternal depression and poor neonatal adaptive behavior [5–8]. This study extends these findings to a specific brain function, the development of cerebral inhibition. Developing neurons are sensitive to cortisol [4]. A source of P50 auditory evoked potentials is in the hippocampus, where inhibition in the sensory gating paradigm is associated with the activity of inhibitory interneurons [23]. Similarly, FHRV is reliant on functioning inhibitory neurocircuits indicative of parasympathetic nervous system influences [24]. Deficiencies in newborn P50 inhibition are associated with subsequent problems in IBQ-R Regulation [4]. Poor infant regulation/orienting predicts poorer effortful control during childhood, which renders risk for adverse outcomes, such as psychopathology and academic difficulties, including reading readiness and conscientiousness [19,25]. Thus, FHRV recorded at the onset of the third trimester may be the first signal of significant developmental issues to be addressed while gestation and fetal development are still ongoing.