Dr B. Jacobsson, Perinatal Center, Department of Obstetrics and Gynecology, Sahlgrenska University Hospital/Östra, SE-416 85 Göteborg, Sweden.
Objective To evaluate the levels of interleukin (IL)-6 and IL-8 in cervical and amniotic fluid in relation to the presence of bacteria in the membranes in women in preterm labour (PTL).
Design A prospective follow up study.
Setting Sahlgrenska University Hospital, Göteborg, Sweden.
Sample Women with singleton pregnancies (<34 weeks) presenting with PTL (n= 30).
Methods Amniotic fluid was retrieved transabdominally and cervical fluid was sampled from the uterine cervix at admission and analysed for IL-6 and IL-8 with enzyme-linked immunosorbent assay (ELISA). At birth, the chorioamniotic membranes were separated and samples for polymerase chain reaction (PCR) for Ureaplasma urealyticum and Mycoplasma hominis and general culture were obtained.
Main outcome measure IL-6 and IL-8 in relation to microbial invasion of the chorioamniotic membranes.
Results Bacteria were found in the membranes in 8 of 21 patients in PTL for whom chorioamnion as well as amniotic fluid PCR and cultures were available. Cervical IL-6 was associated with detectable bacteria in the chorioamniotic membranes in women in PTL (median 8.2 ng/mL vs 0.73 ng/mL; P= 0.01). The IL-6 (median 13 ng/mL vs 1.7 ng/mL; P= 0.004) and IL-8 (median 7.2 ng/mL vs 0.28 ng/mL; P= 0.01) levels in amniotic fluid were higher in PTL cases in which bacteria were found in the chorioamniotic membranes.
Conclusion IL-6 in cervical fluid and IL-6 and IL-8 in amniotic fluid were higher in those PTL cases in which bacteria were found in the chorioamniotic membranes.
Preterm birth (PTB), a condition related to increased perinatal mortality and morbidity, including cerebral palsy and suboptimal school achievement, is still a major problem in modern obstetrics.1–4 In Sweden, the incidence of PTB has been 5.6–6.3% since the beginning of the 1970s2; but in the United States and many other countries, it is much higher (12%) and has increased slightly from the 1980s to the mid-1990s.5 Several recent studies have shown that microbial invasion of the amniotic fluid is one important risk factor for PTB.6,7 The preterm labour (PTL) syndrome has been characterised as an inflammatory-like condition.7 It has also been proven in several studies that proinflammatory cytokines such as interleukin (IL)-6 and chemokines such as IL-8 (new terminology CXCL-8) are elevated in amniotic fluid preceding PTB.8,9 We have previously shown that the prevalence of microbial invasion of the amniotic fluid and the relationship between certain cytokines and chemokines are the same in a low risk Swedish population as in the United States.10–13
The main hypothesis is that bacteria ascend from the vagina, through the chorioamniotic membranes to the amniotic fluid and eventually gain access to the fetus. The microbial invasion could then evoke a fetal inflammatory response syndrome.14–16 This means that bacteria can be detected between the maternal tissue and the fetal membranes (choriodecidual space), in the fetal membranes (chorioamnion), within the placenta, within the amniotic fluid or in the fetus. Infection of the fetal membranes is called chorioamnionitis and is documented by histology or culture.17 Some studies have previously been published on the relationship between bacterial invasion of the chorioamniotic membranes and PTB.18,19 Only two studies have investigated the relationship between bacterial invasion of the chorioamniotic membranes and cytokine levels in amniotic fluid, arriving at conflicting results.20,21 No study has previously described the cytokine levels in cervical fluid in relation to the presence of microorganisms in the chorioamniotic membranes.
The aim of this study was therefore to compare the levels of IL-6 and IL-8 in cervical and amniotic fluid in women in PTL with or without the presence of microorganisms in the chorion and amniotic membranes. Such basic information is required in order to understand the pathophysiological process and to develop strategies for prevention and treatment of PTL.
The study population consisted of women with singleton pregnancies in PTL (n= 30) who presented at Sahlgrenska University Hospital/East, Göteborg, at less than 34 weeks of gestation. PTL was defined as regular uterine contractions (at least two uterine contractions/10 minutes during 30 minutes) in combination with cervical changes:
1≤2 cm length +≥1 cm dilatation;
2≤2 cm length + cervical softening;
3≥1 cm dilatation + cervical softening; or
4cervical length <30 mm at endovaginal ultrasound.
