Here we report a casuistry of 15 stillbirths and 46 newborns, used as cohort of interest and comparisons, respectively. As expected, stillbirths and newborn did not significantly differ except for the gestational age and weight at birth, since the included stillbirths all occurred in the third trimester, while the included newborns were closer or at the physiological term of the pregnancy.

The absence of significant associations between the cohort/comparison and antepartum clinical data, labor parameters or neonatal outcomes suggests that there are no spurious factors that could affect the biochemical analyses, except for those shown in Table 1, i.e. gestational age, mode of delivery with emergency CS and birth weight.

Serum and urinary CA have been studied in adult humans as biomarkers of cause of death and agonal time for forensic purposes, showing sometimes contradictory results [28,29,30]. Indeed, the interpretation of biochemical analyses in the post-mortem setting could be affected by interference due to pre-existing factors disorders during life, cause of death, survival period, post-mortem changes (e.g., autolysis and protein degradation), environmental factors, chemical properties, distributions, and locations of the analytes and analytical procedures [31, 32]. The major drawback in the dosage of CA has been reported as the poor stability at room temperature and the degradation connected to post-mortem changes [32, 33].

In the present study, CA in blood of stillbirths were significantly higher than in blood of newborns and exceeded clinical reference intervals, requiring dilutions. This seems to indicate a stability or a significant release of adrenaline and noradrenaline in the peri-mortem period, rather than a degradation. A stability of CA up to 60 h of storage in a cool environment was shown by Berg and Bonte [34]. In another study, adrenaline showed a mild tendency toward post-mortem increase, but this was not observed for cardiac blood and was not confirmed by further research [29, 35]. A study on 11 human cadavers detected a significant post-mortem increase of noradrenaline in heart blood from the first to the third day postmortem [35]. However, in our casuistry with comparable sample size, this was not confirmed when testing levels of CA with different interval between delivery and autopsy (Table 3).

Our results are consistent with the findings of Zhu et al. [29], who detected no significant changes in CA levels in cadaveric blood samples at ambient temperature and showed reproducible measurements for frozen samples. The findings are also consistent with those of Lee et al. [36], who highlighted no blood CA differences with various post-mortem intervals. Moreover, they seem to confirm the trials performed on cadaveric and healthy volunteers’ blood at room and refrigerated temperature, that demonstrated a reliable CA determination up to 72 h, despite a moderate decline [30].

If an increase in the delivery-autopsy interval is not explanatory of the findings, then the higher levels of CA detected in stillbirths in our casuistry might be connected to the processes leading to death, in the absence of other possible explanatory factors at the post-mortem examination.

Stillbirths in the third trimester mostly involve acidosis, systemic hypoxia, and ultimately an asphyxia mechanism. Some studies on adult humans did not find a distinct CA profile in relationship with the cause of death [30, 37], while others have shown a significantly higher amount of adrenaline and noradrenaline in deaths due to injuries, hyperthermia, poisonings and asphyxiation [29]. Lee et al. found higher femoral noradrenaline in asphyxia compared to deaths due to injuries [36]. In fatal asphyxia, systemic hypoxia and acidosis were considered major triggers for the release of CA, occurring from sympathetic nerve terminals in the extremities. The peripheral release explained why, in asphyxiation, CA blood levels from the external iliac veins were higher than those from the left heart [29].

Our results, although performed on stillbirths and not on adults, might suggest that a post-mortem increase of CA is indicative of the stress reaction during the processes leading the fetus to intrauterine death.

Within stillbirths, CA levels were significantly higher in acute death compared to cases with advanced fetal maceration and placental insufficiency, classified as chronic stillbirths. This might reflect a difference related to the agony time.

In adult autopsies, CA concentrations were shown to be elevated in prolonged agony time, with levels comparable to humans under acute maximal stress [34, 38]. Wilke et al. on the contrary found no significant difference of CA heart blood levels in relation to the length of agony, although mean noradrenaline values were higher in deceased with a short agony compares to the group with long agony [30]. Studies on strangulated animals showed significant higher levels of noradrenaline and adrenaline compared to the control (intoxicated) group (noradrenaline 5.4 ± 2.6 ng/mL vs. 2.8 ± 0.1 ng/ml; adrenaline 6.0 ± 3.4 ng/ml vs. 3.8 ± 3.0 ng/mL), suggesting a possible application of the detection of postmortem catecholamines serum levels in establishing the agonal period. In the same study, CA were suggested as good markers of hypoxia or acute stress, even without pain [28].

