Fetal distress increases interleukin-6 and interleukin-8 and decreases tumour necrosis factor-α cord blood levels in noninfected full-term neonates

Authors


Correspondence: Dr M. Jokie, Service de Néonatologie, CHU Clemenceau, Caen, 14033 Caen Cedex, France.

Abstract

Objective To assess the influence of fetal distress on interleukin-1β, interleukin-6, interleukin-8 and on tumour necrosis factor-α blood levels in noninfected full-term neonates.

Study design In a multicentre prospective study, cord blood samples were obtained at time of delivery from 234 noninfected full-term neonates for the purposes of measuring serum levels of interleukin-1β, interleukin-6, interleukin-8 and tumour necrosis factor-α using immunoassays. Women were classified into four groups according to the mode of delivery (vaginal delivery or caesarean section) and the presence or absence of fetal distress. The role of labour was also investigated.

Results No significant relationship was found between cytokine cord blood levels and the mode of delivery. Fetal distress was associated with an increase in interleukin-6 (P= 0.01) and interleukin-8 (P < 0.001) levels, and a decrease in tumour necrosis factor-α (P < 0.001). Labour was also associated with a significant increase in interleukin-6 and interleukin-8 cord blood levels (P= 0.01 and P < 0.001, respectively).

Conclusion Fetal distress and labour were each associated with elevated interleukin-6 and interleukin-8 cord blood levels in noninfected full term neonates while only fetal distress was associated with decreased tumour necrosis factor-α levels.

INTRODUCTION

Many studies have demonstrated that interleukin-1 β (IL-1β), interleukin-6 (IL-6), interleukin-8 (IL-8) and tumour necrosis factor-α (TNF-α) are released in infected neonates1–6. Although previous studies found that inflammatory cytokines are released in noninfected infants with fetal distress, the results obtained so far remain discordant7,8. Miller et al.7 found that IL-1β blood level was higher in noninfected infants with severe perinatal complications and/or in those born by urgent caesarean section or by vaginal-induced delivery. However, in neonates born with fetal distress, Hata et al. did not detect IL-1β and IL-1 receptor antagonist levels but found elevated levels of IL-6 and IL-88,9. Even though the relationship between fetal distress and cytokine production has not been clearly established, this relationship has been highlighted by several experimental studies. Prolonged hypoxia induces cytokine production in human umbilical venous endothelial cells and in blood mononuclear cells in vitro10–17. Elevated IL-8 secretion has been reported in human umbilical cord segments and in the culture of adult mononuclear cells exposed to hypoxia15,16. Previous studies reported elevated or unchanged levels of TNF-α released from hypoxic endothelial and mononuclear cells10,15 and, unlike IL-6 produced from human umbilical venous endothelial cells which increased during reoxygenation, hypoxia had no influence on IL-6 produced from mononuclear cells12. Elevated serum TNF-α and IL-6 levels have recently been reported in mice exposed to nonlethal hypoxia, in the absence of blood loss or significant hypotension18. The aim of this study was to assess the effects of fetal distress on the cord blood level of the most frequent inflammatory cytokines (IL-1β, IL-6, IL-8 and TNF-α) in noninfected full-term neonates. The influence of the mode of delivery, the presence and length of labour were also investigated.

METHODS

In a prospective study undertaken at Brest University Hospital, Cherbourg General Hospital and Caen University Hospital (France), cord blood samples were collected from noninfected full-term and normal for gestational age singleton neonates born after an uneventful gestation, with and without fetal distress. Fetal distress was defined according to Pontonnier's criteria as the presence of a fetal heart rate > 180 bpm or < 100 bpm for at least 10 minutes, or a fetal scalp pH lower than 7.2019. The following parameters were studied: mode of delivery (vaginal delivery or caesarean section), presence or absence of labour, birth weight, and presence or absence of fetal distress. The start of labour was defined as regular contractions occurring < 10 min apart and progressive cervical dilation or effacement. Neonates were prospectively classified into four groups of at least 50, according to the mode of delivery and the presence or absence of fetal distress: group 1 = caesarean section with fetal distress, group 2 = caesarean section without fetal distress, group 3 = vaginal delivery with fetal distress and group 4 = vaginal delivery without fetal distress. The influence of labour and its length on the cord blood level of cytokines was retrospectively investigated. The presence or absence of infection was determined by the results obtained from the culture of gastric fluid and/or blood or intrathecal fluid samples. Neonates were considered as noninfected when they presented with sterile blood and cerebral fluid, normal CRP and a good clinical course up to the seventh day of life without antibiotic therapy. Infection was presumed or confirmed according to the following set of criteria:

  • 1The presence of maternal risk factors and of clinical signs in neonates: fever, capillary refilling time > 3 s, apnea; elevated CRP levels > 20 mg/L with absent or negative blood, gastric fluid and intrathecal fluid cultures.
  • 2CRP levels of the neonate > 20 mg/L during the first 24 hours of life and bacteriologically-positive gastric fluid smears.
  • 3Positive blood and/or intrathecal fluid cultures.

Additional data were obtained from neonates admitted to the neonatalogy unit for fetal distress. They included the presence or absence of seizures during the first week of life, cranial sonography and electroencephalogram on the third and sixth day of life and neurological evaluation at the end of the first month of life and during the ninth.

Venous cord pH was measured. Lactates were determined using an enzymatic assay coupling lactate oxidase, peroxidase and amino-4 antipyrine enzymes (Lactate PAP, Biomérieux, Lyon, France). To measure cytokine levels, venous umbilical cord blood was collected aseptically within 20 minutes of delivery. Preliminary studies showed that cytokine levels were stable during that period of time (data not shown). Blood was collected without additive and after decantation, and all sera were kept at −80°C until assayed. Cytokines were detected using enzyme-linked immunosorbent assays in microplates. Intra and interassay variability were minimised by adding dilutions of the recombinant cytokine (standardisation curve), and samples of previously-studied sera in each microplate. Commercial kits were used as recommended by the manufacturer for IL-1β (R&D Systems, Abingdon, UK), TNF-α (Medgenix, Rungis, France), IL-8 (Medgenix), and IL-6 (R&D Systems); the lower limits of detection were 3 pg/mL, 3 pg/mL, 0.03 IU/mL and 18 pg/mL, respectively.

Statistical analysis was performed using statistical software (StatView, Abacus Concept Inc, Berkeley, California, USA). Weight, pH and lactate values were expressed as median and range. Results of cytokine cord blood levels were expressed as median and inter-quartile range (25%−75% and 90%). Descriptive data were first analysed using the Kruskall-Wallis test. When a significant difference was found, pairwise comparison of each group was done with the Mann-Whitney U test. Significance was established at P < 0.05. However, because of these multiple comparisons, P values close to 0.05 should be interpreted with caution.

The study was approved by the local Ethical Committee. Informed consent was obtained from the parents of all enrolled children.

RESULTS

Clinical characteristics

Between January 1993 and December 1994, 271 neonates were initially enrolled in this study. Of this number, 119 came from Caen, 41 from Brest and 74 from Cherbourg, respectively. Thirty-seven others were excluded from the study because of incomplete blood sampling or incomplete clinical data or presence of infection (six neonates). The classification of the 234 neonates (130 males and 104 females) into four groups as described above is shown in Table 1. Birthweights ranged from 2450 g-5000 g (median 3390 g). In the caesarean section group, 80 out of 125 children were born in the course of labour. Sixty of them had fetal distress (group 1) and the remaining 20 had no fetal distress (group 2). In five cases, data on labour were not available. As expected, mean pH values were significantly lower and lactates significantly higher in groups with fetal distress as compared with groups without fetal distress (group 1 vs 2 and group 3 vs 4, Table 1). Of 48 neonates with clinical symptoms admitted to a neonatology unit, 32 had fetal distress. No intracranial haemorrhage nor seizure was reported. All the children enrolled in our study had a normal physical examination at the end of the first month of life and during the ninth month except for one who died at birth as a result of a knotted cord.

