Urinary Tetrahydrocortisol/Tetrahydrocortisone Ratio at Early Postnatal Period Predict Steroid Dependency and Neonatal Condition in Preterm Infant

Insucient adrenal function in preterm infants affects poor neonatal outcome, owing to the immaturity of their adrenal enzyme. While 11β-hydroxysteroid dehydrogenase (11βHSD) type 1 and type2 act as gatekeepers for cell steroid action. This study aimed to investigate the effects of early postnatal urinary tetrahydrocortisol/tetrahydrocortisone (F/E) ratio, used as an alternative indicator of 11βHSDs activity, in preterm infants on their subsequent clinical course. In 80 preterm infants of ≤ 34 weeks gestational age admitted to our hospital, urinary F/E ratio was measured within 24 hours of birth. Furthermore, the relationship between this ratio and neonatal outcomes was estimated. Univariate analysis revealed that the high F/E ratio group had signicantly higher morbidity in terms of duration of ventilatory support for more than 14 days, hypotension requiring inotropes and hydrocortisone, and symptomatic patent ductus arteriosus. On multivariate analysis, the incidence of hypotension requiring hydrocortisone was higher in the high F/E group, despite the absence of elevated dehydroepiandrosterone, a precursor of cortisol. Conclusion: The urinary F/E ratio in the early postnatal period in preterm infants may contribute to the understanding of the pathogenesis of infant condition after birth by estimating the amount of local steroid action in the organs. MR: NICU: neonatal intensive care unit, PDA: patent ductus arteriosus, ROC curve: receiver operating characteristic curve, RQ: relative quantication, SGA: small for gestational age, THE: tetrahydrocortisone, THF: tetrahydrocortisol, 3βHSD: 3β-hydroxysteroid dehydrogenase, 11βHSD: 11β-hydroxysteroid dehydrogenase


Introduction
Glucocorticoid action on target tissues is determined by the hypothalamic-pituitary-adrenal (HPA axis) activation and the density of "nuclear" receptors, and intracellular metabolism by the two isozymes of 11β-hydroxysteroid dehydrogenase (11βHSD). While 11βHSD type 1 is a reductase that catalyses the regeneration of active glucocorticoids thereby amplifying the cellular action, 11βHSD type 2 is a higha nity dehydrogenase that inactivates glucocorticoids. These enzymes act as gatekeepers for cell steroids [1].
On the other hand, prenatal exposure of the foetus to endogenous and exogenous glucocorticoids promotes foetal organ maturation in exchange for foetal growth [1]. Foetal steroid metabolism is dependent on maternal cortisol passing through the placenta in early pregnancy, and the foetal HPA axis system is immature for steroidogenesis on its own and responds to negative feedback by maternal cortisol. Later in pregnancy, as the maternal cortisol is inactivated to cortisone through slowly activated 11βHSD type 2 in the placenta, the foetus is protected from maternal glucocorticoid exposure. Thereafter, the negative feedback is gradually attenuated, the foetal HPA axis is activated, and various organs undergo further maturation [2]. Therefore, the postnatal adrenal function may be affected in preterm infants born when placental 11βHSD type 2 activity is insu cient.
However, the role of these systems in the maturation of preterm infants and their in uence on morbidity later in life are still not completely understood. This study aimed to investigate the effects of placental and early postnatal 11βHSDs in preterm infants on their subsequent clinical course. The novelty of this study is that it is the rst report on the relationship between early postnatal urinary F/E ratio and morbidity in preterm infants.

Materials And Methods
Eighty preterm infants with a gestational age of ≤ 34 weeks who were admitted to Asahikawa Medical University hospital, survived for at least one month after birth, and with a consent to participate, were included in this study. This study was performed in accordance with the principles of the Declaration of Helsinki. Approval was granted by the local institutional review board (ethics committees at Asahikawa Medical University), and informed parental consent was obtained before studying all the cases.
We were able to measure the placental HSD11b2 mRNA levels in 21 patients. Placental HSD11b2 expression was measured using real-time polymerase chain reaction after delivery of the placenta. To examine whether urine collected within 24 hours of birth would re ect 11βHSD type 2 activity in the placenta during the foetal period, we investigated whether the urinary F/E ratio could be an indicator of placental 11βHSD type 2 activity, and the relationship between the two in 21 cases in which placental HSD11b2 expression was measured.
Thereafter, we divided the patients into two groups according to the mean urinary F/E ratio and compared the patient characteristics and the incidence of each disease in the neonatal period in the two groups using univariate and multivariate analyses.

