Hair cortisol concentration (HCC) as a measure for prenatal psychological distress — A systematic review
Introduction
Starting from the Barker hypothesis (Barker, 1986) on the developmental origins of adult disease, the importance of fetal environment for human development has been increasingly recognized (O’Donnell and Meaney, 2017). The characteristics of prenatal environment prepare the fetus for postnatal circumstances but the impact may also be maladaptive, ultimately leading to negative developmental outcomes in the offspring. One of the environmental factors gaining increasing interest is exposure to maternal prenatal psychological distress (PD), a heterogenic concept that comprises varying types of maternal distress, such as symptoms of depression or anxiety, experiences of stress related to either pregnancy itself or to everyday life situations and major life events (Dunkel Schetter, 2011; Scheinost et al., 2017). The topic is clinically very relevant as it has been estimated that the effects of prenatal PD could explain up to 17% of the variance in childhood cognitive abilities (Bergman et al., 2007) and that exposure to prenatal anxiety may double the prevalence of child psychiatric disorders (O’Donnell et al., 2014). Although PD has sometimes been associated with accelerated development of the offspring (e.g. DiPietro et al., 2006; Li et al., 2013), it has more consistently been linked to impaired neurological and psychosocial development (for a review, see van den Bergh et al., 2017; Capron et al., 2015; Grace et al., 2016; Pearson et al., 2016; Rijlaarsdam et al., 2017). Thus, prenatal PD is an important target for focused prevention and intervention programs (Glover, 2014), and identification of phenotypes with the greatest risk to affect the fetal programming is essential.
Cortisol, the hormonal end-point of the main human stress regulation system hypothalamic-pituitary-adrenal (HPA) axis, has been suggested to have a significant role in mediating the effects of maternal stress on the fetus. In general, increased maternal prenatal glucocorticoid levels as measured by maternal salivary, blood or urine cortisol concentrations have been linked with similar child developmental outcomes as prenatal PD (Braun et al., 2013; Moisiadis and Matthews, 2014; Painter et al., 2012). Elevated maternal prenatal cortisol concentrations are reportedly associated with compromised cognitive and motor development, affective problems, blunted cortisol reaction to stress, and alterations in regional brain volumes and connectivity in children (Buss et al., 2012; Davis and Sandman, 2012; Huizink et al., 2003; Kim et al., 2017; O’Connor et al., 2013). However, the data are inconsistent (Zijlmans et al., 2015), and also positive associations between maternal cortisol concentration and child cognitive performance have been reported (Davis et al., 2017).
Associations between maternal cortisol concentrations and prenatal PD are also inconsistent, partially due to the biological heterogeneity of and variety between the phenotypes of PD. Maternal symptoms of pre- and postnatal depression were correlated with maternal salivary, blood or urine cortisol concentrations in only 24 out of 47 studies assessing their associations (Seth et al., 2016). Some studies have reported increased prenatal maternal salivary cortisol concentrations in the context of pregnancy-related anxiety (Kane et al., 2014; Obel et al., 2005). Interestingly, maternal exposure to traumatic events during pregnancy has been associated with lower maternal plasma cortisol concentration (Perroud et al., 2014). However, assessing long-term cortisol levels with momentary measurements is challenging and multiple sampling is varyingly applied (Seth et al., 2016; Short et al., 2016).
One factor potentially producing variance in these findings is that maternal HPA axis functioning alters significantly during the normal course of pregnancy (see Fig. 1; Benediktsson et al., 1997; Challis et al., 2001; Petraglia et al., 1992). It is known that the physiological 2–3-fold increase in maternal cortisol levels towards the end of pregnancy (Jung et al., 2011) is essential to the maturation of several organ systems of the fetus and plays an important role in initiating parturition (Moisiadis and Matthews, 2014). Thus, the timing of assessments is of special importance during pregnancy and findings in a given trimester cannot necessarily be generalized to other trimesters (Kane et al., 2014). On the other hand, HPA axis overall reactivity is attenuated during pregnancy (de Weerth and Buitelaar, 2005; Schulte et al., 1990), further illustrating the complexity of the picture.
