Objective
The purpose of this study was to compare intraamniotic inflammation vs microbial invasion of the amniotic cavity (MIAC) as predictors of adverse outcome in preterm labor with intact membranes.
Study Design
Interleukin-6 (IL-6) was measured in prospectively collected amniotic fluid from 305 women with preterm labor. MIAC was defined by amniotic fluid culture and/or detection of microbial 16S ribosomal DNA. Cases were categorized into 5 groups: infection (MIAC; IL-6, ≥11.3 ng/mL); severe inflammation (no MIAC; IL-6, ≥11.3 ng/mL); mild inflammation (no MIAC; IL-6, 2.6-11.2 ng/mL); colonization (MIAC; IL-6, <2.6 ng/mL); negative (no MIAC; IL-6, <2.6 ng/mL).
Results
The infection (n = 27) and severe inflammation (n = 36) groups had similar latency (median, <1 day and 2 days, respectively) and similar rates of composite perinatal morbidity and mortality (81% and 72%, respectively). The colonization (n = 4) and negative (n = 195) groups had similar outcomes (median latency, 23.5 and 25 days; composite morbidity and mortality rates, 21% and 25%, respectively). The mild inflammation (n = 47) groups had outcomes that were intermediate to the severe inflammation and negative groups (median latency, 7 days; composite morbidity and mortality rates, 53%). In logistic regression adjusting for gestational age at enrollment, IL-6 ≥11.3 and 2.6-11.2 ng/mL, but not MIAC, were associated significantly with composite morbidity and mortality rates (odds ratio [OR], 4.9; 95% confidence interval [CI], 2.2–11.2, OR, 3.1; 95% CI, 1.5–6.4, and OR, 1.8; 95% CI, 0.6–5.5, respectively).
Conclusion
We confirmed previous reports that intraamniotic inflammation is associated with adverse perinatal outcomes whether or not intraamniotic microbes are detected. Colonization without inflammation appears relatively benign. Intraamniotic inflammation is not simply present or absent but also has degrees of severity that correlate with adverse outcomes. We propose the designation amniotic inflammatory response syndrome to denote the adverse outcomes that are associated with intraamniotic inflammation.
Intrauterine infection and inflammation are well-documented causes of preterm labor with intact fetal membranes, especially at very early gestational ages. Cultures for microorganisms in amniotic fluid demonstrate microbial invasion of the amniotic cavity (MIAC) in 20-60% of women with preterm labor at <28 weeks of gestation and 10-25% at 28-32 weeks of gestation. Culture-proven MIAC is associated with short latency to delivery and high rates of perinatal morbidity and mortality.
Even with culture-negative amniotic fluid, however, women in preterm labor often have intraamniotic inflammation, which is evidenced by elevated amniotic fluid levels of inflammatory markers such as interleukin-6 (IL-6), other proinflammatory cytokines and chemokines, tumor necrosis factor alpha, or matrix metalloproteinase-8 (MMP-8). Whether or not the amniotic fluid culture is positive, intraamniotic inflammation is associated with short latency and high rates of perinatal morbidity and mortality.
One explanation for the morbidity that is associated with culture-negative intraamniotic inflammation is that many cases actually have MIAC but that the amniotic fluid cultures are falsely negative. Using polymerase chain reaction (PCR) amplification, several groups have demonstrated prokaryotic 16S subunit ribosomal RNA or the DNA coding for it (rDNA) in amniotic fluid in many culture-negative preterm labor cases. The microbes that are identified by 16S PCR techniques are often facultative organisms that are difficult to culture with standard techniques. Preterm labor cases with 16S PCR-proven MIAC have similar outcomes to cases with culture-proven MIAC, which suggests true infection and not simply detection of nonviable microbial degradation products.
In principle, MIAC and the intraamniotic inflammatory response are distinct entities. In the simplest model, each of them can be either present or absent; therefore, states of amniotic fluid are possible: (1) infection (MIAC and inflammatory response both present), (2) inflammation (inflammatory response present, MIAC absent); (3) colonization (MIAC present, inflammatory response absent); (4) negative (both absent).
Moreover, the inflammatory response is not simply present or absent but is a continuum. A recent report suggested that clinical outcomes correlated with gradations in inflammatory response that varied from “no” to “minimal” to “severe,” with categories defined by the number of biomarkers that are present in amniotic fluid. To our knowledge, there has been no previous report that has investigated whether the severity of outcomes might be graded similarly based on the concentration of a single inflammatory marker.
