Article Text
Abstract
Objective: To determine the fetal cortisol response to intrauterine infection.
Study design: 16s ribosomal RNA genes or the urease genes of Ureaplasma spp were identified by the polymerase chain reaction in intrauterine samples. Cord blood cortisol, interleukin 6 (IL6) and IL8 were measured in samples from 31 infants born at <32 weeks gestation.
Results: 13 infants (median gestation 30 weeks, birth weight 1350 g) had at least one positive intrauterine sample for microbial genes and 18 infants (31 weeks, 1320 g) did not. The cord blood cortisol concentration was significantly higher in fetuses exposed to intrauterine infection and significantly increased in fetuses/mothers presenting in preterm labour with intact membranes compared with infants delivered by elective prelabour caesarean section (p<0.05). The cord blood cortisol concentration was increased in the mothers with prelabour premature rupture of membranes but this was not significant compared with the caesarean section group. The cord blood cortisol concentration was significantly increased in the presence of chorioamnionitis or funisitis and was moderately correlated with cord blood IL6 (r = 0.64, p<0.01) and IL8 (r = 0.52, p<0.01).
Conclusions: In this study, cord blood cortisol was increased in the colonised group compared with non-colonised infants. It is unclear if infants born following prelabour premature rupture of the membranes mount an adequate anti-inflammatory response.
Statistics from Altmetric.com
Intrauterine infection and inflammation are frequent findings in spontaneous preterm deliveries of less than 33 weeks.1 ,2 A decrease in acute lung disease in premature infants has been observed in infants exposed to intrauterine infection and chorioamnionitis.3–5 Chorioamnionitis also appears to improve early survival after preterm birth.6 Despite the possible short-term beneficial effects, chorioamnionitis may have deleterious effects, such as pulmonary inflammation with subsequent chronic lung disease (CLD)3 ,7–10 and neurological injury.9–12 The fetal responses to intrauterine infection and inflammation need to be further characterised if we are to prevent the pulmonary inflammation that is so often observed in infants who progress to develop CLD of prematurity.13 ,14
An increase in glucocorticoids near to term in most animals is thought to contribute to maturation of the lungs. In mice, the absence of corticotrophin-releasing hormone or glucocorticoid receptors results in delayed maturation of distal lung structures.15 ,16 Exogenous glucocorticoids induce the maturation of premature lungs in animal models17 and humans.18 Based on this evidence, it has been proposed that chorioamnionitis might induce accelerated lung maturation by increasing endogenous cortisol secretion.19 However, this observation was not confirmed in a model of intra-amniotic inflammation in preterm lambs in which improvements in lung function were seen in the absence of a rise in cord blood cortisol concentrations.20
We extended our previous study21 by investigating the fetal cortisol responses to the presence of intrauterine microbial genes detected by using the polymerase chain reaction (PCR) and to intrauterine inflammation identified by presence of interleukin 6 (IL6) and IL8.
METHODS
Our previous study, which was approved by the local ethics committee, had included 41 preterm infants of less than 33 weeks, and the presence of intrauterine infection was identified by detecting the presence of microbial genes (16s ribosomal RNA (rRNA) and the urease gene for Ureaplasma spp) by PCR and intrauterine inflammation by the presence of IL6 and IL8 in cord blood and other intrauterine samples.21 Cord blood was available in a subset of 31 infants in whom cortisol in umbilical cord blood was determined by ELISA (DRG Instruments, Germany). The assay was based on a monoclonal antibody directed at a unique site on the cortisol molecule. There was 45% cross-reactivity with corticosterone, less than 9% cross-reactivity with progesterone, less than 2% cross-reactivity with deoxycortisol and dexamethasone, and no remarkable cross-reactivity with oestriol, oestrone and testosterone. The lowest sensitivity was 6.9 nmol/l.
Methods used for detecting the presence of 16s rRNA genes and the urease gene for Ureaplasma spp, measurement of IL6 and IL8 in intrauterine samples and presence of chorioamnionitis and funisitis are described in more detail in our previous paper.21 The results for blood counts and C-reactive protein (CRP) were obtained by routine venepuncture within an hour of birth for all infants. Chest x rays were obtained in infants who had clinical signs of respiratory distress beyond 4 h of age and independently reported by radiologists who were unaware of other study results.
