Association Between Mother-Infant Hair Cortisol Levels,Perinatal Stressors,and Adverse Birth Outcomes Among Kenyan Mother-Infant Pairs

Abstract

Background:

Few data on maternal-infant hair cortisol are available from HIV high-burden settings. We evaluated maternal hair cortisol and its relationship with infant hair cortisol, perinatal stressors, and pregnancy outcomes in Kenya.

Materials and Methods:

We analyzed data from an HIV prevention trial (NCT03070600). Pregnant women without HIV were enrolled and followed through 9 months postpartum. We utilized hair samples collected at 6 weeks postpartum from mother-infant pairs to quantify hair cortisol concentrations (HCC) reflecting cortisol levels over the prior ∼3 months. HCC was measured by enzyme immunoassay. We tested whether mean maternal HCC was associated with infant HCC, stressors, and/or birth outcomes using linear regression.

Results:

Overall, 153 women had HCC available at 6 weeks postpartum. Median maternal age was 25 years (Interquartile range [IQR] 22–30), median completed education was 12 years (IQR: 8–12), and 95% were married. Median birthweight was 3.4 kilograms (IQR: 3.1–3.7), 13% of women experienced preterm birth (PTB), and 2% of infants had low birthweight (LBW); 11% were small-for-gestational age (SGA). Median maternal HCC was 12.0 pg/mg (IQR: 6.3–20.4); and median infant HCC was 10.4 pg/mg (IQR: 5.5–23.7). Infant HCC was correlated with maternal HCC (r = 0.498, p < 0.001). Stressors associated with maternal HCC (p < 0.05) included having a partner living with HIV or of unknown HIV status, relationship instability, and having high behavioral risk for HIV. Maternal HCC was not significantly associated with PTB, LBW, or SGA.

Conclusion:

In conclusion, peripartum maternal HCC was correlated with infant HCC and associated with experiencing HIV acquisition-related stressors.

Introduction

Exposure to stressors during the prenatal period is linked to adverse birth outcomes such as preterm birth (PTB) and low infant birthweight (LBW),¹ as well as neurodevelopmental impairments into childhood.^2,3 Cisgender women experiencing high stress are 1.5 to 3 times more likely to have PTB than less distressed women.^4,5 Data on mechanisms underlying these associations are growing, although most studies to date are retrospective⁶ and mostly conducted in North America or Europe. PTB and LBW are leading causes of neonatal morbidity and mortality worldwide⁷ and over 25% of global PTBs occur in the World Health Organization (WHO) Africa Region.⁸ The burden of PTB- and LBW-related disability in the WHO's Africa Region is double that of higher income countries, mainly due to inaccessibility of neonatal intensive care.⁹

Accruing data on maternal stress and its relationship to birth and infant outcomes among populations in the WHO Africa Region could help establish an evidence base in this setting, which is disproportionately affected by adverse perinatal outcomes.

One potential casual mechanism for the relationship between maternal stress and adverse birth outcomes is that stress alters the hypothalamic-pituitary-adrenal (HPA) axis.^10–12 Cortisol is a glucocorticoid steroid hormone released by the HPA axis in response to stress. Maternal/fetal cortisol plays a critical role in fetal development, with cortisol steadily rising throughout pregnancy and peaking at delivery^13–15; chronic stress (i.e., allostatic load) may disrupt normal cortisol release.¹⁰ Cortisol extracted from hair reliably measures HPA axis activity over several months.^16,17 Data suggest that women who spontaneously deliver preterm have lower hair cortisol levels, potentially due to a cortisol response blunted by allostatic overload.^18,19 In settings with high HIV prevalence in East and Southern Africa,^20,21 pregnancy is associated with increased relationship instability,²² intimate partner violence (IPV),²³ and risk of human immunodeficiency virus (HIV) acquisition²⁴ which uniquely contribute to chronic stress.

HPA axis activation is also linked to inflammatory biomarkers that are implicated in increased HIV vulnerability in nonpregnant populations in Kenya.²⁵ To date, no studies have examined the relationship between maternal and infant hair cortisol, stressors in the perinatal period, or adverse pregnancy or child outcomes in a setting of the WHO Africa Region with high HIV prevalence.

We characterized hair cortisol levels among Kenyan mother-infant pairs and evaluated the association between hair cortisol levels in late pregnancy and maternal stressors in pregnancy and adverse birth outcomes (PTB, LBW, and small-for-gestational-age, [SGA]) with the overall objective of exploring possible mechanisms for adverse birth and other child health outcomes influenced by maternal stress in this setting.

Methods

Study setting and participants

We analyzed data and hair specimens collected from participants enrolled in the pre-exposure prophylaxis (PrEP) Implementation for Mothers in Antenatal Care (PrIMA) study.^26,27 The PrIMA study is a recently completed cluster randomized trial which compared PrEP for HIV prevention delivery models for pregnant cisgender women (NCT03070600). Recruitment, enrollment, and study procedures have previously been described.²⁶ In brief, cisgender women who were HIV-negative enrolled during pregnancy at any gestational age at 20 sites in Siaya County and Homa Bay County, Kenya—counties with 20% HIV prevalence among women^20,21—within routine antenatal care clinics were followed through 9 months postpartum. Enrollment began in January 2018 and follow-up completed in May 2021.

Protocols were approved by the Kenyatta National Hospital-University of Nairobi Ethics and Research Committee (P73/02/2017) and the University of Washington Institutional Review Board (STUDY00000438) before the initiation of the study. Hair samples collected at 6 weeks postpartum from PrEP-unexposed mother-infant pairs enrolled in the PrIMA study were utilized in this substudy to quantify maternal and infant hair cortisol concentrations (HCC).

For exploratory analyses, we also utilized data collected between October 2020 and March 2023 from an ongoing PrIMA Extension Study. The extension cohort was established to evaluate child health outcomes among mother-child pairs enrolled at four PrIMA sites to be followed until the child's 5th birthday with study visits every 6 months. Extension cohort participants were included in the exploratory analyses if they had data available on perinatal hair cortisol at 6 weeks postpartum and developmental outcomes at 36 months.

Data collection and measures

Study nurses collected information on demographic and behavioral characteristics, pregnancy history, partner and relationship characteristics, and psychosocial factors through questionnaires administered in Kiswahili, Dholuo, or English languages using tablet-based REDCap surveys. Birth outcomes were abstracted from medical records at the 6-week postpartum visit as previously described.^26,27 PTB was defined as <37 weeks gestation at birth. LBW was defined as birth weight <2.5 kilograms, and SGA was defined as a birth weight of less than 10th percentile for gestational age. Partner characteristics, including partner HIV status, were ascertained via self-report by women.

Potential maternal stressors included having a partner known to be living with HIV and/or of unknown HIV status, relationship instability, high HIV risk score, high perceived HIV risk, depression, IPV, low social support, household crowding, and maternal adverse childhood experiences (ACEs), based on characteristics ascertained in the parent study. Relationship instability was defined as responding “yes” to the question, “Do you think your primary partner had sex with anyone besides you in the past 3 months?” We used a validated risk score developed to predict HIV acquisition among perinatal women in Kenya to define high HIV risk; a HIV risk score of >6 was considered “high” (corresponding to 8.9 HIV infections per 100 person-years).²⁹ To measure self-perceived risk for HIV acquisition, we asked participants “What is your gut feeling about how likely you are to get infected with HIV?,” with four Likert response options (very likely, somewhat likely, very unlikely, or extremely unlikely).³⁰

We evaluated depression using the 2-item Patient Health Questionnaire-2 (PHQ-2), defining depression as scores ≥3.²⁸ We used the 4-item Hurt, Insult, Threaten, Scream Scale with a cutpoint of 10 or greater (absolute score range: 4–20) to determine history of IPV.²⁷ The 18-item Medical Outcomes Study Social Support Survey (MOS-SSS; higher scores, higher social support, range: 18–90) was used to collect social support data.²⁶ We dichotomized this variable with a cutpoint of less than 72 denoting low social support based on prior studies that used MOS-SSS in Kenya.^29,30 Household crowding was collected as a marker of socioeconomic status,²⁸ defined as the ratio of people per room in a residence greater than the median (>2 people/room).

