Comparative steroid profiling of newborn hair and umbilical cord serum highlights the role of fetal adrenals, placenta, and pregnancy outcomes in fetal steroid metabolism

Previous steroid hormone studies concerning pregnancy and newborns have mainly focused on glucocorticoids; wider steroid profiles have been less commonly investigated. Here, we performed a comparative analysis of 17 steroids from newborn hair and umbilical cord serum at the time of delivery. The study participants (n = 42, 50% girls) were a part of the Kuopio Birth Cohort and represent usual Finnish pregnancies. The hair and cord serum samples were analyzed with liquid chromatography high resolution mass spectrometry and triple quadrupole tandem mass spectrometry, respectively. We detected high individual variations in steroid hormone concentrations in both sample matrices. The concentrations of cortisol (F), corticosterone (B), estrone (E1), estradiol (E2), dehydroepiandrosterone (DHEA), 11 β -hydroxyandostenedione (11bOHA4), 5 α -androstanedione (DHA4), and 17 α -hydroxypregnenolone (17OHP5) correlated positively between cord serum and newborn hair samples. In addition, F and 11bOHA4 concentrations correlated positively with each other in both newborn hair and cord serum samples. The cortisone-to-cortisol ratio (E/F) was significantly higher in cord serum than in newborn hair samples reflecting high placental 11 β HSD2 enzyme activity. Only minor sex differences in steroid concentrations were observed; higher testosterone (T) and 11-deoxycortisol (S) with lower 11bOHA4 in male cord serum, and higher DHEA, androstenedione (A4) and 11bOHA4 in female newborn hair samples. Parity and delivery mode were the most significant pregnancy-and birth-related parameters associating with F and some other adreno-cortical steroid concentrations. This study provides novel information about intrauterine steroid metabolism in late pregnancy and typical concentration ranges for several newborn hair steroids, including also 11-oxygenated androgens.


Introduction
The presence of steroid hormones is essential for gestation and fetal development, and their secretion increases throughout pregnancy [1,2].
The analysis of steroid hormones from human scalp hair has become more popular during pregnancy because hair analysis can provide valuable insights into long-term endogenous steroid metabolism. Especially glucocorticoid (F and cortisone [E]) measurements during gestation and after pregnancy have been performed widely from maternal hair [10,11]. Some investigators have analyzed also the glucocorticoids present in newborn hair [12][13][14]. In addition, Kapoor et al. [15] examined the steroid milieu by analyzing eight different steroids, including F, E, estradiol (E2), estrone (E1), testosterone (T), dihydrotestosterone (DHT), dehydroepiandrosterone (DHEA) and P4, in newborn and maternal monkey hair concluding that newborn hair can be useful as a way of assessing fetal in utero hormone exposure. The newborn hair present at birth is thought to reflect the steroid metabolism occurring during the last trimester of pregnancy (after the 28th week of gestation) [16]. A hair sample from a newborn represents a unique matrix and, unlike maternal hair, it has not been exposed to environmental substances. While during the last decade the role of 11-oxygenated androgens such as 11β-hydroxyandrostenedione (11bOHA4), 11-ketoandrostenedione (11KA4) and 11-ketotestosterone (11KT) has also attracted more interest in human physiology and pathology [17], their role in fetal steroidogenesis remains still largely unclear [18].
In our recently published study, we quantified a large panel of steroids from human scalp hair comprising glucocorticoids, progestogens, estrogens, and androgens including also 11-oxygenated androgens [19]. In this present study, we have investigated the newborn hair steroid milieu by assaying the concentrations of 17 steroids in newborn hair samples at the time of delivery. One aim was to determine if there were correlations between steroid profiles in newborn hair and umbilical cord serum. In addition, we analyzed the association of pregnancy-and birth-related parameters with newborn steroid profiles. We believe that this study advances our understanding of fetal steroid metabolism and the usability of newborn hair as a sample matrix in steroid hormone analysis in pregnancy research.