Women with pPROM (amniorrhexis, visible amniotic fluid in the vagina, before the onset of spontaneous labour), known uterine abnormalities, fetal malformations, significant vaginal bleeding, imminent delivery or fetal distress were not included.
Gestational age was determined in all patients by routine ultrasound in the second trimester (16th–19th weeks of gestation). Tocolytic therapy (intravenous terbutaline and/or indomethacin if the pregnancy was <28 weeks of gestation) was administered according to the local protocol at the clinic.
Cervical fluid was obtained with a Cytobrush (Cytobrush Plus GT, Medscan Medical, Malmö, Sweden) from the external os of the cervix in all patients in PTL except one (n= 29). The cervical mucus was weighed and kept in a refrigerator (+4°C) until processed within 5 hours. The Cytobrush with the cervical mucus was submerged in 1.0 mL 0.9 mg/mL NaCl, shaken for 30 minutes at +4°C, followed by centrifugation at 855 ×g at +4°C for 10 minutes and storage at −80°C until analysis.
Ultrasound-guided transabdominal amniocentesis, aspirating 30–50 mL of amniotic fluid, was performed in 30 women in PTL. After amniotic fluid retrieval, the amniotic fluid was immediately placed in a refrigerator (+4°C) and was centrifuged within 5 hours of sampling at 855 ×g in +4°C for 10 minutes. The supernatant was stored at −80°C until analysis.
After delivery, each placenta was immediately placed in a refrigerator (+4°C) until processed, most often, within 5 hours. The chorioamniotic membrane was separated with sterile instruments and sampled between the membranes at the rupture site and in an area close to the insertion between the membranes and the placenta. One sample each for polymerase chain reaction (PCR) for Ureaplasma urealyticum and Mycoplasma hominis and general cultures were obtained by a viscose swab from both sites.
A sample of uncentrifuged amniotic fluid and the swabs from the chorioamniotic interface were then immediately transported to the microbiological laboratory for culture for aerobic and anaerobic bacteria (not for U. urealyticum and M. hominis), and PCR analyses of U. urealyticum and M. hominis were performed. Bacterial isolates were characterised biochemically by using the API Rapid ID32STREP, Rapid ID32A, ID32STAPH, ID32E, API ZYM, API Coryne, API50CHL and ID32C according to the manufacturer's instructions (API bioMérieux). These identification systems identified anaerobes, coryneform bacterium, yeasts, streptococci, Enterobacteriaceae and other gram-negative rods, the genera Staphylococcus, Micrococcus and the genus Lactobacillus and related organisms. PAGE analysis of whole-cell proteins was performed as described by Pot et al.22 For densitometric analysis, normalisation and interpretation of protein patterns, the GelCompar 4.1 software package (Applied Maths) was used. Anaerobic bacteria were further analysed by fatty acid analysis using the MIDI specifications.23 Some of the difficult bacterial isolates were identified by DNA sequencing. The 16S rRNA genes of the isolates were amplified by PCR and directly sequenced by using the Big dye terminator cycle sequencing kit (Applied Biosystems) and an automatic DNA sequencer (model 310, Applied Biosystems).
Microbial invasion of the amniotic fluid was defined as positive PCR for U. urealyticum and M. hominis and/or growth of any bacteria in the amniotic fluid except for coagulase-negative Staphylococcus, which was considered to be a skin contamination. However, coagulase-negative Staphylococcus in amniotic fluid from patients with an intra-amniotic inflammation (higher levels of IL-6 and IL-8) was considered to be microbial invasion of the amniotic fluid.10,11 Microbial invasion of the membranes was defined as a positive PCR for U. urealyticum and M. hominis and/or growth of any bacteria from the chorioamniotic membranes.