The aims of the present investigation were to compare the outcomes of preterm labor in women with intraamniotic infection, inflammation, or colonization and to examine whether the outcomes are related to the severity of the inflammatory response as defined by intraamniotic IL-6 levels.
Materials and Methods
This report involved a subset of subjects from a larger multicenter study, the goal of which was to develop a noninvasive test to screen for intraamniotic infection based on cervicovaginal proteins. The protocol was approved by the local institutional review board at each participating site.
Inclusion/exclusion criteria
We included consenting women who were at least 18 years old with singleton pregnancies at 15.0-36.9 weeks of gestation in spontaneous preterm labor with intact fetal membranes and who underwent amniocentesis to evaluate for intraamniotic infection and to measure amniotic fluid IL-6. Preterm labor was defined as regular uterine contractions plus at least 1 of the following: cervical dilation ≥2 cm; cervical length by transvaginal sonography ≤30 mm; or a positive cervicovaginal fetal fibronectin test. The protocol required cervical length by transvaginal sonography or fetal fibronectin test only if cervical dilation was <2 cm. Each of the participating sites was a tertiary perinatal center where amniocentesis was offered routinely to women with preterm labor, although several centers restricted the procedure to <34 weeks of gestation. Exclusion criteria were ruptured membranes, major fetal anomaly, fetal aneuploidy, or a medical indication for preterm birth.
Specimens
Amniotic fluid was obtained by transabdominal amniocentesis with the use of sonographic guidance and antiseptic skin preparation. A 5-mL aliquot was sent to the local hospital laboratory for assessment of glucose concentration, white blood cell count (WBC), Gram stain, and aerobic and anaerobic culture, which included genital mycoplasmas at some laboratories. A 10-mL aliquot of amniotic fluid was frozen at –80°C and shipped to the central laboratory at ProteoGenix, Inc (Costa Mesa, CA).
If delivery occurred at the study hospital, a full-thickness placental biopsy that included chorionic plate and a segment of umbilical cord within 2-3 cm from the placental insertion site were obtained and fixed in formalin. Biopsy specimens were interpreted by a placental pathologist (T.K.M.) who was blinded to clinical and laboratory findings.
Additional specimens that were obtained as part of the parent study but were not considered for the present study included samples of maternal plasma, cervicovaginal secretions, urine at enrollment, and cord blood at delivery.
Treatment
Management of preterm labor, duration of hospitalization, route of delivery, diagnosis and management of infection, and other clinical decisions were left to the discretion of the caregivers and not specified by the protocol. The only study-specific procedures were the collection of specimens: the extra aliquot of amniotic fluid, cervicovaginal swabs, maternal plasma and urine, cord blood, and placental biopsy specimens. Caregivers had access to local hospital laboratory results but were blinded to all results from the central laboratory.
Central laboratory methods
Amniotic fluid IL-6 concentration was assayed with a quantitative sandwich enzyme immunoassay (Quantikine ELISA #D6050; R&D Systems, Minneapolis, MN).
Cultures of amniotic fluid were performed on an aliquot of amniotic fluid that had been sent to reference laboratories at the University of Washington (through May 2008) or Focus Diagnostics Inc (Cypress, CA; after June 2008). Broth enrichment techniques were used to detect low levels of aerobic and anaerobic bacteria, including genital mycoplasmas, as described elsewhere.
PCR for amniotic fluid 16S rDNA was performed on DNA that was extracted with the use of an FFTE kit and Maxwell 16 system (Promega, Madison, WI). The optimized PCR reaction included universal 16S rDNA primers 5F and 531R (Invitrogen Life Technologies, Carlsbad, CA). All extraction runs included known positive and negative controls that were generated by spiking the water sample with Staphylococcus aureus . Positive samples, which were identified by the presence of amplicon of 460-560 base pair, were ligated into the pCR4 TOPO vector and transformed into Escherichia coli TOP10 cells with the use of the TOPO-TA cloning kit (Invitrogen Life Technologies). After overnight incubation on ampicillin-containing agar, up to 24 clones per sample were selected. Plasmids were isolated from bacterial colonies with the QIAprep Spin kit (Qiagen Sciences, Germantown, MD). Inserts from 24 plasmids that were generated from each positive PCR were sequenced with an AB 3730xl sequencer (Applied Biosystems, Foster City, CA) with the custom primers Bac16S_8F2 and Bac16S_R518. For bacterial identification, bidirectional sequences were processed with MicroSeq analysis software (Applied Biosystems), and MicroSeq ID 16S rDNA 500 Library v2.0. Ureaplasma urealyticum and Sneathia sanguinegens sequences were added to create custom libraries.