Subjects and sample collection
The hair and umbilical cord serum samples analyzed in this study were from the Kuopio Birth Cohort (KuBiCo) [20]. The participants included in this study (n = 42, 50% girls) represent usual Finnish pregnancies and there were no strict exclusion criteria. A more detailed description of the whole study cohort has been reported previously [20]. The hair samples of the newborns (n = 42) were collected within the first 48 h after birth. The umbilical cord blood samples (n = 42) from umbilical vein were collected from the same newborns after birth, with the serum samples being frozen and stored at − 80 • C until analyzed. This study was conducted according to the ethical guidelines laid down in the Declaration of Helsinki and approved by the Ethics Committee of the Hospital District of Central Finland in Jyväskylä. All pregnant women provided informed consent during recruitment.
The scalp hair samples from newborns were cut from the posterior vertex and 3 cm of hair closest to the scalp was analyzed. If the hair length was less than 3 cm, the whole hair lock was used in the analysis. The hair samples were stored in the dark, wrapped in aluminum foil at ambient temperature (+22 ± 2 •C) [19].

Quantitative steroid hormone analyses
The steroid hormone analysis of the newborn hair and umbilical cord serum samples was performed as described previously [19,21]. Briefly, 5-10 mg of newborn hair was weighed and washed with isopropanol. The samples were then extracted with methanol (1.5 ml) overnight at ambient temperature after addition of 20 µl of internal standard (IS) solution. The extracts were purified first with solid phase extraction (SPE) and subsequently with liquid-liquid extraction (LLE) with toluene. Finally, the samples were reconstituted in 100 µl of 100 mM hydroxylamine hydrochloride solution, heated at 60 • C for 30 min and analyzed by ultra-performance liquid chromatography high resolution mass spectrometry (UPLC-HRMS). The umbilical cord serum was prepared and analyzed as we described previously for plasma [21]. The samples were diluted (1:10, v/v) before LLE with toluene as follows: 15 µl of serum and 20 µl of IS were pipetted into 135 µl of NaCl solution (9 mg/ml, B. Braun) and extracted with 1 ml of toluene for 10 min. After the toluene phase had been dried under nitrogen, the samples were reconstituted in 50 µl of 100 mM hydroxylamine hydrochloride solution, heated at 60 • C for 30 min and analyzed by liquid chromatography triple quadrupole tandem mass spectrometry (LC-MS/MS). Standard and quality control (QC) samples for hair analyses were prepared similarly as the authentic samples, except that 40 µl of standard working solution was added to methanol for overnight extraction instead of hair. In the cord serum analyses, standards and QCs were also prepared similarly in 150 µl of NaCl (9 mg/ml, B. Braun) solution with 20 µl of standard working solution and 20 µl of IS with toluene extraction. In each analytical batch, we analyzed two calibration curves and 6-8 parallel QC samples at three different concentration levels.
Hair steroid analyses were performed with an ultra-highperformance liquid chromatograph (Vanquish Flex UHPLC system, Thermo Scientific, Bremen, Germany) connected online to a highresolution Orbitrap mass spectrometer (HRMS, Q Exactive Classic, Thermo Scientific, Bremen, Germany). Umbilical cord serum analyses were conducted with a liquid chromatograph (1290 Rapid Resolution LC system, Agilent Technologies, San Jose, CA, USA) coupled to a triple quadrupole (QQQ) mass spectrometer (model 6495 QQQ, Agilent Technologies, San Jose, CA, USA). In both methods, steroids were separated in an Acquity UPLC CSH C18 column, 1.7 µm, 100 × 2.1 mm, (Waters Corporation, Milford, MA, USA) with the same mobile phase composition being used: mobile phase A consisted of 0.2 mM ammonium fluoride (NH 4 F) in water and B was a mixture of 0.2 mM NH 4 F in methanol:water 95:5 (v/v). All steroids were measured as their hydroxylamine derivatives in the positive ionization mode except for E2 which was assessed as the free form in the negative ionization mode. These two analytical methods were utilized due to the high matrix background in hair analysis. With HRMS technology, we could achieve better selectivity and sensitivity with hair samples [19].