Statistical analysis
Data are given as median (interquartile (IQR)) ranges. Two-group comparisons were made by using the non-parametric Mann–Whitney U test and multiple groups were compared by using the Kruskal–Wallis test. Correlation between cord blood IL6, IL8 and cortisol concentrations was assessed using the Spearman rank correlation test for non-parametric data. The receiver operating characteristics curves were plotted with SPSS version 9.
RESULTS
Patient characteristics
Cortisol and cytokine concentrations in cord blood were available from 31 preterm infants. Of these infants, 9 were delivered after prelabour premature rupture of the membranes (pPROM), 8 were delivered after spontaneous onset of preterm labour with intact membranes, 5 were spontaneous preterm twin deliveries and 9 were delivered by caesarean section or maternal or fetal indications without onset of labour. All but two mothers had received antenatal steroid therapy (dexamethasone). Median time of administration of the first dose of dexamethasone was 4 days prior to delivery (IQR 0.6–13.6 days). For the different groups the medians (IQRs) were: caesarean section 3.1 days (0.8–7.2), pPROM 5.3 days (2.2–11.9), preterm labour with intact membranes 0.4 days (0.1–8.1) and twin group 22 days (11.1–26.0). Twelve mothers had received antibiotics prior to delivery with the greatest use in the pPROM group: caesarean section 2/9 (22%), pPROM 7/9 (78%), preterm labour with intact membranes 1/9 (11%) and twin 2/5 (40%). However, none of the infants had positive blood cultures during the first week following delivery. Microbial genes were detected in intrauterine samples from 13 (42%) infants (6 in the pPROM group, 6 in the spontaneous onset of preterm labour with intact membranes group, 1 in the caesarean section group and none in the preterm twin group).
Table 1 shows the patient characteristics. Histological chorioamnionitis was noted in 8 (62%) and funisitis in 7 (54%) of the 13 infants with microbial colonisation but no chorioamnionitis or funisitis was noted in any of the 18 infants in the non-colonised group. When the complete study group was considered, radiological evidence of respiratory distress syndrome (RDS) was less frequent in the colonised group than in the uncolonised group although this difference was not reflected clinically (table 1).
Measurement of umbilical cord blood cortisol
Cortisol concentrations in cord blood were significantly higher in the colonised group than in the non-colonised group (fig 1, table 2). As previously reported in this same group of babies, the current subset of preterm infants had increased cord blood IL6 and IL8 in the presence of intrauterine microbial colonisation.21 Although there were no significant differences between the two groups for CRP or total white cell counts, monocytes were increased in the colonised group compared with uncolonised infants (table 3). The cord blood cortisol levels were significantly higher if funisitis (the histological correlate of fetal systemic inflammation) or chorioamnionitis were present (table 4). There was a moderate correlation between cortisol and IL6 (r = 0.64 after log10 transformation of IL6, p<0.01) and IL8 concentrations (r = 0.52 after log10 transformation of IL8, p<0.01) in cord blood (fig 2). This relationship persisted when only vaginal deliveries were considered. Since there may be a relationship between gestation and cord blood cortisol, we repeated the analyses controlling for gestational age: r = 0.71, p<0.001 between cortisol and IL8, and r = 0.60, p<0.001 between cortisol and IL8.
Table 5 shows the sensitivity and specificity obtained by a receiver operator curve for cortisol, IL6, IL8 and monocytes to identify the presence of microbial genes as detected by PCR. Cord blood IL6 and IL8 seemed to show the greatest sensitivity and specificity at a concentration of 45.4 pg/ml and 35.3 pg/ml, respectively.
When the results in the four preterm delivery categories were compared, the cord blood cortisol concentrations were significantly higher in the spontaneous onset of preterm labour with intact membranes group than in the caesarean section or twin groups (p<0.02) (fig 3). Concentrations of cord blood IL6 and IL8 are also shown.21 Interestingly, in the pPROM group the cortisol was increased when compared to the caesarean section group but this difference did not reach statistical significance (p = 0.23).
DISCUSSION
What is already known on this topic
Intrauterine infection and inflammation are frequent findings in spontaneous preterm deliveries of less than 33 weeks.
Fetal cortisol is important in responding to antenatal infection.
What this study adds
Cord blood cortisol is increased in colonised preterm infants compared with non-colonised infants.