Groups defined by amniotic fluid results
To define groups, we used the cut-off amniotic fluid IL-6 concentrations of 11.3 ng/mL and 2.6 ng/mL 3 that had been determined by previous investigators. MIAC was defined by a positive 16S rDNA result and/or a positive culture from either the local hospital and/or reference laboratory.
For analysis, cases were divided into 5 groups based on amniotic fluid results: (1) infection group: MIAC plus IL-6 concentration of ≥11.3 ng/mL; (2) severe inflammation group: no MIAC, IL-6 concentration of ≥11.3 ng/mL; (3) mild inflammation group: no MIAC, IL-6 concentration of 2.6-11.2 ng/mL; (4) colonization group: MIAC, IL-6 concentration of <2.6 ng/mL; and (5) negative group: no MIAC, IL-6 concentration of <2.6 ng/mL. These groups were compared regarding baseline characteristics, pregnancy outcomes, and neonatal outcomes.
Definitions
Clinical chorioamnionitis was defined as antenatal maternal fever (≥100.4°F) plus either maternal leukocytosis (WBC >15,000/mm 3 ), uterine tenderness to palpation, and/or fetal tachycardia (baseline fetal heart rate >160 beats/min).
Histologic chorioamnionitis was defined by the presence of neutrophils that marginated into the placental chorionic plate; funisitis was defined by the presence of neutrophils in the umbilical cord vessel walls.
Composite perinatal morbidity and death was defined as any ≥1 the following events: stillbirth, neonatal death, respiratory distress syndrome, grade 3 or 4 intraventricular hemorrhage, necrotizing enterocolitis, or culture-positive neonatal sepsis. Neonatal outcomes were assessed from birth to 28 days of life or until hospital discharge, whichever came first.
Statistical analyses
Between-group differences were tested with χ 2 or Fisher exact test for categoric variables, t test and analysis of variance for continuous variables (with log-transformation as appropriate), and log-rank test for time-to-event curves. Multivariable logistic regression analyses were performed to examine predictors of short latency, composite perinatal morbidity and mortality rates, or composite morbidity and mortality rates that excluded respiratory distress syndrome and were adjusted for gestational age at enrollment or gestational age at delivery. One set of regression models included MIAC (present or absent) and amniotic fluid IL-6 concentration (≥11.3, 2.6-11.2, or <2.6 ng/mL) as predictors. Another set of models included the 5 defined groups as predictors. Probability values of < .05 were considered significant.
Results
From September 2007 through November 2009, 338 women who met all the inclusion criteria were enrolled at 16 sites. Of these, 18 women withdrew or were lost to follow-up evaluation, and 15 were excluded because of inadequate specimens, which left 305 women for the analyses presented here. The number of women at each site and the gestational ages at amniocentesis are tabulated in Table 1 .
| Site no. | Gestational age, wk | |||
|---|---|---|---|---|
| <30 | 30-33.9 | ≥34 | Total | |
| 1 | 7 | 9 | 4 | 20 |
| 3 | 9 | 5 | 0 | 14 |
| 4 | 14 | 13 | 0 | 27 |
| 5 | 15 | 4 | 0 | 19 |
| 6 | 3 | 4 | 0 | 7 |
| 7 | 2 | 2 | 1 | 5 |
| 8 | 21 | 7 | 3 | 31 |
| 9 | 1 | 0 | 0 | 1 |
| 10 | 15 | 40 | 26 | 81 |
| 11 | 26 | 10 | 9 | 45 |
| 12 | 11 | 12 | 8 | 31 |
| 13 | 2 | 2 | 2 | 6 |
| 14 | 4 | 5 | 1 | 10 |
| 16 | 0 | 1 | 0 | 1 |
| 19 | 0 | 1 | 1 | 2 |
| 20 | 3 | 1 | 1 | 5 |
| Total | 133 | 116 | 56 | 305 |
Groups defined by amniotic fluid results
MIAC was found in 31 of 305 women (10.1%). As summarized in Table 2 , 20 of these cases (65%) were both 16S rDNA PCR-positive and culture-positive; 6 cases (19%) were PCR-positive but culture-negative, and 5 cases (16%) were culture-positive but PCR-negative.