Statistical analyses
The statistical analyses were performed with IBM SPSS Statistics Version 27. Non-parametric tests were used due to the relatively small number of study participants (n = 42, 50% girls) and the non-normal distributions of several steroid concentrations. Mann-Whitney U test was used to compare differences of steroid concentrations between sexes and Wilcoxon signed rank test when comparing the steroid ratios between sample matrices. Correlation analyses between steroid concentrations in umbilical cord serum and newborn hair were performed with Spearman correlation test which was also used in the analysis of associations between pregnancy-and birth-related parameters (i.e., maternal age, parity, mode of delivery, gestational age at birth, sex, and birth weight) and steroid concentrations with unadjusted p-values and correlation coefficients being reported. In addition, linear regression models were applied to logarithmically transformed steroid data to identify if there were parameters associating independently with umbilical cord serum or newborn hair steroid concentrations. The results were considered statistically significant when p < 0.05.

Characteristics of the study participants
Selected pregnancy-related parameters and characteristics of the participants are shown in Table 1. The study participants represent normal Finnish pregnancies. There was an equal number of male and female newborns. Most of the newborns were born full-term with vaginal delivery (90%) and with normal median birth weight (3.7 kg). The median maternal age was 31 years (range 17 -44 years) and 45% of women were nulliparous and 55% primi-or multiparous.

Steroid concentrations and their correlations between newborn hair and umbilical cord serum
A wide panel of steroids including F, E, corticosterone (B), 11-deoxycortisol (S), pregnenolone (P5), E1, E2, DHEA, androstenedione (A4), T, 11bOHA4, 11KA4, 5α-androstanedione (DHA4), 17α-hydroxypregnenolone (17OHP5), 17α-hydroxyprogesterone (17OHP4), and 11-deoxycorticosterone (11DOC, also known as 21-hydroxyprogesterone), were successfully quantified from cord serum and newborn hair samples ( Table 2). There were extensive inter-individual variations in steroid concentrations in both sample matrices. F, E, P5, P4 and E1 belonged to the steroids with the highest concentrations in both sample matrices. However, the concentrations of P4 were above the upper limit of quantification (ULOQ) (1326 nmol/L in cord serum and 625 pg/mg in hair when measured from 5 mg of hair) while aldosterone (A) was below the lower limit of quantification (LLOQ) (3.2 nmol/L in cord serum and 6.9 pg/mg in hair when measured from 5 mg of hair) in both sample matrices. In addition, the concentrations of F, B, E1, E2, DHEA, 11bOHA4, DHA4, and 17OHP5 correlated positively between cord serum and newborn hair samples (Table 2). Statistically significant correlations were also evident between F and 11bOHA4 concentrations in both cord serum and newborn hair samples (r = 0.810, p < 0.001 and r = 0.645, p < 0.001, respectively).
Six androgens were quantified from cord serum and newborn hair samples. DHEA was the most abundant classical androgen, followed by A4, DHA4 and T in both sample matrices (Table 2), while DHT concentrations were below the LLOQ (0.33 nmol/L in cord serum and 0.6 pg/mg in hair when measured from 5 mg of hair). Of the 11-oxygenated androgens, 11bOHA4 and 11KA4 concentrations exceeded those of their precursor A4 in both cord serum and newborn hair, but the concentrations of the other 11-oxygenated androgens 11OHT, 11KT and 11KDHT were all below the LLOQ in both cord serum (0.33, 0.67 and 0.33 nmol/ L, respectively) and hair samples (1.2 pg/mg when measured from 5 mg of hair for the above mentioned compounds).
Of the measured estrogens, there were about 6-and 4-fold higher median concentrations of E1 than E2 in cord serum and newborn hair samples, respectively. It is noteworthy that the median concentrations of E1 and E2 were clearly higher than those of their immediate precursors A4 and T, respectively, in both cord serum and newborn hair ( Table 2). In newborn hair, P4 clearly displayed the highest median concentration of the measured progestogens, followed by P5, 17OHP4 and 17OHP5. In cord serum, the order of the median progestogen concentrations was similar to hair samples except that the level of 17OHP4 was slightly higher than that of P5 (Table 2).