In the extension cohort, study nurses assessed development using the Ages and Stages Social-Emotional Questionnaire (ASQ-SE),^31,32 an early developmental screener. ASE-SE scores at 36 months of life were dichotomized with scores >105 indicating that some milestone attainment was not evident and further evaluation was recommended.³³ Maternal ACEs were also measured among extension cohort participants using the Adverse Childhood Experiences International Questionnaire (ACE-IQ); high ACE-IQ scores were defined as scores ≥6 based on established cutpoints.^34,35

Laboratory procedures

Starting in early 2019, we simultaneously collected small scalp hair samples (∼100 strands) from all willing participants and their infants enrolled in PrIMA at every postnatal follow-up visit (6 weeks; 3, 6, 9 months). The only exclusion criteria for hair collection were failure to obtain maternal consent for hair collection and lack of scalp hair to collect in neonates. Overall, acceptability of hair collection was high (>90%). For hair collection, ∼100 strands of hair were cut as close as possible to the scalp in the occipital region, placed in aluminum foil with the distal end labeled, and stored at room temperature in a dark environment until shipment to the University of Washington in Seattle, Washington, USA. Hair cortisol is stable in hair samples stored at room temperature indefinitely if the samples are protected from ultraviolet light.³⁶

We utilized hair samples collected at 6 weeks postpartum from mothers and infants to quantify HCC in 3 cm of proximal hair (reflecting cortisol levels over the past 3 months, including the last 6 weeks of pregnancy).³⁶ We followed standard washing and processing steps of HCC analysis^36,37 as described by Meyer et al. to extract cortisol from hair at the University of Washington Office of Nursing Research Laboratory.³⁸ In brief, the proximal 3 cm of hair was identified in each sample, and distal segments were discarded. Based on an average hair growth rate of 0.79 cm per month among populations in African countries,^39,40 each hair sample represented an assumed cumulative cortisol concentration of ∼2.5–3 months.

After standard washing and processing of hair samples,^36,37 HCC was measured by enzyme immunoassay (ELISA).³⁸ A purified polyclonal anticortisol antibody and cortisol reference calibrators were used.²⁵ Two dilutions of a commercial quality-control sample were run on every plate. Intra- and interassay coefficients of variation were calculated and reported using a variance components model for all plates used in this study.

Analyses

We compared characteristics of women with and without paired infant hair samples using Wilcoxon rank-sum tests for continuous measures and chi-square tests for proportions with Fisher exact tests for small cell counts. HCC was log₁₀-transformed in all analyses, as is typical with hair cortisol data. Values below the lower limit of quantification (LLOQ) were set equal to the LLOQ value before log-transformation. The median, mean, and ranges of maternal and infant HCC were calculated, and scatterplots were used to describe the overall distributions. We generated a composite maternal stressor index using variables associated (p < 0.25) with HCC in univariate models (partner HIV status, high HIV risk score, and IPV during pregnancy), given the exploratory nature of the study. Each stressor was assigned a value of “1” and a cumulative stressor index was calculated by summing the number of stressors for each woman.

A separate dichotomized variable was generated for ≥2 stressors versus <2 based on frequency distributions of the composite index score. Separate linear regression models for mothers and infants were performed to test whether mean HCC was associated with individual maternal stressors. In exploratory analyses, linear regression was also performed to test whether there is an association between mean HCC and perinatal outcomes. Multivariate models adjusted for HIV partner status, relationship instability, and IPV because of their known association with both stress and/or mental health issues and HIV acquisition risk during the perinatal period.

We used scatterplots and Spearman correlation coefficients to assess correlations between maternal and infant hair cortisol levels. In the primary correlation analysis, we included continuous HCC values from all mother-infant dyads together. We conducted separate sensitivity analyses by restricting the sample to each respective quartile of maternal hair cortisol. We also tested whether extreme outliers drive the correlations that we observe by trimming observations at the 95th percentile (i.e., including only observations within two standard deviations of the median). We conducted exploratory analyses among the subset of extension cohort participants to test whether there is an association between maternal or infant HCC at 6 weeks postpartum and child development outcomes at 36 months.

Results

Participant characteristics

Overall, 153 women had HCC available at 6 weeks postpartum and were included in the analysis; all had live infants at 6 weeks and 106 women also had paired infant HCC available. The most frequent reason for lack of infant hair sample collection was inadequate amount of infant hair (72%). Median maternal age was 25 years (interquartile range [IQR]: 22–30), median years of completed education was 12 (IQR: 8–12), and 95% were married. At the time of enrollment into the parent study, the median gestational age was 20 weeks (IQR: 16–26), 22% were primigravida, and 12% had experienced a prior pregnancy loss. Median gestational age at birth was 39 weeks (IQR: 38–40) and 13% of women experienced PTB. Median birthweight was 3.4 kilograms (IQR: 3.1–3.7) and 2% infants had low birthweight; 11% were SGA.

Maternal stressors during pregnancy were common (Table 1) and the most frequently occurring stressors were relationship instability (61%), having a high HIV risk score (40%), and household crowding (35%). A majority of women (84%) had at least one stressor during pregnancy and 64% had 2 or more stressors. Among women who had information on ACEs (n = 74), 30% had high ACE scores >6. Women who had paired infant hair samples available were more likely to have a male partner of unknown HIV status and have high HIV risk scores; there were no differences in other demographic characteristics, pregnancy history or outcomes, or stressors between women with and without paired infant hair samples (Table 1).

Table 1.

Characteristics of Human Immunodeficiency Virus-Negative Women with Hair Cortisol Concentration Available at 6 Weeks Postpartum (n = 153)¹

Demographic characteristics	N	n (%) or Median (IQR)
		Overall ( n = 153)	Has paired infant hair sample
		Overall ( n = 153)	Yes ( n = 106)	No ( n = 47)	p -value¹⁰
Age (years)	152	25.0 (21.8, 30.2)	25.0 (21.2, 30.3)	24.9 (22.0, 29.3)	0.899
Age <24 years	152	88 (57.9%)	61 (58.1%)	27 (57.5%)	0.940
Gestational age at enrollment (weeks)	153	20 (16, 26)	20 (15, 26)	21 (17, 26)	0.549
Currently married	151	143 (94.7%)	99 (94.3%)	44 (95.7%)	0.730
Completed education (years)	152	12 (8, 12)	12 (8, 14)	10.5 (8, 12)	0.439
Regularly employed	152	29 (19.1%)	22 (20.8%)	7 (15.2%)	0.425
Pregnancy history
Primigravida	153	33 (21.6%)	24 (22.6%)	9 (19.2%)	0.628
No. of living children	148	1 (0, 2)	1 (0, 2)	1 (1, 3)	0.378
Prior pregnancy loss	153	19 (12.4%)	15 (14.2%)	4 (8.5%)	0.329
Prior preterm birth	153	0 (0.0%)	0 (0.0%)	0 (0.0%)	—
Syphilis diagnosis	149	0 (0.0%)	0 (0.0%)	0 (0.0%)	—
Maternal stressors in pregnancy
Partner HIV status unknown	153	57 (37.3%)	47 (44.3%)	10 (21.3%)	0.006
Relationship instability²	145	89 (61.4%)	58 (58.0%)	31 (68.9%)	0.213
High HIV risk³	153	61 (39.9%)	49 (46.2%)	12 (25.5%)	0.016
High self-perceived HIV risk⁴	153	5 (3.3%)	5 (4.7%)	0 (0.0%)	0.324
Depression⁵	152	16 (10.5%)	11 (10.4%)	5 (10.9%)	0.928
Intimate partner violence⁶	152	6 (4.0%)	5 (4.7%)	1 (2.2%)	0.668
Low social support⁷	150	26 (17.3%)	19 (18.5%)	7 (14.9%)	0.594
Household crowding⁸	152	53 (34.9%)	35 (33.0%)	18 (39.1%)	0.468
High maternal ACES⁹	74	22 (29.7%)	18 (31.0%)	4 (25.0%)	0.640
Perinatal outcomes
Gestational age at birth (weeks)	152	38.9 (38, 40)	38.9 (38, 40)	38.6 (38, 40)	0.479
Gestational age at birth <37 weeks	152	20 (13.2%)	14 (13.2%)	6 (13.0%)	0.978
Infant sex at birth (female)	150	75 (50.0%)	58 (55.2%)	17 (37.8%)	0.050
Birthweight (kilograms)	106	3.4 (3.1, 3.7)	3.4 (3.1, 3.7)	3.4 (3.1, 4)	0.749
Birthweight <2.5 kilograms	106	2 (1.9%)	0 (0.0%)	2 (6.7%)	0.078
Small-for-gestational age	105	11 (10.5%)	8 (10.5%)	3 (10.3%)	0.978

Bold indicates statistically significant differences between mothers with and without paired infant hair samples.