Table 1.  Clinical and biological data in the four groups of neonates. Values are given as n or median (range) [min-max values] {interquartile range: 25%; 75%; 90%}. IL = interleukin; TNF = tumour necrosis factor.
 Caesarean section
 Group 1 Fetal distressGroup 2 No fetal distressGroup 3 Fetal distressGroup 4 No fetal distress
  1. Statistical analysis was performed using the Mann-Whitney U test when the Kruskall-Wallis test showed significant difference in the distribution of each data.

  2. *P < 0.001 vs Group 2; **P < 0.001 vs Group 4; P= 0.05 vs Group 4; P < 0.05 vs Group 4; §P < 0.001 vs Group 3; ¶P < 0.01 vs Group 2.

No. of neonates66595356
No. born during course    
of labour60205356
Birthweight (g)3285 (2470–4370)3330 (2450–4920)3350 (2610–5000)3430 (2650–4140)
Cord blood pH7.23 (6.93–7.39)7.27(7.11–7.38)7.23 (6.97–7.41)7.29(7.05–7.47)§
Cord blood lactates (mmol/L)5.3 (1.1–22)*2.6(1–124)**§6.6 (2.9–14.8)4.6(1 .2–1 1.5)§
Cord blood concentrations (pg/mL)    
IL–1β0[0–12]{0;0; 1}0[0–41] {0;0;1}0[0–64] {0;0;3}0 [0–27] {0; 0; 0}
IL-60 [0–278] {0; 11.2; 26}0 [0–64] {0; 3; 18.8}0 [0–357] {0; 4.5; 70}0 [0–334] {0;0; 7.3}
IL–830.5 [0–1402] {0; 63;110}*‡0 [0–117] {0; 0; 0}‡§44 [0–887]{0; 87; 277}**0 [0–105] {0; 0; 50}
TNF–α8 [0–33] {0; 15.5; 24]*‡12.5 [0–44] {7.5; 18; 28}§7 [0–39] {0; 13; 20}12 [0–57] {4; 17; 25}

Cytokine blood levels are detailed in tables 1 and 2. Significant difference in the distribution of the cytokines cord blood concentrations was assessed with the Kruskall-Wallis test in Table 1 (IL-6: P= 0.05; IL-8: P < 0.001; TNF-α: P < 0.001) and in Table 2 (IL-1β: P < 0.01; IL-6: P < 0.01; IL-8 < 0.001; TNF-α= 0.01) but only the results of the Mann-Whitney test is shown in these tables.

Table 2.  Influence of the mode of delivery, labour and fetal distress on cytokine cord blood concentrations. Values are given as median (min-max values) [interquartile range: 25%; 75%; 90%]. IL = interleukin; TNF = tumour necrosis factor; VD = vaginal delivery; CS = caesarean section; U = Mann-Whitney U test; NS = not significant.
 IL–1β (pg/mL)IL–6 (pg/mL)IL–8 (pg/mL)TNF–α (pg/mL)
Mode of delivery    
VD (n= 109)0 (0–64) [0; 0; 1.2]0 (0–357) [0; 0; 24]0 (0–887) [0; 60.5; 101]8 (0–57) [4; 17; 23]
CS (n= 125)0 (0–41)[0; 0; 1]0 (0–278) [0; 7; 22]0 (0–1402) [0; 38.5; 78]10.5 (0–44) [2; 17; 27]
U6404725770106997
PNSNSNSNS
Labour    
Yes (n= 189)0 (0–64) [0; 0; 2]0 (0–357) [0; 6 27]0 (0–1402) [0; 60.5; 102.5]9 (0–57) [0.2; 17; 24]
No (n= 40)0 (0–2) [0; 0; 0]0 (0–26) [0; 0; 6.5]0 (0–44) [0; 0; 0]11 (04) [5; 18; 29]
U4054443447094355
PNS0.01< 0.001NS
Fetal distress    
Yes (n= 119)0 (0–64) [0; 0; 3]0 (0–357) [0; 8; 45]33 (0–1402) [0; 72; 209]7.5 (0–39) [0; 14.5; 23]
No (n= 115)0 (0–41)[0; 0;1]0 (0–334) [0; 0; 16]0 (0–117) [0; 0; 40]12 (0–57) [5; 18; 27]
U6608762791708615
PNS0.01< 0.001< 0.001