| Group classification | Organism identified | 16S-ribosomal DNA result | Culture result | Amniotic fluid interleukin-6, ng/mL |
|---|---|---|---|---|
| Infection | Bacteroides ureolyticus | Positive | Positive | 206 |
| Sneathia sanguinegens | Positive | Negative | ||
| Beta-lactamase positive bacterium | Negative | Positive | ||
| Infection | Sneathia sanguinegens | Positive | Negative | 199 |
| Unknown bacterium | Negative | Positive | ||
| Infection | Fusobacterium nucleatum | Positive | Positive | 166 |
| Infection | Gardnerella vaginalis | Positive | Negative | 161 |
| Diphtheroids | Negative | Positive | ||
| Mobiluncus sp | Negative | Positive | ||
| Infection | Haemophilus influenzae | Positive | Positive | 147 |
| Beta-lactamase positive bacterium | Negative | Positive | ||
| Infection | Fusobacterium nucleatum | Positive | Positive | 106 |
| Infection | Fusobacterium nucleatum | Positive | Negative | 102 |
| Infection | Bacteroides ureolyticus | Positive | Positive | |
| Sneathia sanguinegens | Positive | Negative | ||
| Gardnerella vaginalis | Negative | Positive | ||
| Acinomyces sp | Negative | Positive | ||
| Peptostreptococcus sp | Negative | Positive | ||
| Ureaplasma urealyticum | Negative | Positive | ||
| Mycoplasma hominis | Negative | Positive | 98.5 | |
| Infection | Ureaplasma urealyticum | Negative | Positive | 71.7 |
| Infection | Ureaplasma urealyticum | Positive | Positive | 54.8 |
| Ureaplasma parvum | Positive | Negative | ||
| Infection | Candida albicans | Negative | Positive | 51.1 |
| Infection | Fusobacterium nucleatum | Positive | Positive | 25.9 |
| Bacteroides ureolyticus | Positive | Negative | ||
| Infection | Ureaplasma urealyticum | Positive | Positive | 24.8 |
| Infection | Streptococcus agalactiae | Positive | Positive | 24.6 |
| Ureaplasma urealyticum | Positive | Positive | ||
| Ureaplasma parvum | Positive | Negative | ||
| Infection | Candida albicans | Negative | Positive | 24.0 |
| Unknown bacterium | Positive | Negative | ||
| Infection | Ureaplasma urealyticum | Positive | Positive | 22.9 |
| Ureaplasma sp | Positive | Negative | ||
| Infection | Sneathia sanguinegens | Positive | Negative | 21.3 |
| Infection | Bacteroides hemolyticus | Positive | Negative | 21.1 |
| Infection | Fusobacterium nucleatum | Positive | Negative | 20.6 |
| Bacteroides ureolyticus | Positive | Negative | ||
| Infection | Listeria monocytogenes | Positive | Negative | 20.5 |
| Ureaplasma sp | Negative | Positive | ||
| Infection | Bergeyella zoohelecum | Positive | Negative | 20.4 |
| Bergeyella sp | Positive | Negative | ||
| Fusobacterium sp | Negative | Positive | ||
| Clostridium sp | Negative | Positive | ||
| Infection | Streptococcus agalactiae | Positive | Positive | 20.4 |
| Infection | Staphylococcus hemolyticus | Positive | Positive | 20.3 |
| Infection | Leptotrichia amnionii | Positive | Negative | 19.7 |
| Infection | Ureaplasma urealyticum | Positive | Positive | 19.5 |
| Ureaplasma parvum | Positive | Negative | ||
| Infection | Streptococcus agalactiae | Positive | Positive | 19.3 |
| Infection | Ureaplasma urealyticum | Positive | Positive | 12.8 |
| Ureaplasma parvum | Positive | Negative | ||
| Colonization-only | Ureaplasma urealyticum | Negative | Positive | 1.8 |
| Colonization-only | Ureaplasma urealyticum | Negative | Positive | 1.0 |
| Colonization-only | Ureaplasma urealyticum | Negative | Positive | 0.6 |
| Colonization-only | Unknown bacterium | Positive | Negative | 0.4 |
The distribution of amniotic fluid IL-6 concentrations that was used in the designation of the 5 groups is shown in Figure 1 . IL-6 concentration of ≥11.3 ng/mL was found in 63 cases (20.7%), which included 27 cases with MIAC (infection group) and 36 cases without MIAC (severe inflammation group); IL-6 concentration of 2.6-11.2 ng/mL (mild inflammation group) was found in 47 cases (15.4%), none of whom had MIAC; and IL-6 concentration of <2.6 ng/dL was found in 195 cases (63.7%), which included 4 cases with MIAC (colonization group) and 191 cases without MIAC (negative group).