Importantly, maternal cortisol is not the only mechanism relating prenatal PD to offspring outcomes. The importance of placental functioning − specifically, the placental enzyme 11β-hydroxysteroid dehydrogenase type 2 (11β-HSD2) in converting cortisol into inactive metabolites − is gaining support from recent evidence (Janssen et al., 2016). Other mechanisms that could mediate the fetal programming effects of PD include direct effects of placental CRH (pCRH), changes in maternal immune system functioning, gut microbiota composition, serotonin levels and epigenetic changes as well as maternal lifestyle during pregnancy (Abbott et al., 2018; Beijers et al., 2014; Elwenspoek et al., 2017; Glover, 2015; Howland et al., 2016; Karlén et al., 2015; Zijlmans et al., 2015).
During the past few years, hair cortisol concentration (HCC) has been presented as a method to assess long-term levels of cortisol (Stalder and Kirschbaum, 2012). Cortisol is accumulated into hair as it grows and thus, with the generally accepted average hair growth rate of one centimeter per month (Wennig, 2000), one or several segments of selected length can be analyzed for the mean levels of cortisol during the corresponding months. Instead of the more traditional short-term measurements of cortisol assessing either the reactivity or diurnal profile of cortisol by repetitive samples of saliva or blood (Adam and Kumari, 2009; Vining et al., 1983), HCC provides a possibility to measure retrospectively cumulative cortisol levels of previous months to gain a more complete picture of the mean cortisol levels during the chosen period, via a single sample (Davenport et al., 2006; Kirschbaum et al., 2009; Raul et al., 2004). Although the comparison of methods depicting different aspects of the HPA axis is not straightforward, salivary cortisol and HCC do correspond when the saliva sampling protocol adequately models the long-term fluctuation of cortisol levels with a minimum of three saliva samples during consecutive days (D’Anna-Hernandez et al., 2011; Short et al., 2016).
When evaluating studies on HCC, it should be noted that there is methodological variation in the laboratory analysis protocols and analytical methods (e.g. utilization of either immunoassay or mass spectrometry techniques) leading to subsequent inter-laboratory variance in the absolute HCC concentrations (Stalder et al., 2017). In addition to the methodological factors, several characteristics of study subjects and hair samples, such as age, sex, hair washing frequency, hair treatments and oral contraceptive use have been identified as significant determinants of HCC (Stalder et al., 2017). The selected segment length is relevant as cortisol concentrations have been shown to decrease along the hair shaft (wash-out effect) up to 30–40% from a 3 centimeter segment to the next (Kirschbaum et al., 2009). The components contributing to this wash-out or degradation of cortisol include hair washing and hair treatments (Hoffman et al., 2014) as well as exposure to sunlight (Wester et al., 2016).
To date, HCC has been reported to be useful in responding to various research questions across different study populations (Chau et al., 2015; Steudte-Schmiedgen et al., 2017; Wright et al., 2015; Yamada et al., 2007). The meta-analysis by Stalder et al. (2017) covering 66 independent HCC studies concluded that HCC was elevated in subjects with ongoing exposure to a chronic stressor such as caregiving stress, unemployment or natural disaster and declined in patients with an anxiety disorder. Self-report measures of perceived stress, social support or depressive symptoms were not associated with HCC (Stalder et al., 2017).
Several studies (D’Anna-Hernandez et al., 2011; Hoffman et al., 2017; Karlén et al., 2013; Kirschbaum et al., 2009; Schreier et al., 2016; Smy et al., 2015; Wikenius et al., 2016) have assessed the change in HCC levels during pregnancy. The physiological increase in cortisol levels was evident in all these studies. This is consistent with previous data based on salivary cortisol measurements (Kirschbaum and Hellhammer, 1994), which lends support for the suitability of HCC as a measure for cortisol concentrations also during pregnancy.
So far, studies using HCC during pregnancy have not been included in the reviews considering the mechanisms of prenatal PD or its outcomes (Seth et al., 2016; Zijlmans et al., 2015). As both these reviews conclude that the comparability and accuracy of short-term cortisol measurements are a major limitation in the studies (Seth et al., 2016; Zijlmans et al., 2015), HCC could serve as a potential means in elucidating the phenomena. Because of potential programming effects on fetal development, assessing long-term levels of maternal prenatal cortisol is of special interest.
