the fetal inflammatory response syndrome

194

The fetal inflammatory response syndrome

Ricardo Gomez, MD,b, c Roberto Romero, MD,a, b Fabio Ghezzi, MD,b Bo Hyun Yoon, MD, PhD,d

Moshe Mazor, MD,e and Stanley M. Berry, MDa

Detroit, Michigan, Bethesda, Maryland, Puente Alto, Chile, Seoul, Korea, and Beer-Sheva, Israel

OBJECTIVE: The objective of this study was to determine the frequency and clinical significance of a sys-temic inflammatory response as defined by an elevated plasma interleukin-6 concentration in fetuses withpreterm labor or preterm premature rupture of membranes.STUDY DESIGN: Amniocenteses and cordocenteses were performed in 157 patients with preterm labor andpreterm premature rupture of membranes. Written informed consent and multi-institutional review board ap-provals were obtained. Amniotic fluid was cultured for aerobic and anaerobic bacteria, as well as mycoplas-mas. Amniotic fluid and fetal plasma interleukin-6 concentrations were measured with a sensitive and spe-cific immunoassay. Statistical analyses included contingency tables, receiver operating characteristic curveanalysis, and multiple logistic regression.RESULTS: One hundred five patients with preterm labor and 52 patients with preterm premature rupture ofmembranes were included in this study. The overall prevalence of severe neonatal morbidity (defined as thepresence of respiratory distress syndrome, suspected or proved neonatal sepsis, pneumonia, bronchopul-monary dysplasia. intraventricular hemorrhage, periventricular leukomalacia, or necrotizing enterocolitis)among survivors was 34.8% (54/155). Neonates in whom severe neonatal morbidity developed had higherconcentrations of fetal plasma interleukin-6 than fetuses without development of severe neonatal morbidity(median 14.0 pg/mL, range 0.5 to 900 vs median 5.2 pg/mL, range 0.3 to 900, respectively; P < .005).Multivariate analysis was performed to explore the relationship between the presence of a systemic fetal in-flammatory response and subsequent neonatal outcome. To preserve a meaningful temporal relationship be-tween the results of fetal plasma interleukin-6 concentrations and the occurrence of severe neonatal morbid-ity, the analysis was restricted to 73 fetuses delivered within 7 days of cordocentesis who survived. Theprevalence of severe neonatal morbidity in this subset of patients was 53.4% (39/73). A fetal plasma inter-leukin-6 cutoff value of 11 pg/mL was used to define the presence of a systemic inflammatory response. Theprevalence of a fetal plasma interleukin-6 level >11 pg/mL was 49.3% (36/73). Fetuses with fetal plasma in-terleukin-6 concentrations >11 pg/mL had a higher rate of severe neonatal morbidity than did those with fetalplasma interleukin-6 levels ≤11 pg/mL (77.8% [28/36] vs 29.7% [11/37], respectively; P < .001). Stepwise lo-gistic regression analysis demonstrated that the fetal plasma interleukin-6 concentration was an independentpredictor of the occurrence of severe neonatal morbidity (odds ratio 4.3, 95% confidence interval 1 to 18.5)when adjusted for gestational age at delivery, the cause of preterm delivery (preterm labor or preterm prema-ture rupture of membranes), clinical chorioamnionitis, the cordocentesis-to-delivery interval, amniotic fluidculture, and anmiotic fluid interleukin-6 results.CONCLUSION: A systemic fetal inflammatory response, as determined by an elevated fetal plasma inter-leukin-6 value, is an independent risk factor for the occurrence of severe neonatal morbidity. (Am J ObstetGynecol 1998;179:194-202.)

Key words: Preterm labor, intrauterine infection, fetal homeostasis, inflammatory response,neonatal morbidity

Microbial invasion of the amniotic cavity is present in10% of patients with preterm labor and intact mem-branes1-4 and in 30% of patients with preterm prematurerupture of membranes.5-8 Moreover, bacterial footprintshave been detected in the amniotic fluid of as many as60% of patients with preterm labor and intact mem-branes.9, 10 Microorganisms in the amniotic cavity or ma-ternal compartment may reach the human fetus andstimulate the biosynthesis of proinflammatorycytokines.11 Increased bioavailability of interleukin-6 (IL-6) during fetal life may lead to the development of anacute-phase response similar to that observed in adult

From the Division of Maternal-Fetal Medicine, Department of Obstetricsand Gynecology, Wayne State University/Hutzel Hospital,a thePerinatology Research Branch, National Institute of Child Health andHuman Development,b the Department of Obstetrics and Gynecology,Sotero del Rio Hospital,c the Department of Obstetrics and Gynecology,Seoul National University,d and the Department of Obstetrics andGynecology, Ben-Gurion University.ePresented in part at the Seventeenth Annual Meeting of the Society ofPerinatal Obstetricians, Anaheim, California, January 20-25, 1997.Received for publication June 23, 1997; revised January 8, 1998; ac-cepted January 15, 1998.Reprint requests: Roberto Romero, MD, Department of Obstetrics andGynecology, Hutzel Hospital, Perinatology Research Branch, NICHD,4707 St. Antoine Blvd, Detroit, MI 48201.6/1/89113

Volume 179, Number 1 Gomez et al 195Am J Obstet Gynecol

patients with systemic inflammatory response syndrome,a serious condition characterized bv multiple organ fail-ure associated with sepsis.12 Circulating human fetal cellsare capable of transcribing messenger ribonucleic acidfor proinflammatory cytokines. We have proposed thatproinflammatory cytokines may serve as fetal signals forthe onset of preterm parturition in the context of in-trauterine infection.13 The objective of this study was todetermine the frequency and clinical significance of asystemic inflammatory response as defined by an ele-vated plasma IL-6 concentration in fetuses with pretermlabor or preterm premature rupture of membranes.