Baseline characteristics
The infection and severe inflammation groups were more prevalent at earlier gestational ages, as shown in Figure 2 , whereas the mild inflammation and colonization groups showed no such trend. Other than the differences in gestational age, the initial clinical presentation was similar in all 5 groups ( Table 3 ). There was no significant difference in maternal age, rate of nulliparity, or cervical dilation. More than 80% of subjects presented with cervical dilation of ≥2 cm. Clinical signs of infection (such as fever, uterine tenderness, maternal leukocytosis, and fetal tachycardia) were uncommon in all 5 groups.
| Variable | Group | P value a | ||||
|---|---|---|---|---|---|---|
| Amniotic fluid interleukin-6, ≥11.3 ng/mL | Amniotic fluid interleukin-6, 2.6-11.2 ng/mL | Amniotic fluid interleukin-6, <2.6 ng/mL | ||||
| Infection (n = 27) | Severe inflammation (n = 36) | Mild inflammation (n = 47) | Colonization (n = 4) | Negative (n = 191) | ||
| Demographics | ||||||
| Maternal age, y b | 25.8 ± 5.4 | 26.6 ± 4.9 | 26.5 ± 5.2 | 25.5 ±5.2 | 25.8 ±5.6 | .92 |
| Nulliparous, n (%) c | 10 (37) | 11 (31) | 19 (41) | 2 (50) | 48 (25) | .17 |
| Presentation | ||||||
| Gestational age, wk b | 27.5 ± 3.9 | 26.8 ± 3.5 | 29.9 ± 3.9 | 30.3 ± 4.7 | 31.1 ± 3.6 | < .0001 |
| Cervical dilation, cm b , d | 3.0 ± 1.6 | 3.2 ± 1.4 | 3.3 ± 1.5 | 2.0 ± 1.4 | 2.7 ± 1.4 | .05 |
| Cervical dilation ≥2 cm, n (%) d | 21 (88) | 27 (90) | 39 (85) | 2 (50) | 134 (80) | .27 |
| Maternal signs, n (%) | ||||||
| Uterine tenderness | 2 (7) | 5 (14) | 2 (4) | 1 (25) | 10 (5) | .13 |
| Fever ≥100.4°F | 1 (4) | 0 | 0 | 1 (25) | 8 (4) | .10 |
| White blood cell count >15,000/mm 3 | 13 (48) | 14 (39) | 5 (11) | 2 (50) | 22 (12) | < .001 |
| Baseline fetal heart rate >160 beats/min | 1 (4) | 0 | 0 | 0 | 4 (2) | .54 |
| Amniotic fluid markers | ||||||
| Glucose <20 mg/dL, n (%) e | 25 (96) | 9 (25) | 6 (13) | 1 (33) | 17 (9) | < .0001 |
| Gram stain positive, n (%) | 14 (52) | 0 | 0 | 0 | 1 (<1) | < .0001 |
| Amniotic fluid interleukin-6, ng/dL f | 41.4 ± 2.5 | 23.0 ± 1.7 | 4.4 ± 16 | 0.8 ± 1.9 | 1.0 ± 2.4 | N/A |
a Comparison of subjects across all 5 groups with analysis of variance or χ 2 test; Fisher exact test was used in lieu of χ 2 for clinical signs of infection because of the small counts
c N = 27, 35, 46, 4, and 191 for the 5 groups, respectively, because of missing values
d N = 24, 30, 46, 4, and 167 for the 5 groups, respectively, because of missing values
e N = 3 for the colonization group and n = 190 for the negative fluid group because of missing values
Amniotic fluid glucose level of <20 mg/dL was highly sensitive for the infection group (96%), but not for the inflammation groups (severe inflammation group, 25%; mild inflammation group, 13%), and had a false-positive rate of 29% (17 of 58 cases with glucose levels of <20 mg/dL were in the negative group). The cutoff value of 20 mg/dL was chosen because some local laboratories do not report a specific value <20 mg/dL because of known nonlinearity in the glucose assay below this level.