IL-6 and IL-8 in cervical fluid and amniotic fluid were analysed with enzyme-linked immunosorbent assay (ELISA) (paired antibodies from R&D Systems, Minneapolis, Minnesota, USA). The amniotic fluid samples were diluted 1:5, 1:20 and 1:100 and run in duplicates. The interassay variation was calculated at <10%, based on analysis of several samples on three separate occasions. Low values (≤700 pg/mL) showed a higher coefficient of variation (29–76%). The detection limit of the ELISA test was 30 pg/mL for both IL-6 and IL-8, but because the samples were run at a 1:5 dilution, the lower limit of detection was 150 pg/mL. We have analysed IL-6 and IL-8 in amniotic fluid in the same population in previous studies.10,11 A receiver–operator characteristic curve (ROC) was then used to identify the best cutoff levels of IL-6 and IL-8 for delivery within seven days, and these cutoff levels were also used to define intra-amniotic inflammation. Intra-amniotic inflammation was defined as IL-6 ≥ 1.5 ng/mL and/or IL-8 ≥ 1.3 ng/mL in women. Two investigators scrutinised the medical records and entered maternal and perinatal data into a database.
Ethical approval for the study was obtained from the local ethical committee in Göteborg. The women gave informed consent before enrollment in the study.
Calculations were made using the computer programmes StatView 5.01 (SAS Institute, Cary, North Carolina, USA) and InStat 2.01 (Graph Pad Software, San Diego, California, USA). Continuous variables were analysed with Kruskal–Wallis test and Mann–Whitney U test and proportions with Fisher's exact test. A P value < 0.05 was considered statistically significant.
Demographics and pregnancy outcome are presented in Table 1. Chorioamnion and amniotic fluid PCR and culture were available in 21 patients. Six patients (28%) had bacteria between the membranes but no bacteria in the amniotic fluid (Table 2). Five patients (24%) had bacteria in the amniotic fluid; two of them also had bacteria (different species in both cases) present in the chorioamniotic membranes (Table 2). Ten patients (48%) had no detectable bacteria either in the amniotic fluid or between the membranes. IL-6 and IL-8 were analysed in cervical fluid (n= 20) and amniotic fluid in all 21 cases. IL-6 was detectable in cervical fluid in 85% (17/20) of the women in PTL and in 95% (20/21) of their amniotic fluid samples. IL-8 was detectable in cervical fluid in 85% (17/20) of the women in PTL and in 81% (17/21) in amniotic fluid samples.
Table 1. Demographics and pregnancy outcome in women in PTL with no bacteria between the membranes or in the amniotic fluid, with presence of bacteria between the membranes but not in the amniotic fluid and with bacteria in the amniotic fluid. Data are presented as median and range except for nulliparity, births <34 weeks, term delivery (≥37 weeks), delivery ≤7 days, use of corticosteroids and antibiotics and intra-amniotic inflammation where the number and percentage [n (%)] of the women in preterm and term population, respectively, are given.
No bacteria between the membranes or amniotic fluid
Bacteria between the membranes but no bacteria in the amniotic fluid
Bacteria within the amniotic fluid
AF = amniotic fluid; CF = cervical fluid; NA = not applicable; NS = non-significant.
Fisher's exact test was used to compare nominal variables between the three groups.
Kruskal–Wallis test was used to compare continuous variables among the three groups.
Cervical fluid IL-6 levels in women with bacteria in the chorioamniotic membranes were higher than in those without bacteria (median 8.2 ng/mL vs 0.73 ng/mL; P= 0.01). IL-8 levels in cervical fluid in women with bacteria in the chorioamniotic membranes were also higher than in those without bacteria but had only a borderline significance (median 19 ng/mL vs 4.0 ng/mL; P= 0.05). Cervical fluid IL-6 levels in women with no detectable bacteria in the amniotic fluid but with bacteria in the chorioamnion were higher than in those without bacteria (median 6.2 ng/mL vs 0.73 ng/mL; P= 0.04) (Fig. 1a). However, IL-8 levels in the cervical fluid in women with no detectable bacteria in the amniotic fluid but with bacteria present in the chorioamnion membranes were not higher than in those without bacteria (median 13.6 ng/mL vs 4.01 ng/mL; P= 0.13; Fig. 1b). Women in PTL with microbial invasion of the amniotic fluid had significantly higher cervical fluid IL-6 and IL-8 levels than in the absence of microbes in the amniotic fluid.24