Characteristics assessed at enrollment during pregnancy unless otherwise indicated.

Relationship instability was defined as responding “yes” or “don't know” to the question, “Do you think your primary partner had sex with anyone besides you in the past 3 months?”

We evaluated behavioral HIV risk using an assessment tool that had an AUC of 0.76 to predict HIV incidence in pregnancy and postpartum in Kenyan women (Pintye et al. CID 2017); a HIV risk score of >6 was considered “high” (corresponding to 8.9 HIV infections per 100 person-years).

To measure self-perceived risk for HIV acquisition, we asked participants “What is your gut feeling about how likely you are to get infected with HIV?,” with four Likert response options (very likely, somewhat likely, very unlikely, or extremely unlikely), defining high self-perceived HIV risk as “very likely.”

We evaluated depression using the 2-item Patient Health Questionnaire-2 (PHQ-2), defining depression as scores ≥3.

We evaluated IPV using the 4-item Hurt, Insult, Threaten, and Scream scale, defining IPV as scores of 10 and above.

We evaluated social support using the 18-item Medical Outcomes Study social support score (MOS-SSS), defining low social support as scores <72.

Household crowding was defined as the ratio of people per room in a residence >2 people/room.

Adverse Childhood Experience score ≥6.

Wilcoxon rank-sum tests for continuous measures and Chi-square tests for proportions. Fisher exact tests used for variables with cell counts <5. Statistical tests compare participants with and without paired infant hair samples (i.e., “yes” vs. “no”).

Maternal and infant HCC

The overall median maternal HCC (n = 153) was 12.0 pg/mg (IQR: 6.3–20.4), the mean HCC was 17.2 pg/mg (SD 16.2), and the absolute range was 1.1–94.9 pg/mg; 3% had HCC below the LLOQ and none had HCC above the upper limit of quantification (ULOQ). The median maternal HCC for the lowest quartile was 4.0 pg/mg (IQR: 3.0–5.3) and 31.2 pg/mg (IQR: 24.9–56.8) for the highest quartile (Fig. 1a). The overall median infant HCC (n = 106) was 10.4 pg/mg (IQR: 5.5–23.7) and the mean HCC was 16.7 pg/mg (SD 17.5), and the absolute range was 1.4–79.2 pg/mg; 5% had HCC below the LLOQ and none had HCC above the ULOQ. The median infant HCC for the lowest quartile was 3.3 pg/mg (IQR: 2.5–4.7) and 37.2 pg/mg (IQR: 28.2–51.3) for the highest quartile (Fig. 1b).

FIG. 1.

(a) Distribution of log₁₀ maternal hair cortisol concentrations at 6 weeks postpartum, overall and by quartile (n = 153). (b) Distribution of log₁₀ infant hair cortisol concentrations at 6 weeks postpartum, overall and by quartile (n = 106).

Among mother-infant pairs (n = 106), infant HCC had a positive correlation with maternal HCC (r = 0.498, p < 0.001, Fig. 2) with similar correlation when stratified by quartiles and trimming outliers (n = 3) in sensitivity analyses (data not shown).

FIG. 2.

Correlation between maternal and infant log₁₀ hair cortisol concentrations at 6 weeks postpartum, overall and by quartile (n = 106) r = 0.498, p < 0.001.

Maternal stressors in pregnancy and HCC

In univariable models, stressors during pregnancy associated with maternal HCC (p < 0.25) included having a partner of unknown HIV status or known to be living with HIV, relationship instability, having a high HIV risk score, experiencing IPV during pregnancy, and high ACE scores (Table 2). In multivariable models, having a partner of unknown HIV status or known to be living with HIV was associated with 35% higher adjusted mean HCC than women with partners believed to HIV-negative (95% CI: 2%–81%, p = 0.04). Having a high HIV risk score was associated with having a 37% higher adjusted mean HCC than women with lower risk scores (95% CI: 3%–83%, p = 0.03).

Table 2.

Association of Stressors During Pregnancy and Mean log₁₀ Maternal Hair Cortisol Concentrations at 6 Weeks Postpartum (n = 153)

	N (%)	Mean log₁₀ HCC concentration	Univariable linear regression			Multivariable linear regression⁴
Maternal stressors¹	N (%)	Mean log₁₀ HCC concentration	Absolute difference² (95% CI)	% difference³ (95% CI)	p -value	Absolute difference² (95% CI)	% difference³ (95% CI)	p -value
Partner HIV status (n = 153)
Unknown/Positive	58 (38%)	1.1621	0.140 (0.017 to 0.263)	38.1 (4.1 to 83.2)	0.026	0.133 (0.007 to 0.259)	35.8 (1.6 to 81.4)	0.039^a
Negative	95 (62%)	1.0220	Ref.			Ref.
Relationship instability (n = 145)⁵
Yes	89 (61%)	1.0303	−0.119 (−0.245 to 0.007)	−24.0 (−43.1 to 1.6)	0.063	−0.136 (−0.260 to −0.011)	−26.8 (−45.1 to −2.5)	0.033^b
No	56 (39%)	1.1494	Ref.
High HIV risk score (n = 153)⁶
Yes	61 (40%)	1.1610	0.143 (0.021 to 0.264)	38.9 (5.0 to 83.8)	0.022	0.137 (0.013 to 0.262)	37.2 (3.1 to 82.7)	0.030^a
No	92 (60%)	1.0182	Ref.			Ref.
High perceived HIV risk (n = 153)⁷
Yes	5 (3%)	0.9944	−0.083 (−0.424 to 0.257)	−17.5 (−62.3 to 80.7)	0.629
No	148 (97%)	1.0778	Ref.
Depression (n = 152)⁸
Yes	16 (11%)	1.0120	−0.067 (−0.265 to 0.130)	−14.4 (−45.7 to 34.9)	0.501
No	136 (89%)	1.0795	Ref.
IPV during pregnancy (n = 153)⁹
Yes	16 (10%)	1.1964	0.136 (−0.061 to 0.332)	36.6 (−13.1 to 114.9)	0.175	0.119 (−0.075 to 0.313)	31.5 (−15.9 to 105.7)	0.228^c
No	137 (90%)	1.0609	Ref.			Ref.
Low social support (n = 150)¹⁰
Yes	26 (17%)	1.1296	0.070 (−0.093 to 0.232)	17.4 (−19.3 to 70.7)	0.399
No	124 (83%)	1.0600	Ref.
Household crowding (n = 152)¹¹
Yes	53 (35%)	1.0972	0.035 (−0.093 to 0.163)	8.4 (−19.2 to 45.5)	0.588
No	99 (65%)	1.0621	Ref.
ACEs score (n = 74)¹²
<6	52 (70.3%)	0.9409	Ref.			Ref.
≥6	22 (29.7%)	1.0503	0.109 (−0.060 to 0.279)	28.6 (−12.9 to 90.0)	0.202	0.137 (−0.035 to 0.310)	37.2 (−7.8 to 104.1)	0.117^e
Number of stressors (n = 153)¹³
≥2	58 (38%)	1.1621	0.140 (0.0171 to 0.263)	38.1 (4.1 to 83.2)	0.026	0.143 (0.018 to 0.268)	38.9 (4.2 to 85.23)	0.026^d
<2	95 (62%)	1.0220	Ref.			Ref.