Despite the limited number of observed stillbirths, our results seem to be in line with these animal models, suggesting that CA could represent a promising biomarker for acute fetal distress. In chronic stillbirths, CA might be released to a lower extent and/or have more time to be degraded in uterus before analysis.

Although glucocorticoids are lipophilic and thus are believed to freely cross the placenta from the maternal to fetal circulation, cortisol levels are controlled in fetuses by the fetal HPA axis and by the placental enzyme 11-β-hydroxysteroid dehydrogenase-type 2 (11β-HSD2), which catalyzes the conversion of active cortisol into inactive cortisone [25, 39, 40]. There is scientific evidence that the activity of the enzyme, whose expression increases at term, is reduced by several pathological conditions, e.g., pre-eclampsia or IUGR, as well as by maternal prenatal stress, permitting a greater amount of maternal cortisol to freely pass the placenta [25, 41, 42]. However, in a recent large study, maternal stress induced a reduction, not an increase, in cortisol cord levels, likely due to blunted fetal response [43]. Placenta also releases corticotropin-releasing hormone (CRH), that can stimulate either the maternal or fetal pituitary adrenal axis, further complicating the fetal regulation of cortisol [44].

In our study, comparing cortisol levels between the stillbirths and comparisons, we did not observe a statistically significant difference. This might be due to the significantly lower gestational age of stillbirths compared to newborns, which might result in an immature HPA axis, unable to increase the production of cortisol despite illnesses [27]. Accordingly, fetal animal studies have shown that in preterm fetuses the cortisol levels, in response to moderate hypoxia or short-lasting asphyxia, are lower than in late-gestation fetuses [27].

However, severe asphyxia caused by 25-min long umbilical cord occlusion caused an upsurge of cortisol even in preterm fetal sheep, but cortisol rapidly returned to basal levels within 72 h [45]. A sustained cortisol elevation is found in ovine fetuses exposed to chronic intrauterine hypoxemia [46].

In our study cortisol levels were not statistically higher in chronic compared to acute stillbirths, suggesting that cortisol is not a good post-mortem marker of fetal distress or acute vs chronic agony. This result is rather in line with previous studies on human infants, showing that CRH, but not cortisol, was increased in cord plasma in cases of chronic fetal distress [44]. This might be due to the complex mechanisms of cortisol balancing involving the placenta and to the presence of variable contributions from fetuses and mothers [25].

A post-mortem decrease of cortisol due to instability is not confirmed by our data, but cannot be totally excluded, given its demonstrated decrease after 48 h at 4 °C [47].

Within our comparison group, cortisol was significantly higher in SVG compared to ELCS, which is consistent with past studies on preterm [44] and term infants [48], suggesting that the stress at delivery might have an influence on the infant HPA axis. Accordingly, the study of Miller et al. on 172 primiparous women with uncomplicated singleton pregnancies found higher cortisol in newborns delivered vaginally than those delivered by caesarean section [49].

The consistency of the results in newborns with past literature data further confirms the reliability of our data.

In the literature, it is known that cortisol has a circadian rhythm, showing normal levels during the early morning hours (8–10 a.m.) and decreasing in the evening and during the early sleep phase (nadir at approximately 4 a.m.), further complicating the use of the hormone as a biomarker for chronic stress [24, 50]. If this rhythm could be confirmed in fetuses, cortisol measurements could provide a clue regarding the timing of death of stillbirths.

However, since the true timing of death in stillbirths is unknown, in our study we evaluated whether cortisol blood levels in newborns could reflect the time of birth, whether delivered during the morning or the afternoon/night.

Only a few studies in the literature evaluated the cortisol circadian rhythm with respect to the time of delivery. Su et al. found low levels of cortisol in samples collected during both daytime and nighttime, although all the cases displayed chronic stress with related adrenal hyporesponsivity, which could have biased the casuistry [43].

In our study, although more research on a wider casuistry is needed, the absence of a significant difference in cortisol levels between newborns delivered at daytime and nighttime does not even allow the application of cortisol analysis to stillbirths for the evaluation of the time of death.