Statistical analysis
Pairwise Pearson's product-moment correlation coe cient was used to assess the correlation between the two groups, and the 2 test and the Wilcoxon signed-rank test were used for comparison between the two groups. Multivariate analysis was performed using logistic regression analysis with adjustment for gestational age, small for gestational age (SGA), and maternal steroid administration as confounders. Statistical analyses were performed using JMP 10.0.2d1 software (SAS Institute Inc., USA).

Our treatment policy and de nition of terms
Our hospital is a tertiary perinatal care centre and provides the most critical care for mothers and infants. Chronic pulmonary disease in newborns is de ned as a child requiring oxygen for more than 28 days after birth; retinopathy of prematurity was evaluated according to the international classi cation; periventricular leukomalacia was diagnosed by magnetic resonance imaging of the brain at discharge; and intraventricular haemorrhage was diagnosed by ultrasonography of the brain during admission to the neonatal intensive care unit (NICU). Circulatory management included the use of hydrocortisone 2-3 mg/kg/dose for postnatal hypotension with poor response to uid loading and inotropes, with additional doses as appropriate. Indomethacin was used for symptomatic patent ductus arteriosus (PDA), and surgical treatment (only in two cases) was used if the response was poor.

Results
In the rst 24 hours of life, the urinary steroid metabolite levels were as follows: urinary THF + allo-THF (ng/ml): 48.43 ± 102.79, THE (ng/ml): 229.07 ± 300.55, and F/E ratio: 0.20 ± 0.22 (mean ± standard deviation). Figure 1 reveals that the F/E ratio was the highest at birth, regardless of the gestational age, and illustrates an immediately decreasing trend with an increasing postnatal age. These trends suggest that local steroid action decreases with age after birth. Table 1 shows the urinary F/E ratio and urinary THF + allo-THF, THE, DHEA level, and HSD11b2 expression in placenta. Urinary THF + allo-THF and THE were measured in 80 cases, DHEA in 59 cases, and HSD11b2 expression in the placenta in 21 cases. There was a signi cant positive correlation between the urinary F/E ratio and urinary THF + allo-THF, but not with urinary DHEA. The relative quanti cation (RQ) of HSD11b2 expression in the placenta was 1.07 ± 0.94; there was no signi cant correlation with the urinary F/E ratio.
Most neonatal outcomes were related to normal adrenal gland function. Table 2 shows the incidence of disease in eligible infants in this study. A nationwide survey in Japan revealed that about 4% of very low birth weight infants require postnatal corticosteroid therapy after 7 days of life. All of them present with at least two of the following ve features, including hypotension, oliguria, hyponatraemia, pulmonary oedema, and increased oxygen supplementation [5]. Late-onset circulatory collapse (LCC) represents 'lateonset' transient adrenocortical insu ciency of prematurity. This symptom of prematurity is often reported in Japan [5,6]. Indomethacin is generally administered three times in one course for PDA in Japan [7]. Therefore, cases of frequent use were de ned as those in which indomethacin was used three or more times in PDA.
To investigate the relationship between morbidity and the F/E ratio in the neonatal period, we compared the patient characteristics between the two groups according to the F/E ratio in Table 3. In the high F/E group, the gestational age was short, and the body size at birth was signi cantly small. Moreover, there were signi cantly more SGA infants in the high F/E group. There were no signi cant differences in other background factors, including prenatal maternal steroid use.
Univariate and multivariate analyses of neonatal morbidity by F/E ratio in the two groups is shown in Table 4. In the univariate analysis, the high F/E group had signi cantly higher morbidity in respiratory and circulatory conditions. On multivariate analysis adjusted for gestational age, anthropometric status at birth (with or without SGA), prenatal maternal steroid administration which may affect the adrenal function of the infants after birth, and the incidence of hypotension requiring hydrocortisone was higher in the high F/E group.
Receiver operating characteristic (ROC) curve drawn between the urinary F/E ratios at birth for hypotension requiring hydrocortisone is depicted in Figure 2. The cut-off value was 0.19 with a sensitivity of 0.7692, speci city of 0.5035, and area under the curve of 0.7826.