Bold indicates statistically significant differences between mothers with and without paired infant hair samples.

Stressors assessed at enrollment during pregnancy unless otherwise indicated.

Estimated absolute difference in adjusted mean log₁₀-transformed HCC concentration women with specific characteristic versus the reference group for categorical covariate measures.

Estimated percent change in adjusted mean HCC concentration between women with specific characteristic versus the reference group for categorical covariate measures.

Multivariable models were adjusted for partner HIV status, relationship instability, and IPV.

Relationship instability was defined as responding “yes” to the question, “Do you think your primary partner had sex with anyone besides you in the past 3 mos?”

HIV risk was evaluated using an assessment tool to predict HIV incidence in pregnancy and postpartum in Kenyan women (Pintye et al. CID 2017); a HIV risk score of >6 was considered “high” and corresponds to 8.9 HIV infections per 100 person-years (high HIV risk: score ≥6 = “Yes,” score <6 = “No”).

To measure self-perceived risk for HIV acquisition, participants were asked “What is your gut feeling about how likely you are to get infected with HIV?,” with four Likert response options (very likely, somewhat likely, very unlikely, or extremely unlikely), defining high self-perceived HIV risk as “very likely.”

Depression was defined using the 2-item Patient Health Questionnaire-2 (depression: PHQ-2 score ≥3 = “Yes,” PHQ-2 score <3 = “No”)

IPV was evaluated using the 4-item Hurt, Insult, Threaten, and Scream Scale (IPV: HITS score ≥10 = “Yes,” HITS score <10 = “No”) at any time in pregnancy

Social support was defined using the 18-item Medical Outcomes Study social support score (MOS-SSS), defining low social support as scores <72 (low social support: MOS-SSS score <72 = “Yes,” MOS-SSS score ≥72 = “No”)

Household crowding was defined as the ratio of people per room in a residence >2 people/room

Adverse Childhood Experience score ≥6.

Each stressor which was positively associated with maternal HCC concentration (partner HIV status, High HIV risk score, and IPV during pregnancy) was assigned a value of “1”, and we calculated a cumulative stressor index by summing the number of stressors for each individual woman. Only women with complete data on all stressors were included in the cumulative stressor index analysis.

Adjusted for relationship instability and IPV.

Adjusted for partner HIV status and IPV.

Adjusted for partner HIV status and relationship instability.

Adjusted for relationship instability.

Adjusted for partner HIV status, relationship instability, and IPV.

ACE, adverse childhood experience; HCC, hair cortisol concentrations; IPV, intimate partner violence.

Having a high ACE score trended toward an association with a 37% higher mean HCC than women with lower ACE scores (95% CI −8% to 104%, p = 0.117), although few women had ACE scores available (n = 74). Women reporting relationship instability had 27% lower adjusted mean HCC than women without relationship instability (95% CI: −45% to −3%, p = 0.033); no other stressors were associated with lower HCC. Having ≥2 stressors during pregnancy was associated with a 39% higher adjusted mean HCC than having <2 stressors (95% CI 4%–85%, p = 0.026).

In univariable models, maternal stressors during pregnancy associated with infant HCC (p < 0.25) included relationship instability, high perceived HIV risk, and high maternal ACE scores (Table 3). In multivariable models, no maternal stressors remained significantly associated with infant HCC, although relationship instability (p = 0.098) trended toward an association with lower mean infant HCC.

Table 3.

Association of Maternal Stressors During Pregnancy and Mean log₁₀ Infant Hair Cortisol Concentrations at 6 Weeks (n = 106)

	N (%)	Mean log₁₀ HCC concentration	Univariable linear regression			Multivariable linear regression⁴
Maternal stressors¹	N (%)	Mean log₁₀ HCC concentration	Absolute difference² (95% CI)	% difference³ (95% CI)	p -value	Absolute difference² (95% CI)	% difference³ (95% CI)	p -value
Partner HIV status (n = 106)
Unknown/Positive	48 (45%)	1.0672	0.079 (−0.085 to 0.242)	19.8 (−17.7 to 74.45)	0.342
Negative	58 (55%)	0.9887	Ref.
Relationship instability (n = 100)⁵
Yes	58 (58%)	0.9720	−0.142 (−0.311 to 0.026)	−28.0 (−51.1 to 6.2)	0.097	−0.141 (−0.309 to 0.026)	−27.8 (−50.9 to 6.3)	0.098
No	42 (42%)	1.1144	Ref.			Ref
High HIV risk score (n = 106)⁶
Yes	49 (46%)	1.0557	0.059 (−0.105 to 0.222)	14.4 (−21.4 to 66.6)	0.478
No	57 (54%)	0.9972	Ref.
High perceived HIV risk (n = 106)⁷
Yes	5 (5%)	0.7604	−0.277 (−0.658 to 0.104)	−47.1 (−78.0 to 27.1)	0.152	−0.283 (−0.662 to 0.097)	−47.8 (−78.2 to 25.0)	0.143
No	101 (95%)	1.0373	Ref.			Ref
Depression (n = 106)⁸
Yes	11 (10%)	1.0012	−0.026 (−0.293 to 0.242)	−5.7 (−49.1 to 74.5)	0.849
No	95 (90%)	1.0269	Ref.
IPV during pregnancy (n = 106)⁹
Yes	13 (12%)	1.0971	0.083 (−0.165 to 0.331)	21.1 (−31.6 to 114.4)	0.508
No	93 (88%)	1.0141	Ref.
Low social support (n = 103)¹⁰
Yes	19 (18%)	0.9315	−0.105 (−0.315 to 0.105)	−21.5 (−51.6 to 27.3)	0.323
No	84 (82%)	1.0368	Ref.			Ref.
Household crowding (n = 106)¹¹
Yes	35 (33%)	1.0121	−0.018 (−0.192 to 0.155)	−4.1 (−35.7 to 43.0)	0.836
No	71 (67%)	1.0302	Ref.
Maternal ACES (n = 58)¹²
<6	40 (69.0%)	1.0332	Ref.
≥6	18 (31.0%)	0.8693	−0.164 (−0.408 to 0.080)	−31.4 (−60.9 to 20.3)	0.184	−0.104 (−0.370 to 0.162)	−21.3 (−57.3 to 45.2)	0.436
Number of stressors (n = 106)¹³
≥2	48 (45%)	1.0672	0.079 (−0.085 to 0.242)	19.8 (−17.7 to 74.4)	0.342
<2	58 (55%)	0.9887	Ref.

Stressors assessed at enrollment during pregnancy unless otherwise indicated.

Estimated absolute difference in adjusted mean log₁₀-transformed HCC concentration between women with specific characteristic versus the reference group for categorical covariate measures.

Estimated percent change in adjusted mean HCC concentration between women with specific characteristic versus the reference group for categorical covariate measures.

The multivariable model for relationship instability was adjusted for high perceived HIV risk. The multivariable model for high perceived HIV risk was adjusted for relationship instability. The multivariable model for maternal ACEs was adjusted for high perceived HIV risk and relationship instability.

Relationship instability was defined as responding “yes” to the question, “Do you think your primary partner had sex with anyone besides you in the past 3 mos?”

Depression was defined using the 2-item Patient Health Questionnaire-2 (depression: PHQ-2 score ≥3 = “Yes,” PHQ-2 score <3 = “No”)

IPV was evaluated using the 4-item Hurt, Insult, Threaten, and Scream Scale (IPV: HITS score ≥10 = “Yes,” HITS score <10 = “No”)

Household crowding was defined as the ratio of people per room in a residence >2 people/room

Adverse Childhood Experience score ≥6.

Each stressor which was positively associated with maternal HCC concentration (partner HIV status, High HIV risk score, and IPV during pregnancy) was assigned a value of “1” and we calculated a cumulative stressor index by summing the number of stressors for each individual woman. Only women with complete data on all stressors were included in the cumulative stressor index analysis.