Gram stain was highly specific for MIAC (positive predictive value, 93%; 14 of 15), but sensitivity was only 45% (14 of 31).
Amniotic fluid WBC could not be tabulated meaningfully because of major differences in the reporting of results by local hospital laboratories: some laboratories reported total WBC; some laboratories reported only neutrophils, and some laboratories used semiquantitative scales such as 1+, 2+.
Outcomes
Survival curves that show latency are plotted in Figure 3 . The infection and severe inflammation groups had similar short latencies (median, <1 and 2 days, respectively), compared with the 2 groups with low IL-6 concentrations (colonization and negative fluid groups; median, 23.5 and 25 days, respectively). The group with intermediate IL-6 concentration (mild inflammation group) had intermediate latency (median, 7 days). Cases in the infection group had the highest rates of histologic chorioamnionitis and funisitis ( Table 4 ). Clinical chorioamnionitis and postpartum endometritis were uncommon in all 5 groups.
| Variable | Group | P value a | ||||
|---|---|---|---|---|---|---|
| Amniotic fluid interleukin-6, ≥11.3 ng/mL | Amniotic fluid interleukin-6, 2.6-11.2 ng/mL | Amniotic fluid interleukin-6, <2.6 ng/mL | ||||
| Infection (n = 27) | Severe inflammation (n = 36) | Mild inflammation (n = 47) | Colonization (n = 4) | Negative (n = 191) | ||
| Gestational age at delivery, wk b | 27.6 ± 3.9 | 27.9 ± 3.9 | 32.3 ± 4.7 | 33.6 ± 4.4 | 35.5 ± 3.3 | < .0001 |
| Median latency, d c , d | <1 (0–1) | 2 (0–7) | 7 (2–24) | 23.5 (11.5–33) | 25 (6–48) | < .0001 |
| Latency <72 h, n (%) c | 23 (88) | 20 (56) | 16 (34) | 1 (25) | 32 (17) | < .0001 |
| Latency <7 d, n (%) c | 26 (100) | 25 (69) | 23 (49) | 1 (25) | 48 (25) | < .0001 |
| Birth <37 wk, n (%) c | 26 (100) | 35 (97) | 39 (83) | 3 (75) | 122 (64) | < .0001 |
| Clinical chorioamnionitis, n (%) | 4 (15) | 2 (6) | 2 (4) | 1 (25) | 11 (6) | .17 |
| Histologic chorioamnionitis, n (%) e | 22 (85) | 13 (38) | 7 (19) | 0 | 10 (8) | < .0001 |
| Funisitis, n (%) e | 19 (73) | 7 (20) | 7 (19) | 0 | 13 (11) | < .0001 |
| Postpartum endometritis, n (%) | 1 (4) | 1 (3) | 1 (2) | 1 (25) | 6 (3) | .26 |
a Comparison of subjects across all 5 groups with analysis of variance (gestational age), nonparametric analysis of variance (latency), or χ 2 test for other outcomes; for histologic chorioamnionitis, funisitis, and postpartum endometritis, the Fisher exact test was used because of small sample sizes
c N = 26 in the infection group because of missing data
d Data are given as median (interquartile range) for latency
e N = 26, 34, 37, 1, and 120 for the 5 groups, respectively, because of missing data.
Perinatal outcomes are summarized in Table 5 . The infection and severe inflammation groups had similar high rates of perinatal morbidity and mortality; the mild inflammation group had an intermediate rate of perinatal morbidity and mortality.