Steroid ratios and their association with steroidogenic pathways
We evaluated indirectly steroidogenic pathways and enzymatic activities in the fetoplacental unit by calculating the ratios of specific steroids (product/precursor ratios) ( Table 3) as previously reported [22]. The steroid ratios reflecting "adrenal-specific" steroid transforming enzyme activities (11β-hydroxylase, 21-hydroxylase, 17-hydroxylase, 17,20-lyase) were significantly higher in the newborn hair samples and those ratios reflecting placentally active steroid conversions (aromatase, 11βHSD2, 3βHSD) were higher in the cord serum samples (Table 3). In addition, the E/F ratio correlated significantly with the 11KA4/11bOHA4 ratio in newborn hair (r = 0.849, p < 0.001), but no statistically significant correlation was detected in cord serum between these ratios.

Differences in steroid concentrations between sexes
Most of the measured steroids revealed no statistically significant concentration differences between male and female newborns in either sample type. Only five steroids (T, DHEA, A4, 11bOHA4, S) had a significant difference between sexes in either hair or umbilical cord serum samples ( Table 4). The concentrations of T were significantly higher in males (p < 0.001) in umbilical cord serum, but not in hair samples. DHEA, A4 and 11bOHA4 concentrations were higher in the hair samples of females, but only 11bOHA4 was slightly higher in cord serum. Higher S concentrations were detected in umbilical cord serum of males, but no significant difference was found in hair samples (Table 4).

The association of newborn steroid concentrations with pregnancyand birth-related parameters
The correlation analyses revealed several associations between both cord serum and newborn hair steroid concentrations and pregnancyand birth-related parameters (Fig. 1). Higher gestational age at birth was associated with higher F, B, P5, DHEA, A4, 11bOHA4, 11KA4, DHA4, 17OHP5 and 11DOC concentrations in cord serum, while similar associations were seen only with F, DHEA, 11bOHA4 and 11DOC concentrations in newborn hair samples (Fig. 1). Nulliparity was significantly associated with higher F, E, B, P5, E1, A4, 11bOHA4, DHA4, 17OHP4, 17OHP5, and 11DOC concentrations in cord serum as compared to primi-and multiparity. A similar association was also seen with F, B, 11bOHA4, and 11DOC concentrations in hair samples. The mode of delivery associated also with several steroid concentrations. Higher concentrations of F, E, B, P5, 11bOHA4 and 11DOC in cord serum were associated with normal or operational vaginal delivery when compared to samples from newborns delivered by elective Caesarean section. This association was also seen in F and 11bOHA4 concentrations in newborn hair samples. Maternal age exhibited a weak association with lower B, P5 and A4 concentrations in cord serum and lower F concentrations in newborn hair samples. Birth weight displayed no significant association with any steroid concentration in cord serum or hair samples (Fig. 1). When steroid concentrations were compared on a group-wise basis of the same gestation-and birth-related factors used above in the correlation analyses, similar types of associations were detected (data not shown).
When linear regression models were applied to logarithmically transformed steroid data to examine which pregnancy related parameters associated independently with steroid concentrations, it was noted that parity and delivery mode were associated with several steroids more often than the other parameters (Table 5). Based on these results, male sex was a significant predictor for higher S and T concentrations in cord serum and for lower DHEA and A4 concentrations in newborn hair samples. Gestational age was observed to be a significant predictor for higher 11KA4 concentrations in cord serum and lower 11KA4 concentrations in newborn hair. Nulliparity was a significant predictor for higher F, B, P5, 11bOHA4, and DHA4 concentrations in cord serum and for higher F, B, 11bOHA4, and 11KA4 concentrations in newborn hair samples when compared to primi-and multiparity. Birth by elective Caesarean section was a significant predictor for lower F, E, B, and P5 concentrations in cord serum and for lower F concentrations in newborn hair samples when compared to vaginal delivery. Birth weight and maternal age showed no significant associations with any steroid in Table 3 Steroid concentration ratios (reflecting steroid transforming enzymatic activities) calculated from umbilical cord serum and newborn hair samples (n = 42).