Pregnancy outcomes and HCC

In univariable models, maternal HCC was not associated with experiencing PTB (p = 0.809), having an infant with LBW (p = 0.971) or having an infant who was SGA (p = 0.254, Table 4). Infant HCC was not associated with PTB or being SGA; no cases of LBW occurred among infants with HCC available. Our results should be interpreted with caution due to the small number of adverse pregnancy outcomes.

Table 4.

Association of Pregnancy Outcomes and Mean log₁₀ Maternal and Infant Hair Cortisol Concentrations at 6 Weeks Postpartum

	N (%)	Mean log₁₀ HCC concentration	Univariable linear regression			Multivariable linear regression³³
	N (%)	Mean log₁₀ HCC concentration	Absolute difference¹ (95% CI)	% difference² (95% CI)	p -value	Absolute difference¹ (95% CI)	% difference² (95% CI)	p -value
Maternal HCC (n = 153)
Gestational age at birth (n = 152)
<37 weeks	20 (13%)	1.0929	0.022 (−0.158 to 0.202)	5.2 (−30.5 to 59.3)	0.809
≥37 weeks	132 (87%)	1.0708	Ref.
Birthweight (n = 106)
<2.5 kilograms	2 (2%)	1.0504	0.010 (−0.568 to 0.548)	−2.4 (−73.0 to 252.8)	0.971
≥2.5 kilograms	104 (98%)	1.0608	Ref.
Small-for-gestational age (n = 105)
Yes	10 (10%)	0.8229	−0.150 (−0.409 to 0.109)	−29.2 (−61.0 to 28.6)	0.254	−0.043 (−0.315 to 0.228)	−9.5 (−51.5 to 69.1)	0.752
No	95 (90%)	1.0727	Ref.			Ref.
Infant HCC (n = 106)
Gestational age at birth (n = 106)
<37 weeks	14 (13%)	1.1793	0.179 (−0.060 to 0.417)	50.9 (−12.9 to 161.2)	0.140	0.195 (−0.061 to 0.452)	56.8 (−13.3 to 183.0)	0.134
≥37 weeks	92 (87%)	1.0001	Ref.			Ref.
Birthweight (n = 76)
<2.5 kilograms	0 (0%)	—	—	—	—
≥2.5 kilograms	76 (100%)	0.9943	—
Small-for-gestational age (n = 76)
Yes	7 (9%)	0.8923	−0.126 (−0.470 to 0.217)	−25.2 (−66.1 to 64.8)	0.466
No	69 (91%)	1.0186	Ref.

Estimated absolute difference in adjusted mean log₁₀-transformed HCC concentration between pregnancies with specific outcome versus the reference group.

Estimated percent change in adjusted mean HCC concentration between pregnancies with specific outcome versus the reference group.

Covariates with p-value <0.25 in univariable analysis of maternal stressors and HCC were included in the multivariable model (HIV partner status, relationship instability, and IPV).

Exploratory analyses: Child developmental outcomes and HCC

Overall, 72 mother-infant pairs were enrolled in the extension cohort and had data on maternal HCC at 6 weeks postpartum and child development outcomes at 36 months of life. Of these 72 children, 10 (14%) had Ages & Stages scores >105, indicating that some milestone attainment was not evident and further evaluation was recommended. Mean maternal HCC at 6 weeks postpartum did not differ between children with Ages & Stages scores >105 and those with lower scores in univariable models (mean percent difference −20.2%, 95% CI: −53.0 to 35.6, p = 0.399). In total, 55 children had data on infant HCC at 6 weeks of life and development outcomes at 36 months of whom 10 (18%) had Ages & Stages >105.

In exploratory univariable models, having an Ages & Stages score >105 at 36 months was not associated with mean infant HCC at 6 weeks (mean percent difference 38.7%, 95% CI: −68.7 to 19.9, p = 0.149, Table 5), although precision was limited.

Table 5.

Association Mean log₁₀ Maternal and Infant Hair Cortisol Concentrations at 6 Weeks Postpartum and Child Developmental Outcomes 36-Months

	N (%)	Mean log₁₀ HCC concentration	Univariable linear regression
	N (%)	Mean log₁₀ HCC concentration	Absolute difference¹ (95% CI)	% difference² (95% CI)	p -value
Maternal HCC (n = 72)
ASQ-SE score
≤105	62 (86%)	0.9469	−0.098 (−0.328 to 0.132)	−20.2 (−53.0 to 35.6)	0.399
>105	10 (14%)	1.0447	Ref.
Infant HCC (n = 55)
ASQ-SE score
≤105	45 (82%)	0.9205	−0.213 (−0.505 to 0.079)	−38.7 (−68.7 to 19.9)	0.149
>105	10 (18%)	1.1334	Ref.

Estimated absolute difference in adjusted mean log₁₀-transformed HCC concentration between pregnancies with specific outcome versus the reference group.

Estimated percent change in adjusted mean HCC concentration between pregnancies with specific outcome versus the reference group.

ASQ-SE, Ages and Stages Social-Emotional Questionnaire.

Discussion

In this study of Kenyan mothers and their infants in a high HIV prevalence setting, we found that maternal HCC during the peripartum period was associated with experiencing stressors during pregnancy, especially stressors associated with maternal HIV acquisition. Similar to prior studies,^37,41,42 we also found that infant HCC was correlated with maternal HCC. We did not find an association between maternal HCC in samples collected at 6 weeks postpartum and any adverse pregnancy outcomes, unlike prior studies of perinatal HCC and birth outcomes.^43,44 Our findings suggest that maternal stressors during pregnancy impact HPA axis activity. Our results also suggest that maternal perinatal stress influences fetal HPA axis activity in the third trimester.

Data from our study contribute to the growing body of literature on mechanisms to explain how perinatal stress exposure influences pregnancy and infant outcomes and, to our knowledge, is the first study of maternal-infant dyadic HCC among a population in a high HIV prevalence setting in the WHO Africa Region.

Prior studies among peripartum women, mainly in North America and Europe, found HCC between 4.99 and 32.00 pg/mg,^45–51 which is within the range of HCC among Kenyan women in our study at 6 weeks postpartum (mean HCC 17.2 pg/mg) and a prior study among pregnant Kenyan women at 22–28 weeks gestation (mean HCC 6.11 pg/mg).⁵² As early as the first trimester, cortisol levels are higher than nonpregnant periods⁵³ and substantial increases in cortisol levels are associated with each subsequent trimester.^43,54 Few studies to date have measured HCC among populations in the WHO African Region,^52,55,56 and to our knowledge, no studies have assessed hair cortisol among mother-infant dyads or pediatric populations in this setting.

Other studies among nonpregnant women in Kenya living in “high stress” situations (i.e., living in settlements and female sex-workers) found mean hair cortisol levels of 6.79 pg/mg and 13.8 pg/mg.^25,55 Our findings suggest that HCC among peripartum Kenyan women are similar to peripartum populations outside of Kenya but potentially elevated relative to other ‘high stress' nonpregnant populations in Kenya.

We found that higher maternal HCC was associated with characteristics associated with HIV acquisition, suggesting that factors such as experiencing relationship instability and having a partner known to be living with HIV or of unknown HIV status contribute to chronic stress in the perinatal period. Conversely, we found that relationship instability was associated with lower maternal HCC, possibly due to a blunting effect of allostatic load, although further study is required to understand the negative association.^18,19 A recent study among female sex workers in Kenya found that exposure to physical, sexual, or emotional violence was associated with increased HCC and inflammatory biomarkers, specifically systemic C-reactive protein and interleukin-6 levels.²⁵

Physiological stress responses, such as inflammation, may alter systemic immune responses.⁵⁷ Data suggest that HPA axis activation may lead to glucocorticoid receptor resistance that could result in failure to downregulate inflammatory responses to viral triggers, such as HIV.⁵⁸ Studies from Kenya and Uganda found a two-fold increase in the risk of HIV acquisition among women during pregnancy and postpartum compared to nonpregnant women.²⁴ Therefore, elucidating pathways to HIV susceptibility during pregnancy, including the contribution of chronic stress and HPA axis activation, warrants further inquiry.