The increase in cord blood cortisol may be inadequate in infants with prelabour premature rupture of the membranes.
Our study showed that cortisol was increased in umbilical cord blood from preterm infants in whom microbial genes were identified in intrauterine samples by using PCR. Furthermore, the cortisol concentrations were increased significantly in the preterm labour with intact membranes group. Such a significant increase was not seen in the pPROM group although the numbers that we studied in each group were low. Cord blood cortisol was also increased in the infants who had funisitis or histological chorioamnionitis. The cortisol concentration correlated significantly with both IL6 and IL8. We also found that IL6 and IL8 had reasonable sensitivity and specificity for predicting presence of microbial genes in the intrauterine environment, but cortisol was only moderately predictive.
These results are consistent with previous findings. Watterberg et al reported increased cortisol concentrations in very low birthweight infants on day 2 following exposure to chorioamnionitis,19 which is closely related to intrauterine infection.21 In addition, there was increased sensitivity of the hypothalamic–pituitary–adrenal (HPA) axis to stimulation in the first week of life following exposure to intrauterine inflammation.19 Another study demonstrated increased fetal adrenal steroid production in response to maternal infection.22 Prior to these studies, a possible link between chorioamnionitis and the HPA axis had been noted when a series of post-mortem examinations revealed that infants exposed to chorioamnionitis had larger adrenal glands than those not exposed.23
In the present study, the finding of a moderate correlation between cortisol and both IL6 and IL8 persisted when vaginal deliveries alone were considered. Such a correlation has previously been reported in preterm infants born following pPROM.24 Presumably the increase in cortisol is partly due to the stresses of labour but greater increases were seen in the colonised group suggesting a potential increase due to the inflammatory processes occurring in the intrauterine cavity. There is evidence that proinflammatory cytokines stimulate pituitary cells in vitro,25 ,26 and the HPA axis in vivo,27 the likely mechanisms for our observation.
Chorioamnionitis has previously been associated with improvements in short-term lung function.3 ,5 ,19 It has been postulated that this effect results from an accompanying increase in cortisol.19 Given the importance of corticosteroids in lung development, it would seem likely that increased cortisol concentrations secondary to intrauterine infection would contribute to accelerated lung development. However, it is also known that intra-amniotic inflammation in the preterm lamb model can induce lung development in the absence of a rise in cortisol, suggesting a role for the direct effect of cytokines.20 Whether two separate pathways exist is currently unclear.
Two recent trials of low-dose hydrocortisone to prevent CLD have been discontinued because of adverse effects, in particular an increased incidence of intestinal perforation.28 ,29 As infants with cortisol concentrations above the median seem more susceptible to side effects, a more selective approach to steroid supplementation has been suggested.30 Infants exposed to intrauterine infection are likely to have higher concentrations of cortisol immediately after birth, so this may be a group in whom prophylactic steroid treatment should be avoided. However, in one trial, the infants exposed to chorioamnionitis seemed to benefit most from hydrocortisone supplementation in preventing CLD. Our data suggest that the infants delivered to mothers with pPROM had a high inflammatory response but did not have a marked increase in cord blood cortisol compared with those preterm infants delivered for maternal or fetal reasons. Our hypothesis that there is an inadequate anti-inflammatory response of cortisol in infants born to mothers with pPROM is consistent with other published data but needs to be formally confirmed in an adequately powered study.
The most sensitive and specific markers of intrauterine infection in our study were the proinflammatory cytokines, IL6 and IL8. An increase of monocyte concentration was also observed. However, total white cell count and CRP did not seem useful in the diagnosis of intrauterine infection or fetal inflammation, although in other studies appreciable associations have been found.19 ,31 The numbers of mother–infant pairings studied here do not allow firm conclusions to be drawn although the colonised group did have decreased radiological evidence of RDS.
SUMMARY
We have shown that intrauterine infection was associated with increased umbilical cord blood cortisol in cohort of preterm infants. In the human preterm infants, an increase in cortisol may have a role in modulating fetal lung development along with the direct effects of proinflammatory cytokines. The most useful markers of antenatal infection were IL6 and IL8 rather than CRP or total white cell count. Further studies on a larger cohort would be appropriate to confirm if specific groups of infants would benefit from cortisol supplementation.
REFERENCES
Footnotes
Competing interests: None.
Ethics approval: The study was approved by the local research ethics committee.