Ratio
Umbilical cord serum (nmol/L) Data are shown as median (range min-max). Wilcoxon signed rank test was used to compare the ratios between sample matrices. Differences between sample matrices are expressed as p-values. either of the sample matrices.

Discussion
Our recently developed method makes possible the analysis of 22 steroids from newborn scalp hair samples [19]. Here we analyzed scalp hair and umbilical cord serum samples from 42 mostly full-term newborns representing normal pregnancies and report the concentrations of 17 steroids from newborn hair and cord serum samples. To the best of our knowledge, this is the first study in which newborn hair and cord serum steroid profiles have been compared with such a wide steroid panel including also 11-oxygenated androgens. Previously reference ranges have been reported for hair F concentrations in children and newborns [12,14]. Our results are consistent with previous data showing higher F than E levels in newborn hair [13,23,24] which is the reverse of hair samples taken from older children and adults [19,25,26]. We also found that there were positive correlations of several steroid concentrations (F, B, E1, E2, DHEA, 11bOHA4, DHA4, and 17OHP5) between newborn hair and cord serum suggesting that the steroid profile in newborn hair reflects the steroid concentrations in the fetal circulation during late gestation. However, we observed that there were no correlations between newborn hair and cord serum samples with respect to the concentrations of those steroids which are known substrates (e.g., P5, A4, T) for or end products of (e.g., E) active placental steroid transforming enzymes.
In addition to hair sampling, cord blood can be drawn non-invasively and the separated serum used for steroid analysis and evaluation of fetoplacental hormonal milieu [13]. Even though there was high inter-individual variation in our cord serum steroid concentrations, the concentrations of F, E, B, S, P4, E1, E2, T, A4, 17OHP4, and 11DOC were consistent with previously published data [8,22,27]. However, we measured higher median concentrations of E1 from cord serum than reported by Banker et al. [27]. As expected, P4 was the steroid which displayed the highest concentrations of those measured in cord serum and newborn hair, exceeding the ULOQ in our assays. This fits well with the role of placental P4 in maintaining pregnancy until delivery [2] and serving as a substrate for further steroid synthesis [28]. There is both in vitro [29,30] and in vivo [31] evidence for a significant role of placentally produced P4 as a precursor for fetal adrenal F synthesis bypassing the inactive 3βHSD step in the fetal zone [32]. This might also explain the high F concentrations that were measured in newborn hair samples. In addition to hair and cord blood, newborn´s urine has been used to study neonatal steroid metabolism due to its easy collection and non-invasive nature [33,34].
As far as we are aware, 11-oxygenated androgen concentrations in newborn hair samples have not previously been analyzed. We were able to quantitate 11bOHA4 and 11KA4 in both newborn hair and cord serum samples. The concentrations of both 11bOHA4 and 11KA4 were higher than those of their precursor A4 in both matrices reflecting high Fig. 1. Correlations of steroid concentrations in umbilical cord serum and newborn hair with pregnancy-and birth-related parameters. Delivery mode represents either vaginal and operational delivery or elective Caesarean section which is negatively associated with several steroid concentrations. Correlations are expressed as Spearman correlation coefficients. The correlations are depicted by a color spectrum, red color indicating a positive correlation and blue color indicating a negative correlation. In addition, the size of the circle reflects the strength of the correlation. Significance of the correlations are expressed as follows: *** p < 0.001, ** p < 0.01, and * p < 0.05.