We observed a correlation between maternal and infant HCC at 6 weeks postpartum, representing the last few months of the third trimester. Few studies have evaluated the relationship between maternal cortisol and neonatal cortisol. In nonhuman primate studies, a negative association between higher maternal HCC and birth weight was observed^59,60 and exposure to maternal stress in pregnancy was associated with lower infant hair cortisol levels.⁶¹ In human studies, a positive relationship between maternal and infant hair cortisol at 6, 9, and 12 months after birth was observed,^41,42 similar to our findings at 6 weeks postpartum. High levels of stress during the third trimester affect the long-term development of the newborn infant⁶² and evidence suggests that maternal cortisol affects fetal cortisol from the first trimester.⁶³

A longitudinal study of 80 women and their infants in Granada, Spain, found that higher maternal hair cortisol levels in the first trimester of pregnancy predicted lower neonatal hair cortisol,⁶³ similar to a prior study among rhesus monkeys.⁶¹ More longitudinal data from early pregnancy through the postpartum period are needed to understand the time-specific relationship between biological measures of stress, such as cortisol, and fetal programming of the HPA and infant stress regulation.⁶³

Unlike prior studies in other populations,^43,44 we did not detect an association between maternal HCC and PTB or any adverse birth outcome among Kenya women, although our statistical power was limited due to the low number of adverse outcomes. A growing evidence base supports the impact of periconception and perinatal stress exposure on birth outcomes.^6,64 Disruption of the HPA axis is one potential casual mechanism linking maternal stress to PTB.¹⁰ Proinflammatory cytokines stimulate the HPA axis and result in secretion of cortisol, the hormonal product of the HPA axis. In turn, cortisol downregulates proinflammatory cytokines⁶⁵ and stimulates production of anti-inflammatory cytokines.⁶⁶ Stress alters the HPA axis, resulting in hyperactivation in the short-term and hypoactivation in the long-term¹¹ (i.e., allostatic overload).¹² A review of 15 studies among European and North American mothers with sample sizes ranging from 36 to 359 concluded that high maternal cortisol levels during pregnancy were associated with PTB.⁶⁷

All 15 studies cross-sectionally evaluated short-term cortisol measures in blood, saliva, or urine. During chronic stress, cortisol is less effective at suppressing the immune system and inhibiting inflammation, which may lead to increased susceptibility to infection and inflammation,^68–70 subsequently predisposing women who acquire infections to PTB.⁷¹ Existing data suggest women who deliver preterm have lower hair cortisol levels, potentially due to a cortisol response blunted by allostatic overload associated with chronic stress.^18,19 More studies evaluating maternal hair cortisol and birth outcomes are needed, ideally in diverse geographical settings and among women who experience chronic stress.^18,19,67 Clinical implications of stress during the perinatal period among women in the WHO Africa Region warrant further investigation in larger, well-powered studies.

Our study has limitations. Our sample size is limited to the number of hair samples already collected from mothers/infants in parent study and prespecified cost restrictions of hair assays. Our sample size is similar to previous studies that found a significant association between maternal hair cortisol and PTB (prior studies ranged from 58 to 137 mothers).^18,19 Larger population-based studies are needing to better elucidate the association between maternal HCC and less common birth outcomes. Perceived stress was not ascertained in the parent study and therefore no subjective measurement of stress was included in analyses. However, a meta–analysis from 2017 of 26 studies that included 2,441 individuals from diverse populations, including pregnant and postpartum women, did not find an association between perceived stress and HCC,⁷² neither did one study among pregnant Kenyan women.⁵⁴

Although our study was conducted in a setting with high HIV prevalence, our sample only included women without HIV who were mostly married and aged >24 years with 1–2 living children. It is possible that women living with HIV, who are young, unmarried, or have several living children may have differences in chronic stress and HCC due to HIV disease-related or psychosocial factors,⁷³ and therefore our results are not generalizable to all women in our study setting. ACEs and developmental outcomes were only available for a subset of women enrolled in the extension cohort which limited our statistical power for these exploratory variables and may have biased these results to children who survived until 36 months.

Despite our limited sample, lower infant HCC trended toward an association with both high maternal ACEs and possible unmet developmental milestones at 36 months, warranting further study as our statistical precision was low due to sample size. We only measured cortisol concentrations in hair without comparison with other biologic measures (i.e., saliva, urine, or blood samples). In our research setting, hair has logistical advantages, including painless and noninvasive sampling without phlebotomy for infants and room temperature storage without biohazard concerns.⁷⁴

In conclusion, we found that maternal HCC during the peripartum period was associated with experiencing stressors during pregnancy, especially stressors associated with maternal HIV acquisition. Similar to prior studies, we also found that infant HCC correlated with maternal HCC.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

Funding Information

Funding for this project was provided by the National Institutes of Health, National Institute of Allergy and Infectious Disease (GJS- R01AI125498), and Eunice Kennedy Shriver National Institute of Child Health & Human Development (J.P.- R01HD100201). J.P. was additionally supported by NICHD (R01HD108041) and the National Institute of Nursing Research (R01NR019220). The team was supported by the University of Washington's Center for AIDS Research Behavioral Sciences Core and Biometrics Core (P30AI027757) and the Global Center for the Integrated Health of Women, Adolescents, and Children (Global WACh). This research was funded by a 2020 Research Intramural Funding Program grant from the University of Washington School of Nursing and the Van Hooser Fund. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the article.

Abbreviations Used

References

Graignic-Philippe

, Dayan

, Chokron

, et al. Effects of prenatal stress on fetal and child development: A critical literature review. Neurosci Biobehav Rev, 2014; 43:137–162; doi: 10.1016/j.neubiorev.2014.03.022

Hipwell

, Fu

, Tung

, et al. Preconception stress exposure from childhood to adolescence and birth outcomes: The impact of stress type, severity and consistency. Front Reprod Health, 2022; 4:1007788; doi: 10.3389/frph.2022.1007788

Talge

, Neal

, Glover

, Early Stress

TaRaPSNFaNEoCaAMH

. Antenatal maternal stress and long-term effects on child neurodevelopment: How and why?. J Child Psychol Psychiatry, 2007; 48(3–4):245–261; doi: 10.1111/j.1469-7610.2006.01714.x

Christian

. Psychoneuroimmunology in pregnancy: Immune pathways linking stress with maternal health, adverse birth outcomes, and fetal development. Neurosci Biobehav Rev, 2012; 36(1):350–361; doi: 10.1016/j.neubiorev.2011.07.005

Brown

, Yelland

, Sutherland

, et al. Stressful life events, social health issues and low birthweight in an Australian population-based birth cohort: Challenges and opportunities in antenatal care. BMC Public Health, 2011; 11:196; doi: 10.1186/1471-2458-11-196

Keenan

, Hipwell

, Class

, et al. Extending the developmental origins of disease model: Impact of preconception stress exposure on offspring neurodevelopment. Dev Psychobiol, 2018; 60(7):753–764; doi: 10.1002/dev.21773

Beck

, Wojdyla

, Say

, et al. The worldwide incidence of preterm birth: A systematic review of maternal mortality and morbidity. Bull World Health Organ, 2010; 88(1):31–38; doi: 10.2471/blt.08.062554

Blencowe

, Cousens

, Oestergaard

, et al. National, regional, and worldwide estimates of preterm birth rates in the year2010 with time trends since 1990 for selected countries: A systematic analysis and implications. Lancet, 2012; 379(9832):2162–2172; doi: 10.1016/s0140-6736(12)60820-4

Lawn

, Blencowe

, Oza

, et al. Every Newborn: Progress, priorities, and potential beyond survival. Lancet, 2014; 384(9938):189–205; doi: 10.1016/S0140-6736(14)60496-7

10.

Field

, Diego

. Cortisol: The culprit prenatal stress variable. Int J Neurosci, 2008; 118(8):1181; doi: 10.1080/00207450701820944

11.

Chrousos

. Stress and disorders of the stress system. Nat Rev Endocrinol, 2009; 5(7):374–381; doi: 10.1038/nrendo.2009.106

12.

McEwen

. Protective and damaging effects of stress mediators: Central role of the brain. Dialogues Clin Neurosci, 2006; 8(4):367–381.