| Variable | Group | P value a | ||||
|---|---|---|---|---|---|---|
| Amniotic fluid interleukin-6, ≥11.3 ng/mL | Amniotic fluid interleukin-6, 2.6-11.2 ng/mL | Amniotic fluid interleukin-6, <2.6 ng/mL | ||||
| Infection (n = 27) | Severe inflammation (n = 36) | Mild inflammation (n = 47) | Colonization (n = 4) | Negative (n = 191) | ||
| Birthweight, g b | 1165 ± 614 | 1335 ± 839 | 2083 ± 858 | 2206 ± 1102 | 2725 ± 772 | < .0001 |
| Composite perinatal morbidity and death: ≥1 of the following, n (%) | 22 (81) | 26 (72) | 25 (53) | 1 (25) | 41 (21) | < .0001 |
| Perinatal death | 5 (19) | 7 (19) | 3 (6) | 0 | 3 (2) | < .0001 |
| Stillbirth | 0 | 3 | 1 | 2 | ||
| Neonatal death | 5 | 4 | 2 | 1 | ||
| Respiratory distress syndrome, n (%) | 17 (63) | 20 (56) | 23 (49) | 1 (25) | 37 (19) | < .0001 |
| Intraventricular hemorrhage, grade 3 or 4, n (%) | 2 (7) | 3 (8) | 2 (4) | 0 | 4 (2) | .15 |
| Necrotizing enterocolitis, n (%) | 1 (4) | 2 (6) | 0 | 0 | 0 | .02 |
| Culture proven neonatal sepsis, n (%) | 6 (22) | 6 (17) | 4 (9) | 0 | 6 (3) | .001 |
a Comparison of the subjects in the 5 groups by analysis of variance (birthweight) or Fisher exact test (morbidities)
Logistic regression analyses were performed to examine whether MIAC or amniotic fluid IL-6 had associations with either latency of <72 hours or composite perinatal morbidity and death that were independent of gestational age. As shown in model 1 in Table 6 , when adjusted for gestational age at enrollment, both MIAC and IL-6 levels had significant associations with latency of <72 hours, but only IL-6 levels had significant associations with composite perinatal morbidity and death. In model 2, the 5 amniotic fluid groups had strong associations with both latency and composite morbidity and death ( P < .0001 for the across-category trends). We also performed logistic regression analyses adjusting for gestational age at delivery rather than gestational age at enrollment. In these analyses, the gestational age variable dominated, and there was no significant additional contribution of MIAC, amniotic fluid IL-6 (model 3), or the 5 amniotic fluid groups (model 4).
| Predictor variables | Outcome | |||||
|---|---|---|---|---|---|---|
| Latency <72 h | Composite perinatal morbidity and mortality rate | Composite perinatal morbidity and mortality rate, excluding respiratory distress syndrome | ||||
| Odds ratio | 95% CI | Odds ratio | 95% CI | Odds ratio | 95% CI | |
| Model 1 a | ||||||
| Microbial invasion of amniotic cavity amniotic fluid interleukin-6 | 4.3 b | 1.4–13.0 | 1.8 | 0.6–5.5 | 1.4 | 0.3–6.6 |
| ≥11.3 ng/mL | 13.2 c | 5.5–31.8 | 4.9 c | 2.2–11.2 | 4.9 c | 1.9–12.8 |
| 2.6-11.3 ng/mL | 3.2 c | 1.5–6.7 | 3.1 c | 1.5–6.4 | 2.5 | 0.8–7.6 |
| <2.6 ng/mL | Referent | — | Referent | — | Referent | — |
| Model 2 a | ||||||
| Infection | 70.5 c | 17.5–279 | 9.8 c | 3.3–29.5 | 7.0 c | 2.4–20.4 |
| Severe inflammation | 11.7 c | 4.7–29.2 | 4.5 c | 1.9–10.7 | 4.4 b | 1.6–11.8 |
| Inflammation | 3.1 b | 1.4–6.6 | 3.0 b | 1.5–6.4 | 2.2 | 0.8–6.7 |
| Mild colonization | 1.8 | 0.2–18.8 | 0.9 | 0.8–11.2 | N/A | N/A |
| Negative | Referent | — | Referent | — | Referent | — |
| Model 3 d | ||||||
| Microbial invasion of amniotic cavity amniotic fluid interleukin-6 | 1.4 | 0.3–6.6 | 1.1 | 0.3–3.4 | ||
| ≥11.3 ng/mL | 0.2 | 0.1–0.9 | 0.9 | 0.3–2.8 | ||
| 2.6-11.3 ng/mL | 1.2 | 0.4–3.6 | 0.6 | 0.2–2.5 | ||
| <2.6 ng/mL | Referent | — | Referent | — | ||
| Model 4 d | ||||||
| Infection | 0.4 | 0.1–2.5 | 0.9 | 0.2–3.4 | ||
| Severe inflammation | 0.2 | 0.0–0.8 | 0.6 | 0.2–2.3 | ||
| Mild inflammation | 1.2 | 0.4–3.5 | 0.5 | 0.1–2.0 | ||
| Colonization | 0.2 | 0.0–16.1 | N/A | N/A | ||
| Negative | Referent | — | Referent | — | ||
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