Table 5
Multiple linear regression results showing those pregnancy-and birth related parameters that independently associated with steroid concentrations.

Parameter
Umbilical cord serum Results calculated from log transformed steroid concentrations. F* = two outliers with concentrations >ULOQ in newborn hair samples were excluded from the analysis (adrenal) 11-hydroxylase and (placental) 11βHSD2 activities in the fetoplacental unit. Our results also found evidence of high (placental) aromatase activity which limits the exposure of the fetus to androgens [35]. In newborn hair samples, the 11bOHA4 concentration was slightly higher than that of 11KA4, but in cord serum samples this difference was not evident, which may be due to the high placental 11βHSD2 activity increasing the production of 11KA4 from 11bOHA4. du Toit et al. [18] analyzed 11-oxygenated steroids in newborn serum with samples taken 0-1 day and 2 days after birth. They found that 11bOHA4 was the most abundant androgen in newborn serum in both sexes and the concentrations of 11bOHA4 were higher than those of A4, T and 11KA4 [18]. However, He et al. [17] recently reported 11KA4 to be the dominant 11-oxyandrogen in cord serum, exceeding the concentrations of 11bOHA4, A4, T, 11KT, and 11OHT. Consistent with our study, He et al. [17] encountered difficulties in the analysis of the low 11KT and 11OHT concentrations in cord serum samples (only 23% and 15% of the measured concentrations were over the LLOQ, respectively). They also suggested that placental 17βHSD2 has a major role in protecting the fetus by catabolizing maternal androgens, such as 11KT, resulting in higher 11KA4 concentrations [17]. In addition, the steroid concentration ratios E/F, F/S, and 17OHP4/S (reflecting 11βHSD2, 11-hydroxylase and 21-hydroxylase activities, respectively) in our cord serum samples were similar to previously reported data [22]. The E/F ratio correlated positively also with the 11KA4/11bOHA4 ratio in newborn hair, reflecting high 11βHSD2 activity. Most steroid concentrations did not differ between the sexes in either hair or umbilical cord serum samples. Therefore, the detailed steroid concentrations ( Table 2) are shown as combined data from both sexes. The most significant difference between the sexes was the higher T concentration in cord serum of boys which is most likely explained by testicular T secretion, a finding consistent with previous reports [8,22]. However, this sex difference was not evident in the hair samples. On the other hand, A4 and DHEA were measured at slightly higher concentrations in the hair of girls but there was no similar sex difference detected in cord serum. In accordance with our results, Allvin et al. [8] did not find any differences in A4 concentrations between sexes in cord serum from healthy newborns. In our previous [19] and this present study, high F concentrations were observed in neonatal hair samples without a significant difference between sexes. This is in accordance with previous studies [12,14].
When examining the association of gestation-and birth-related parameters with newborn hair or cord serum steroid concentrations by correlation analyses, subgroup comparisons or regression models, it was noted that gestational age, parity, and the mode of delivery revealed the most significant associations. In our study, these associations seemed to be slightly more common with cord serum than newborn hair steroid concentrations. Similar associations with cord blood glucocorticoids were recently reported by Banker et al. [27], who described a positive association of cord blood F concentration with gestational age and delivery mode. In our newborn hair analyses, the F concentrations associated positively with gestational age and negatively with multiparity and birth by elective Caesarean section. Similar associations have been reported previously [12,13,24]. The newborns in our study were mostly full-term and born with normal median birth weights, which might explain the fewer associations between newborn parameters and steroid concentrations.