Amniotic fluid IL-6 levels in women with bacteria in the chorioamniotic membranes were higher than in those without bacteria (median 13 ng/mL vs 1.7 ng/mL; P= 0.004). The levels of IL-8 in amniotic fluid in women with bacteria in the chorioamniotic membranes were also higher than in those without bacteria (median 7.2 ng/mL vs 0.28 ng/mL; P= 0.01). Amniotic fluid IL-6 levels in women with no detectable bacteria in the amniotic fluid but with bacteria in the chorioamnion were higher than in those without bacteria (median 9.2 ng/mL vs 1.7 ng/mL; P= 0.02). The IL-8 levels in amniotic fluid in women with no detectable bacteria in the amniotic fluid but with bacteria present in the chorioamnion membranes were higher than in those without bacteria (median 4.7 ng/mL vs 0.28 ng/mL; P= 0.02). Women in PTL with microbial invasion of the amniotic fluid had significantly higher amniotic fluid IL-6 and IL-8 levels than in the absence of microbes in the amniotic fluid.11
This is the first study to document an association between high levels of IL-6 in cervical fluid and the presence of bacteria in the chorioamniotic membranes. This study is also the first to document a significant association between amniotic fluid IL-8 and the presence of bacteria in chorioamnion. We have also confirmed the findings from a previous study concerning the association between IL-6 in the amniotic fluid and bacterial presence in the membranes.20
Two studies have previously addressed the issue of microorganisms in the chorioamnion without microbial invasion of the amniotic fluid and its relation to presence of an inflammatory response in the amniotic fluid,20,21 yielding conflicting results. Hillier et al.21 did not find any association when they examined IL-6, IL-1α, IL-1β, tumour necrosis factor-α, interferon-γ and prostaglandin E2 in relation to the discovery of bacteria between the membranes. As sampling of amniotic fluid was not simultaneous with the membrane sampling, the authors also stratified for delivery within seven days but this did not change the results.21 Their sampling delay concurred with ours. However, Andrews et al.20 found a relationship between IL-6 in the amniotic fluid and the presence of bacteria between the membranes in their study in which they sampled the amnion and membranes at the same time. Our findings regarding amniotic fluid confirm Andrews' and increase existing knowledge with the finding that the chemokine IL-8 is also increased in the presence of microorganisms in the membranes. This indicates that the bacteria in our study might have been present in the chorioamnion when sampling occurred or might have entered later during the latency period or delivery. It also suggests that high levels of IL-6 and IL-8 might either be markers of microbial invasion of the membranes or of a process leading to colonisation of the membranes. On the other hand, the detection of bacteria both in the membranes and in the amniotic fluid can be a sign of a more pronounced colonisation that provokes a more intense inflammatory response with contributions from different production sites.
All patients with bacteria detected in the chorioamnion but not in the amniotic fluid had signs of intra-amniotic inflammation with increased levels of cytokines. We have previously found that 16% of a Scandinavian population presenting with PTL had microorganisms in the amniotic fluid and 49% had intra-amniotic inflammation.11 The increased levels of IL-6 and IL-8 in these cases could either be explained by a diffusion of cytokines over an inflamed membrane or by the presence of bacteria activating cells in the amniotic membrane to produce cytokines that are released into the amniotic fluid. Kent et al.25 found that amniotic fluid contents do not reflect the cytokines produced by the decidua in cases with intact and non-inflamed fetal membranes, but conditions might be different in an inflamed intrauterine in vivo situation. Another possibility is that bacteria are present in the amniotic fluid but escaped detection. There is evidence that the detected prevalence of microbial invasion of the amniotic fluid doubles if 16S rDNA PCR is used, especially in cases with high IL-6 levels.26 We only used PCR for U. urealyticum and M. hominis so 16S rDNA PCR would possibly reveal that some of the negative amniotic fluid samples and chorioamniotic samples contain bacteria.
The novel finding of a relationship between the cervical fluid IL-6 level and discovery of microorganisms between the membranes supports the theory of an ascending infection.14 The source of cervical fluid IL-6 may be the cervical tissue itself. The decidua and membranes may also be sources of IL-6, as the microorganisms in the chorioamniotic membranes might cause a local inflammatory process leading to a disruption of the choriodecidual interface, thus facilitating the release of IL-6 and other proteins into the cervical fluid.27
In conclusion, the IL-6 in cervical fluid and IL-6 and IL-8 amniotic fluid were higher in women in PTL in whom bacteria were detected in the chorioamniotic membranes.
The study was supported by the Swedish Medical Research Council (09455), The Göteborg Medical Society, The Frimurare Barnhus Foundation and by Swedish government grants to researchers in public health service (ALF). The technical assistance of Ellen Samuelsson and Jolanta Bonislawska are very much appreciated.