Discussion
This is the rst report to reveal a relationship between early postnatal urinary F/E ratio and morbidity in preterm infants. Serum cortisol levels are mainly regulated by the HPA axis. However, during development, the foetal adrenal gland is not highly capable of producing its steroids because of the reduced activity of steroid synthetic enzymes, including 3β-hydroxysteroid dehydrogenase (3βHSD), and foetal adrenal function is dependent on maternal adrenal function and placenta [8,9]. Maternal glucocorticoid concentrations during this period are approximately 5-10 times higher than those of the foetus, mainly due to the activity of 11βHSD type 2 in the placenta, which regulates steroid exposure to the foetus.
During development, 11βHSD type 2 is highly expressed in both the foetal and placental tissues [10][11][12][13]. However, preterm infants who are born unexpectedly at this time are more likely to have impaired adrenal function, which can be associated with severe neonatal conditions such as hypotension [14][15][16].
In addition to the regulation of adrenal function by the HPA axis, the intensity of cortisol action in individual cells is precisely controlled by the balance of the activity between the intracellular glucocorticoid activating enzyme 11βHSD type 1 and the inactivating enzyme 11βHSD type 2 [17]. In most cells, 11βHSD types 1 and 2 coexist, with 11βHSD type 1 being highly expressed in the liver, adipose tissue, central nervous system, and skeletal muscle, and 11βHSD type 2 in the renal tubular epithelium, colon, sweat glands, and placenta, which are involved in water and electrolyte metabolism [18]. 11βHSD types 1 and 2 act as pre-receptor gateways for the glucocorticoid receptor (GR), a system that ne-tunes the intensity of glucocorticoid action on a cell-by-cell basis. Since 11βHSD type 1 is predominantly responsible for reactivating cortisone to cortisol in vivo, it can be regarded as a local ampli cation mechanism of glucocorticoid action. In contrast, 11βHSD type 2 is an enzyme that inactivates cortisol to cortisone and is expressed predominantly in the distal nephron in adults, suppressing the binding of cortisol to the mineralocorticoid receptor (MR) in the kidney. Cortisol also acts as a ligand for the MR and has a comparable a nity to aldosterone. Apparent mineralocorticoid excess syndrome (AME; OMIM #218030) is caused by an abnormality in 11βHSD type 2, which fails to inactivate cortisol in the kidney and binds to the MR, resulting in excessive activation of the MR by glucocorticoids, leading to AME [3,19].
The urinary F/E ratio is used as an alternative indicator of renal 11βHSD type 2 activity in the diagnosis of AME [3]. However, in foetuses and neonates, renal 11βHSD type 2 activity is less than one-tenth of that in adults, and unlike in adults, the foetus converts cortisol to cortisone in many tissues [11]. Therefore, the urinary F/E ratio in the neonatal period may re ect the total activity of 11βHSD types 1 and 2 not only in the kidneys, but also in the whole neonatal body and may be an indicator of local steroid action in the whole body. 11βHSDs are expressed in many foetal cells during this period and, in general, prenatal exposure to endogenous and exogenous glucocorticoids promotes maturation of foetal organs in exchange for overall growth [20,21]. Glucocorticoid exposure in the cells is precisely controlled for each tissue maturation. We hypothesised that urinary F/E ratios at birth could be used to infer local glucocorticoid effects, organ maturation, and adrenal maturation after birth, and to predict the impact on subsequent morbidity.
In the univariate analysis of this study, a higher urinary F/E ratio in the early postnatal period was associated with a longer duration of ventilatory care and a higher incidence of chronic lung disease, symptomatic PDA, and hypotension requiring inotropes and steroids after birth. The results of this study suggest that the high urinary F/E ratio (i.e. low 11βHSD type 2 activity) in preterm infants in the early postnatal period is related to the reduced activity of local steroids in each organ. The fact that hypotension in the early postnatal period requiring steroids was signi cantly more common in the group with a higher F/E ratio in multivariate analysis supports this conjecture. Since several respiratory and circulatory conditions were affected as revealed by the univariate analysis, it was suggested that the F/E ratio in early neonates may estimate the maturation of each organ of the neonate and the maturation of the HPA axis. The present results indicate that there is no correlation between the F/E ratio in the early postnatal period and placental HSD11b2 expression. Therefore, the relationship between each estimation was independent, and the degree of maternal glucocorticoid transfer had less effect on the HPA axis in infants. In other words, it should be noted that the regulation of the intensity of glucocorticoid action in the local organs of the neonate can affect the infant's condition. Based on these results, we speculate the following pathophysiology.