The role of amniotic fluid as a source of steroids incorporated into newborn hair remains to be clarified. In contrast to the hair samples from older children and adults, newborn hair has been exposed to "external" steroids present in the amniotic fluid. Even though the hair samples are washed during sample preparation to remove any external steroid contamination from the surface of the hair [19], this will not exclude the possibility of steroid absorption from amniotic fluid into the fetal hair matrix. The composition of amniotic fluid changes during gestation due to fetal development [36]. It is known that the steroid concentrations in amniotic fluid reflect those in the fetal circulation during embryogenesis but after the fetal skin keratinization, amniotic fluid composition is affected by fetal swallowing, urination, and steroid transport through the surfaces of amnion, placenta, and umbilical cord [36,37]. Partsch et al. [38] detected a higher F concentration in umbilical arterial than venous blood already at early mid-gestation suggesting the presence of active fetal adrenal F production. In mid-gestational amniotic fluid, the concentration of E was reported to be higher than that of F [37], but at the time of delivery, the level of F was slightly higher than that of E [38,39]. When our newborn hair and umbilical cord serum steroid profiles are compared to the above mentioned umbilical arterial/venous and amniotic fluid steroid profiles, it strengthens the view that newborn hair might be influenced also by the amniotic fluid steroid milieu near to parturition. In addition, there is increasing indirect evidence that acute increases in the F concentrations in the fetal circulation during normal vaginal labor or extra birth-related stress are reflected in newborn hair F concentrations [13,24]. This strongly indicates that increased fetal F concentrations due to normal or extra birth-related stress could be reflected in newborn hair due to absorption of these compounds from the amniotic fluid.
This study provides novel results of newborn hair steroids. The strength of the study is the extensive steroid panel including 11-oxygenated androgens measured from both newborn hair and cord serum samples originating from the same individuals. However, our study has some limitations. It is an observational study with a relatively small sample size (n = 42) which reduces the power of the statistical analyses. The umbilical cord serum analyzed was separated from venous cord blood representing the circulation and steroid passage from the placenta to the fetus. Serum samples separated from umbilical cord arterial [31] or peripheral venous blood might have given a steroid profile closer to that of newborn hair than that found in umbilical cord venous samples. However, practical and ethical reasons complicate this kind of sampling for research purposes. In general, hair steroid concentrations represent the steroid milieu of the individual during the last few weeks or months, depending on the length of the hair bundle analyzed. However, it is somewhat unclear what actual time window the hair samples collected from newborns represent. In future studies, the analysis of the same steroid panel from amniotic fluid samples in late pregnancy could give additional information about fetal steroid metabolism. There are also some limitations concerning the interpretation of the product-to-substrate ratios of steroids rather than directly assessing the tissue-specific activities of steroid transforming enzymes. Due to the large steroid network and interconnecting pathways, steroids are at the same time substrates for and products of enzymatic reactions [22]. Thus, these steroid ratios as a measure of enzymatic activities should be cautiously interpreted.

Conclusions
The determination of multiple steroids simultaneously provides a more precise view of the hormonal metabolism compared to detection of a single hormone. To the best of our knowledge, the current study is the first to report such a wide panel of steroid concentrations from newborn hair samples and their correlations to umbilical cord serum samples, providing novel information about fetal and placental steroid metabolism. Even though newborn hair seems to be influenced by steroids present in amniotic fluid, newborn hair samples could serve as a source of information when studying the physiology of pregnancy or the etiology of pregnancy complications. As Banker et al. [27] also concluded, alterations in the steroid milieu during pregnancy can affect fetal outcomes and serve as biomarkers for these changes. Collection of newborn hair samples is easier than that of amniotic fluid at the time of delivery, which makes it a useful sample matrix for examining the metabolic changes occurring during late pregnancy. In the future, the analysis of larger study populations including healthy and complicated pregnancies could give new information on whether the assessments of steroids from newborn hair could be used as biomarkers for later outcomes and metabolic diseases.

Declaration of Competing Interest
None.

Data Availability
Data will be made available on request.