13.

Morsi

, DeFranco

, Witchel

. The hypothalamic-pituitary-adrenal axis and the fetus. Horm Res Paediatr, 2018; 89(5):380–387; doi: 10.1159/000488106

14.

Wang

, Chen

, Myatt

, et al. 11β-HSD1 in human fetal membranes as a potential therapeutic target for preterm birth. Endocr Rev, 2018; 39(3):241–260; doi: 10.1210/er.2017-00188

15.

Campbell

, Linton

, Wolfe

, et al. Plasma corticotropin-releasing hormone concentrations during pregnancy and parturition. J Clin Endocrinol Metab, 1987; 64(5):1054–1059; doi: 10.1210/jcem-64-5-1054

16.

Russell

, Koren

, Rieder

, et al. Hair cortisol as a biological marker of chronic stress: Current status, future directions and unanswered questions. Psychoneuroendocrinology, 2012; 37(5):589–601; doi: 10.1016/j.psyneuen.2011.09.009

17.

Stalder

, Kirschbaum

. Analysis of cortisol in hair—state of the art and future directions. Brain Behav Immun, 2012; 26(7):1019–1029; doi: 10.1016/j.bbi.2012.02.002

18.

Gelaye

, Kirschbaum

, Zhong

, et al. Chronic HPA activity in mothers with preterm delivery: A pilot nested case-control study. J Neonatal Perinatal Med, 2019; 13(3):313–321; doi: 10.3233/NPM-180139

19.

Karakash

, Tschankoshvili

, Weedon

, et al. Hypocortisolism and preterm birth. J Neonatal Perinatal Med, 2016; 9(4):333–339; doi: 10.3233/NPM-161640

20.

National AIDS and STI Control Programme (NASCOP). Kenya AIDS Indicator Survey 2012: Final Report. 2014.

21.

National

AIDS

and STI Control Programme (NASCOP). Preliminary Kenya Population-based HIV Impact Assessment (KENPHIA) 2018 Report. NASCOP: Nairobi, 2020.

22.

Abrahams

, Lund

, Field

, et al. Factors associated with household food insecurity and depression in pregnant South African women from a low socio-economic setting: A cross-sectional study. Soc Psychiatry Psychiatr Epidemiol, 2018; 53(4):363–372; doi: 10.1007/s00127-018-1497-y

23.

Peltzer

, Jones

, Weiss

, et al. Sexual risk, serostatus and intimate partner violence among couples during pregnancy in rural South Africa. AIDS Behav, 2013; 17(2):508–516; doi: 10.1007/s10461-012-0185-2

24.

Thomson

, Hughes

, Baeten

, et al. Increased risk of female HIV-1 acquisition throughout pregnancy and postpartum: A prospective per-coital act analysis among women with HIV-1 infected partners. J Infect Dis, 2018; 218(1):16–25; doi: 10.1093/infdis/jiy113

25.

Heller

, Roberts

, Masese

, et al. Gender-based violence, physiological stress, and inflammation: A cross-sectional study. J Womens Health (Larchmt), 2018; 27(9):1152–1161; doi: 10.1089/jwh.2017.6743

26.

Kinuthia

, Dettinger

, Stern

, et al. Risk-based versus universal PrEP delivery during pregnancy: A cluster randomized trial in Western Kenya from 2018 to 2019. J Int AIDS Soc, 2023; 26(2):e26061; doi: 10.1002/jia2.26061

27.

Dettinger

, Kinuthia

, Pintye

, et al. PrEP Implementation for Mothers in Antenatal Care (PrIMA): Study protocol of a cluster randomised trial. BMJ Open, 2019; 9(3):e025122; doi: 10.1136/bmjopen-2018-025122

28.

Monahan

, Shacham

, Reece

, et al. Validity/reliability of PHQ-9 and PHQ-2 depression scales among adults living with HIV/AIDS in western Kenya. J Gen Intern Med, 2009; 24(2):189–197; doi: 10.1007/s11606-008-0846-z

29.

Larsen

, Pintye

, Abuna

, et al. Risks of adverse perinatal outcomes in relation to maternal depressive symptoms: A prospective cohort study in Kenya. Paediatr Perinat Epidemiol, 2023; 37(6):489–504; doi: 10.1111/ppe.12978

30.

Sila

, Larsen

, Kinuthia

, et al. High awareness, yet low uptake, of pre-exposure prophylaxis among adolescent girls and young women within family planning clinics in Kenya. AIDS Patient Care STDS, 2020; 34(8):336–343; doi: 10.1089/apc.2020.0037

31.

Rah

, Jung

, Lee

, et al. Systematic review and meta-analysis: Real-world accuracy of children's developmental screening tests. J Am Acad Child Adolesc Psychiatry, 2022; 62(10):1095–1109; doi: 10.1016/j.jaac.2022.12.014

32.

van Heerden

, Hsiao

, Matafwali

, et al. Support for the feasibility of the ages and stages questionnaire as a developmental screening tool: A cross-sectional study of South African and Zambian children aged 2–60months. BMC Pediatr, 2017; 17(1):55; doi: 10.1186/s12887-017-0802-3

33.

Kyerematen

, Hamb

, Oberhelman

, et al. Exploratory application of the Ages and Stages (ASQ) child development screening test in a low-income Peruvian shantytown population. BMJ Open, 2014; 4(1):e004132; doi: 10.1136/bmjopen-2013-004132

34.

Pace

, Muzi

, Rogier

, et al. The Adverse Childhood Experiences - International Questionnaire (ACE-IQ) in community samples around the world: A systematic review (part I). Child Abuse Negl, 2022; 129:105640; doi: 10.1016/j.chiabu.2022.105640

35.

World Health Organization. Adverse Childhood Experiences International Questionnaire. In Adverse Childhood Experiences International Questionnaire (ACE-IQ).. WHO. Available from: https://www.who.int/publications/m/item/adverse-childhood-experiences-international-questionnaire-(ace-iq) [Last accessed: November 8, 2023].

36.

Greff

MJE

, Levine

, Abuzgaia

, et al. Hair cortisol analysis: An update on methodological considerations and clinical applications. Clin Biochem, 2019; 63:1–9; doi: 10.1016/j.clinbiochem.2018.09.010

37.

Hoffman

, D'Anna-Hernandez

, Benitez

, et al. Cortisol during human fetal life: Characterization of a method for processing small quantities of newborn hair from 26 to 42weeks gestation. Dev Psychobiol, 2017; 59(1):123–127; doi: 10.1002/dev.21433

38.

Meyer

, Novak

, Hamel

, et al. Extraction and analysis of cortisol from human and monkey hair. J Vis Exp, 2014;(83):e50882; doi: 10.3791/50882

39.

Loussouarn

. African hair growth parameters. Br J Dermatol, 2001; 145(2):294–297; doi: 10.1046/j.1365-2133.2001.04350.x

40.

Loussouarn

, Lozano

, Panhard

, et al. Diversity in human hair growth, diameter, colour and shape. An in vivo study on young adults from 24 different ethnic groups observed in the five continents. Eur J Dermatol, 2016; 26(2):144–154; doi: 10.1684/ejd.2015.2726

41.

Flom

, St John

, Meyer

, et al. Infant hair cortisol: Associations with salivary cortisol and environmental context. Dev Psychobiol, 2017; 59(1):26–38; doi: 10.1002/dev.21449

42.

Liu

, Fink

, Brentani

, et al. An assessment of hair cortisol among postpartum Brazilian mothers and infants from a high-risk community in São Paulo: Intra-individual stability and association in mother-infant dyads. Dev Psychobiol, 2017; 59(7):916–926; doi: 10.1002/dev.21557

43.

D'Anna-Hernandez

, Ross

, Natvig

, et al. Hair cortisol levels as a retrospective marker of hypothalamic-pituitary axis activity throughout pregnancy: Comparison to salivary cortisol. Physiol Behav, 2011; 104(2):348–353; doi: 10.1016/j.physbeh.2011.02.041

44.