Cytokine levels and mode of delivery

IL-1β, IL-6, IL-8 and TNF-α were not affected by the mode of delivery (Table 2). In contrast, we found that levels of IL-6 and IL-8, unlike those of IL-1β and TNF-α, increased significantly after labour as compared with cases in which labour was absent (IL-6: P= 0.01; IL-8: P < 0.001). Moreover, in the study population, a labour length > 10 hours led to a significant increase in IL-6 levels as compared with a labour length ≤ 10 hours (36.7 ± 87 pg/mL vs 11.4 ± 37 pg/mL, P < 0.01), although fetal distress was less frequent in the group with prolonged labour than in the group with short labour (56%vs 64%, respectively). In addition, cord blood levels of cytokines between the two subgroups of children born after labour did not differ statistically no matter what the mode of delivery (details not shown).

Cytokine levels and fetal distress

The presence or absence of fetal distress did not affect IL-1β cord blood levels. In contrast, IL-6 and IL-8 levels were significantly higher in the group with fetal distress compared with the group without fetal distress (P= 0.01 and P < 0.001, respectively) while TNF-α cord blood levels were significantly lower in groups with fetal distress (P < 0.001). No correlation was found between cytokine levels and pH or lactate values (Table 2).

The combined effects of labour and fetal distress were assessed using the nonparametric Kruskal-Wallis test (Table 3) which showed significant differences in the distribution of IL-6 (P < 0.05), IL-8 and TNF-α (P < 0.001, respectively). These tests showed that fetal distress and labour individually enhanced IL-8, but had no effect on IL-6 levels. We also observed that fetal distress and labour, when combined, enhanced IL-6 and IL-8 but decreased TNF-α levels.

Table 3.  Influence of fetal distress and labour on cytokine cord blood concentrations. Values are given as median (min-max values) [interquartile range: 25%; 75%; 90%]. Key as for Table 1.
 IL1–β (pg/mL)IL–6 (pg/mL)IL–8 (pg/mL)TNF–α (pg/mL)
  1. Statistical analysis was performed using the Mann-Whitney U test when the Kruskall Wallis-test was significant (i.e. for IL-6, IL-8 and TNF-α). In five cases, data on labour were not available.

  2. *P < 0.01 vs ‘fetal distress and labour’ group; P < 0.05 vs ‘no fetal distress and labour’ group; P < 0.01 vs ‘fetal distress and no labour’ groups; §P < 0.05 vs ‘fetal distress and no labour’ group.

No fetal distress and no labour (n= 39)0 (0–2) [0; 0; 0]0 (0–64) [0; 0; 6.7]*0 (0–44) [0; 0; 0]*†‡12 (2–44)(6.3;18;32)*§
No fetal distress and labour (n= 76)0 (0–41) [0; 0; 1]0 (0–334) [0; 0.5; 17]0 (0–117) [0; 0; 46.5]13 (0–57) (5; 17.8; 25)
Fetal distress and no labour (n= 5)0 (0–0) [0; 0; 0]0 (0–7) [0; 1 .8; 7]0 (0–38) [0; 32.8; 38]1 (0–29) (0; 8.8; 29)
Fetal distress and labour (n= 109)0 (0–64) [0; 0; 3]0 (0–357) [0; 10.3; 50.4]38 (0–1402) [0; 74.3; 212]8 (0–39) [0; 15; 23]

DISCUSSION

Pro-inflammatory cytokines are thought to be involved in normal pregnancy and labour20–23. They are constitutively secreted by various placental cells24–26 and have all been detected in placenta or amniotic fluid27. A progressive increase in IL-1β, IL-6 and TNF-α intra-amniotic levels has been observed during pregnancy21. IL-1, TNF-α and IL-6 levels increased in amniotic fluid during labour20–22,28,29. Chorioamnionitis causes intra-amniotic IL-1β and TNF-α levels to increase about ten-fold, and IL-6 and IL-8 levels about 100 fold. A concomittant increase in IL-6 levels has been observed in maternal and fetal blood while IL-8 increased in fetal blood only27. In any case, no link has been established between maternal and fetal IL-1β, IL-6, IL-8 and TNF-α blood levels and/or amniotic fluid level21,27,30,31. We did not measure maternal and amniotic cytokines in the course of our study.