The aim of this systematic review is to explore the existing literature on the associations between HCC and maternal prenatal PD.
Section snippets
Methods
To identify all relevant studies, we conducted a systematic search (see Fig. 2). The databases PUBMED, Ovid MEDLINE®, Embase, PsycINFO, WorldCat and WEB of SCIENCE were used for a literature search (latest update December 26, 2017). The following search terms were used: (hair cortisol OR (hair AND cortisol)) AND (pregnancy OR prenatal OR antenatal OR gestational). The inclusion criteria were 1. original human studies 2. written in English 3. assessing maternal HCC and 4. maternal PD 5. during
Results
Here, we review the results of the identified studies. We start with an overview of study characteristics of the identified papers. Then, we present results on the associations between different types of prenatal PD and HCC.
Discussion
Our aim was to systematically review studies that investigated the associations between maternal prenatal self-reported symptoms of different types of prenatal PD and HCC across pregnancy. To sum up, the existing data imply that in a population with low levels of self-reported prenatal PD, the associations with HCC are rather weak or can be seen only or most clearly in mid-pregnancy. However, this association might be strengthened when studying populations with a greater variance in PD and
Conclusions
As the number of studies of meeting the criteria of this review was low, it is difficult to draw definite conclusions on the role of HCC as a measure for PD based on these data alone. The incongruity of the existing results is likely related to both the complexity of the prenatal PD phenomenon itself and the variation in the study protocols and methods in the current database. In fact, considering all the aspects on which HCC differs from the self-reported questionnaires assessing prenatal PD,
Role of the funding source
The preparation of the review has been financially supported by Academy of Finland (grant # 134950 to HK and grant # 308176 to LK), Finnish Brain Foundation (PM), Jane and Aatos Erkko Foundation (HK), Päivikki and Sakari Sohlberg Foundation (PM), Signe and Ane Gyllenberg Foundation (LK, HK, NMS), State Research Funding (LK, NMS, HK), Yrjö Jahnsson Foundation (PM, LK), and Finnish Cultural Foundation (PM). AJR has been funded by the project POCI-01-0145-FEDER-016428 and BC by grant
Conflicts of interest
The authors declare no conflicts of interest.
Authors’ contributions
The design of the review was conceived by PM, LK, AJR, NMS and HK. PM conducted the identification of the papers for the review and prepared the draft and was responsible for writing and revising the manuscript. LK, AJR, NMS, SK, BC and HK contributed in the interpretation of the papers and advised on the manuscript. All authors contributed to writing and critical revisions and approved the final version of the manuscript.
References (82)
- et al.
Prenatal stress and genetic risk: how prenatal stress interacts with genetics to alter risk for psychiatric illness
Psychoneuroendocrinology
(2018) - et al.
Assessing salivary cortisol in large-scale, epidemiological research
Psychoneuroendocrinology
(2009) Infant mortality, childhood nutrition, and ischaemic heart disease in England and Wales
Lancet
(1986)- et al.
Maternal stress during pregnancy predicts cognitive ability and fearfulness in infancy
J. Am. Acad. Child Adolesc. Psychiatry
(2007) - et al.
Determinants of maternal hair cortisol concentrations at delivery reflecting the last trimester of pregnancy
Psychoneuroendocrinology
(2015) - et al.
Associations of maternal and paternal antenatal mood with offspring anxiety disorder at age 18 years
J. Affect. Disord.
(2015) - et al.
The behavioural, cognitive, and neural corollaries of blunted cardiovascular and cortisol reactions to acute psychological stress
Neurosci. Biobehav. Rev.
(2017) - et al.
The fetal placental hypothalamic-pituitary-adrenal (HPA) axis, parturition and post natal health
Mol. Cell. Endocrinol.
(2001) - et al.
Hair cortisol levels as a retrospective marker of hypothalamic-pituitary axis activity throughout pregnancy: comparison to salivary cortisol
Physiol. Behav.
(2011) - et al.
The three-hit concept of vulnerability and resilience: toward understanding adaptation to early-life adversity outcome
Psychoneuroendocrinology
(2013)