Karlén

, Ludvigsson

, Hedmark

, et al. Early psychosocial exposures, hair cortisol levels, and disease risk. Pediatrics, 2015; 135(6):e1450–e1457; doi: 10.1542/peds.2014-2561

45.

Romero-Gonzalez

, Caparros-Gonzalez

, Gonzalez-Perez

, et al. I am pregnant. Am I different?“: Psychopathology, psychological stress and hair cortisol levels among pregnant and non-pregnant women. J Psychiatr Res, 2020; 131:235–243; doi: 10.1016/j.jpsychires.2020.09.023

46.

Kirschbaum

, Tietze

, Skoluda

, et al. Hair as a retrospective calendar of cortisol production—Increased cortisol incorporation into hair in the third trimester of pregnancy. Psychoneuroendocrinology, 2008; 34(1):32–37; doi: 10.1016/j.psyneuen.2008.08.024

47.

Horan

, Cheyney

, Torres

, et al. Maternal hair cortisol concentrations across pregnancy and the early postpartum period in a Puerto Rican sample. Am J Hum Biol, 2022; 34(11):e23718-n/a; doi: 10.1002/ajhb.23718

48.

García-León

MÁ

, Caparrós-González

, Romero-González

, et al. Resilience as a protective factor in pregnancy and puerperium: Its relationship with the psychological state, and with Hair Cortisol Concentrations. Midwifery, 2019; 75:138–145; doi: 10.1016/j.midw.2019.05.006

49.

Orta

, Tworoger

, Terry

, et al. Stress and hair cortisol concentrations from preconception to the third trimester. Stress (Amsterdam, Netherlands), 2019; 22(1):60–69; doi: 10.1080/10253890.2018.1504917

50.

Caparros-Gonzalez

, Romero-Gonzalez

, Strivens-Vilchez

, et al. Hair cortisol levels, psychological stress and psychopathological symptoms as predictors of postpartum depression. PLoS One, 2017; 12(8):e0182817–e0182817; doi: 10.1371/journal.pone.0182817

51.

Lang

, Stickel

, Gaum

, et al. Predicting hair cortisol and cortisone concentration in postpartum women through repeated measurements of perceived stress. Metabolites, 2021; 11(12):815; doi: 10.3390/metabo11120815

52.

Musana

, Cohen

, Kuppermann

, et al. Association of differential symptoms of stress to hair cortisol and cortisone concentrations among pregnant women in Kenya. Stress, 2020; 23(5):556–566; doi: 10.1080/10253890.2019.1696305

53.

Abbassi-Ghanavati

, Greer

, Cunningham

. Pregnancy and laboratory studies: A reference table for clinicians. Obstet Gynecol, 2009; 114(6):1326–1331; doi: 10.1097/AOG.0b013e3181c2bde8

54.

Obel

, Hedegaard

, Henriksen

, et al. Stress and salivary cortisol during pregnancy. Psychoneuroendocrinology, 2005; 30(7):647–656; doi: 10.1016/j.psyneuen.2004.11.006

55.

Henley

, Lowthers

, Koren

, et al. Cultural and socio-economic conditions as factors contributing to chronic stress in sub-Saharan African communities. Can J Physiol Pharmacol, 2014; 92(9):725–732; doi: 10.1139/cjpp-2014-0035

56.

Afulani

, Ongeri

, Kinyua

, et al. Psychological and physiological stress and burnout among maternity providers in a rural county in Kenya: Individual and situational predictors. BMC Public Health, 2021; 21(1):453–453; doi: 10.1186/s12889-021-10453-0

57.

Kalokhe

, Ibegbu

, Kaur

, et al. Intimate partner violence is associated with increased CD4. Pathog Immun, 2016; 1(1):193–213; doi: 10.20411/pai.v1i1.120

58.

Cohen

, Janicki-Deverts

, Doyle

, et al. Chronic stress, glucocorticoid receptor resistance, inflammation, and disease risk. Proc Natl Acad Sci U S A, 2012; 109(16):5995–5999; doi: 10.1073/pnas.1118355109

59.

Dettmer

, Rosenberg

, Suomi

, et al. Associations between parity, hair hormone profiles during pregnancy and lactation, and infant development in rhesus monkeys (Macaca mulatta). PLoS One, 2015; 10(7):e0131692; doi: 10.1371/journal.pone.0131692

60.

Dettmer

, Rosenberg

, Menard

, et al. Differential relationships between chronic hormone profiles in pregnancy and maternal investment in rhesus monkey mothers with hair loss in the neonatal period. Am J Primatol, 2017; 79(1):1–8; doi: 10.1002/ajp.22489

61.

Kapoor

, Lubach

, Ziegler

, et al. Hormone levels in neonatal hair reflect prior maternal stress exposure during pregnancy. Psychoneuroendocrinology, 2016; 66:111–117; doi: 10.1016/j.psyneuen.2016.01.010

62.

Van den Bergh

BRH

, van den Heuvel

, Lahti

, et al. Prenatal developmental origins of behavior and mental health: The influence of maternal stress in pregnancy. Neurosci Biobehav Rev, 2020; 117:26–64; doi: 10.1016/j.neubiorev.2017.07.003

63.

Romero-Gonzalez

, Caparros-Gonzalez

, Gonzalez-Perez

, et al. Newborn infants' hair cortisol levels reflect chronic maternal stress during pregnancy. PLoS One, 2018; 13(7):e0200279; doi: 10.1371/journal.pone.0200279

64.

Witt

, Cheng

, Wisk

, et al. Maternal stressful life events prior to conception and the impact on infant birth weight in the United States. Am J Public Health, 2014; 104 (Suppl 1):S81–S89; doi: 10.2105/AJPH.2013.301544

65.

Geiss

, Varadi

, Steinbach

, et al. Psychoneuroimmunological correlates of persisting sciatic pain in patients who underwent discectomy. Neurosci Lett, 1997; 237(2–3):65–68; doi: 10.1016/s0304-3940(97)00810-0

66.

Barber

, Coyle

, Fischer

, et al. Influence of hypercortisolemia on soluble tumor necrosis factor receptor II and interleukin-1 receptor antagonist responses to endotoxin in human beings. Surgery, 1995; 118(2):406–410; discussion 410–411; doi: 10.1016/s0039-6060(05)80352-6

67.

Giurgescu

. Are maternal cortisol levels related to preterm birth?. J Obstet Gynecol Neonatal Nurs, 2009; 38(4):377–390; doi: 10.1111/j.1552-6909.2009.01034.x

68.

Elenkov

. Glucocorticoids and the Th1/Th2 balance. Ann N Y Acad Sci, 2004; 1024:138–146; doi: 10.1196/annals.1321.010

69.

Gennaro

, Hennessy

. Psychological and physiological stress: Impact on preterm birth. J Obstet Gynecol Neonatal Nurs, 2003; 32(5):668–675; doi: 10.1177/0884217503257484

70.

Ruiz

, Fullerton

, Dudley

. The interrelationship of maternal stress, endocrine factors and inflammation on gestational length. Obstet Gynecol Surv, 2003; 58(6):415–428; doi: 10.1097/01.OGX.0000071160.26072.DE

71.

Burdet

, Rubio

, Salazar

, et al. Inflammation, infection and preterm birth. Curr Pharm Des, 2014; 20(29):4741–4748; doi: 10.2174/1381612820666140130202224

72.

Stalder

, Steudte-Schmiedgen

, Alexander

, et al. Stress-related and basic determinants of hair cortisol in humans: A meta-analysis. Psychoneuroendocrinology, 2017; 77:261–274; doi: 10.1016/j.psyneuen.2016.12.017

73.

Swann

, King

, Côté

HCF

, et al. Stressing the need for validated measures of cortisol in HIV research: A scoping review. HIV Med, 2022; 23(8):880–894; doi: 10.1111/hiv.13272

74.

Nwogu

, Babalola

, Ngene

, et al. Willingness to donate hair samples for research among people living with HIV/AIDS attending a tertiary health facility in Ibadan, Nigeria. AIDS Res Hum Retroviruses, 2019; 35(7):642–648; doi: 10.1089/AID.2018.0242