Our study showed that length of labour and not the mode of delivery are determinant factors in detecting cytokine cord blood levels. To the best of our knowledge, no other study has simultaneously investigated all the effects of fetal distress, labour and mode of delivery on cytokines levels. Opsjon et al. reported elevated IL-6 cord blood levels in neonates born after spontaneous labour as compared with those born by elective caesarean section21. Stallmach et al. recently demonstrated low and stable IL-1, IL-6, IL-8 and TNF-α cord blood levels in neonates born without labour27. However, Miller et al. found elevated cord blood levels of IL-1β in infants born by urgent caesarean section or by induced vaginal delivery but not in those born by elective caesarean section or by normal vaginal delivery8. Hata et al. also found previously nondetectable IL-1β and IL-1 receptor antagonist levels in neonates born by elective caesarean section and by vaginal delivery32. Buoconore et al. reported that IL-1β levels remained unchanged in cord blood irrespective of the mode of delivery30. In our study, IL-6 and IL-8 levels both increased when labour and fetal distress were present. If prolonged labour or uterine contractures occur, fetal oxygenation can rapidly be impaired which could account for the increased levels of IL-6 and IL-8 we observed.

In our study, neonates with fetal distress had moderate and transient hypoxia, severe enough to warrant caesarean section or instrumental extraction delivery. This condition was associated with an increase in IL-6, IL-8 and a decrease in TNF-α cord blood levels. Interactions between fetal distress and cytokines have been previously reported in clinical studies but have thus far led to divergent results. Hata et al.'s finding which is consistent with our own, reported an increase in IL-6 and IL-8 cord blood levels in neonates born with fetal distress8, while Miller et al. found no change in IL-6 and TNF-α but found an increase in IL-1β levels in noninfected infants born with severe complications7. These discrepancies could be due to the different clinical situations and the variability with regard to the amount of cytokine produced by different neonates as has been demonstrated in adults32. Our results are partially consistent with those of Metinko et al. who reported, in vitro, an increase in IL-8 but no variation of IL-1β and TNF-αc production by mononuclear cells in various hypoxia situations15. The concomitant decrease in TNF-α and increase in IL-6 could be attributed to an inhibiting effect of IL-6 on TNF-α production as observed by Schindler et al.32 It could also be explained by the interaction between catecholamines and inflammatory cytokine production since previous studies have particularly shown that epinephrine and norepinephrine secretion could modulate TNF-α and IL-6 production in stressful situations33,34. In vitro, production of cytokines depends on the duration of hypoxia13,16, the maximal secretion occurring during reperfusion11,13,14. Mechanisms of cytokine production involve reactive oxygen intermediates during hypoxia12,16 or during reperfusion11,15. Our results are also consistent with those of Deforge et al. who demonstrated, in vitro, that oxidant stress was an important regulator of IL-8 gene expression since the addition of reactive oxygen intermediate scavengers inhibit IL-8 production by cultured cells17. In these studies, cytokine release does not appear to be the result of altered cell viability. Rather it reflects an increase in transcriptional15,16 or post transcriptional events10. The absence of correlation between pH values and IL-6 or IL-8 cord blood levels was unexpected, one explanation could be that fetal distress was not severe or long enough to induce pronounced acidosis. No previous study has evaluated the relationship between cytokine levels and pH.

Our study underscores the fact that fetal distress is associated with an increase in IL-6 and IL-8 but a decrease in TNF-α cord blood levels. Even if these results do not have any immediate clinical relevance, they should serve as a stimulus to carry out further studies with the aim to measure the cytokine levels in neonates with asphyxia.

Acknowledgements

Our special thanks to Dr L. Adamon, Professor M. Collet and Dr D. Ulmann, and midwives from the three Departments of Obstetrics for their contributions to this study. We thank Mr N. Korney Akrong for translating this article from French into English. This study was funded by the Programme Hospitalier de Recherche Clinique de la Direction Générale de la Santé (no. 940001).

Ancillary