First, there may be inadequate glucocorticoid production in response to the infant's medical condition after birth, resulting in relative adrenal insu ciency, which may enhance the action of local glucocorticoids [14,22]. Previous reports have indicated that insu cient adrenal function in preterm infants causes poor neonatal outcomes because of the immaturity of adrenal enzymes in preterm infants [14]. However, the absence of a correlation between the F/E ratio and DHEA, a precursor of cortisol, suggests that the F/E ratio at birth is not only related to adrenal immaturity at birth, but also to poor responsiveness of the HPA axis. The fact that urinary F/E showed an immediate decrease with postnatal age in the present study, similar to that reported by Midgley et al. [23], may be because of the enhanced local action that becomes unnecessary as the HPA axis matures. In order to compensate for the poor responsiveness of the HPA axis, local steroid action may be enhanced, thus reducing 11βHSD type 2 activity. Heckmann et al. reported similar actions in cardiac paediatric surgery. Our results are in agreement with that of these studies [24].
Second, we considered the immaturity of the regulation of 11βHSD type 2 activity in the kidney. As mentioned above, in the kidney, local 11βHSD type 2 inactivates glucocorticoids, which are also ligands for MR, preventing them from incorrectly activating the MR. However, it has been reported that preterm infants have a low capacity to produce aldosterone but a high expression of MR [25][26][27]. The low renal activity of 11βHSD type 2 in preterm infants (one-tenth of that in adults) may be due to the effect of glucocorticoids on the MR, compensating for the low aldosterone levels. Hypotension requiring steroids in the early postnatal period in preterm infants is not accompanied by classical adrenal insu ciency ndings such as hypoglycaemia, hyponatraemia, and hyperkalaemia, while small doses of hydrocortisone are effective [14]. This phenomenon may be due to the effect of glucocorticoids on MR in the kidneys.
In the present study, the relationship between placental 11βHSD type 2 gene expression and neonatal morbidity could not be examined due to the small number of samples. The urinary F/E ratio in the early postnatal period did not correlate with placental 11βHSD type 2 gene expression and did not re ect placental 11βHSD type 2 activity, suggesting that the urinary F/E ratio can be evaluated independently without placental 11βHSD type 2 activity. The urinary F/E ratio is the most convenient means of measuring the activity of 11βHSD types 1 and 2. However, there are several caveats to the use of these urinary steroid ratios: 1) it is the sum of both 11β-HSD activities; 2) the ratio is also in uenced by 5α-and 5β-reductase activities; and 3) the ratio does not represent a tissue-speci c effect but re ects a systemic effect.
In conclusion, the premature infants suffered several conditions, especially respiratory and circulatory problems, did not only show an increased adrenal stress response, but also an enhancement of glucocorticoid action in local organs due to the reduction in cortisol inactivation. The urinary F/E ratio in the early postnatal period in preterm infants contribute to the understanding of the pathogenesis of the infant condition after birth by estimating the amount of local steroid action in the organs.
Con icts of interest/Competing interests: The authors have no relevant nancial or non-nancial interests to disclose.
Availability of data and material: Not applicable Code availability: Not applicable Authors' contributions: All authors contributed to the conception and design of the study. Material preparation, data collection, and analysis were performed by Toshio Okamoto and Fumikatsu Nohara. The rst draft of the manuscript was written by Ken Nagaya, and all authors commented on the previous versions of the manuscript. All authors read and approved the nal manuscript. Ken Nagaya contributed to the conception of work, and the acquisition, analysis, and interpretation of data.  The receiver operating characteristic curve of hydrocortisone use and urinary tetrahydrocortisol/tetrahydrocortisone (F/E) ratio for hypotension in the early neonatal period F/E ratio were calculated as postnatal days. F/E ratio at birth was the higher level regardless gestational age, then immediately decline.