Patients and methods

Patients and eligibility. Women who were examined atHutzel Hospital with preterm labor and intact mem-branes or with preterm premature rupture of mem-branes, over a 4-year period of time, were offered amnio-centesis for the diagnosis of microbial invasion of theamniotic cavity and the assessment of fetal lung maturity.Patients who consented to have an amniocentesis wereasked to participate in a research management protocolthat included the aspiration of additional amniotic fluidfor research purposes and cordocentesis. Criteria for eli-gibility into this study included (1) preterm labor with in-tact membranes or preterm premature rupture of mem-branes, (2) consent to have an amniocentesis, and (3)availability of skilled medical staff to perform cordocen-tesis. Exclusion criteria were (1) clinical chorioamnioni-tis, (2) multiple gestation, (3) fetal distress, or (4) signifi-cant vaginal hemorrhage. Oligohydramnios was not anexclusion criterion. The diagnosis of preterm labor wasmade in the presence of regular uterine contractions (atleast 3 in 30 minutes) and documented cervical changein patients with a gestational age of 20 to 36 weeks.Preterm premature rupture of membranes was diag-nosed by sterile speculum examination with a combina-tion of vaginal pooling and Nitrazine and ferning tests.Written informed consent was obtained from each pa-tient who agreed to participate in the research protocolbefore any procedures were performed. A group of fe-tuses (n = 29) undergoing diagnostic cordocentesiswhose perinatal outcome was normal was used to gener-ate a reference range of fetal plasma IL-6. This controlgroup consisted of fetuses who underwent cordocentesisbetween 19 and 42 weeks of gestation for clinical indica-tions (eg, abnormal ultrasonographic findings, plateletisoimmunization, and suspected fetal infection) and inwhom the clinical conditions for which they had the pro-cedure were excluded after fetal blood and amnioticfluid analysis. This protocol was approved by the humaninvestigation committees of Wayne State University andHutzel Hospital. Similar protocols have been approvedby the institutional review boards of the following institu-

tions: Yale University (New Haven, Connecticut), BenGurion University (Beer-Sheva, Israel), Seoul NationalUniversity (Seoul, Korea), and Sotero del Rio Hospital(an affiliate of the Catholic University of Santiago,Chile). Included in this manuscript are 41 patients withpreterm premature rupture of membranes who were thesubject of separate communications14, 15 about the rela-tionship between fetal plasma IL-6 and the admission-to-delivery interval.

Amniocentesis, cordocentesis, and assays for IL-6. Allpatients had a detailed ultrasonographic examination be-fore amniocentesis and cordocentesis were performed.Electronic fetal monitoring was performed before andafter the procedure to evaluate fetal well-being.Amniocentesis and cordocentesis procedures were per-formed with the freehand technique described byNicolaides et al.16 One percent lidocaine was given as alocal anesthetic, but no sedative drugs were adminis-tered. A 22-gauge needle was used, and a path was cho-sen for needle insertion that allowed the amniocentesisand cordocentesis procedures to be carried out with asingle percutaneous needle insertion in approximately95% of patients. Amniotic fluid studies included a Gramstain and microbial cultures for aerobic or anaerobic bac-teria and mycoplasmas, as well as the lecithin/sphyn-gomyelin ratio. The results of these tests were used forsubsequent clinical management decisions. Fetal bloodwas collected in ethylenediaminetetra-acetic acid(EDTA). Kleihauer-Betke stains were performed on fetalblood, and all specimens were found to be free of mater-nal blood. Fetal blood was analyzed for pH and gases,and complete white blood cell count, platelet count, anddifferential cell count were performed. Results weremade available for clinical management. Amniotic fluidand fetal plasma IL-6 concentrations were determinedwith commercially available enzyme-linked immunoas-says (R & D Systems, Minneapolis). Two assays were run,one for all fetal blood samples and the other for all am-niotic fluid samples. The assays were conducted by thesame person in one laboratory. For amniotic fluid sam-ples, the sensitivity of the immunoassay was 0.5 pg/mL(intra-assay and interassay coefficients of variation 2.7%and 9.3%, respectively). For fetal plasma, the sensitivityof the assay was 0.058 pg/mL (intra-assay and interassaycoefficients of variation 3.3% and 8.3%, respectively).The results of cytokine concentrations in amniotic fluidor fetal plasma reported herein were neither availablenor used for clinical decision-making.

Criteria for the diagnosis of chorioamnionitis and se-vere neonatal morbidity. Clinical chorioamnionitis was di-agnosed in the presence of a temperature elevation to37.8°C or higher and two or more of the following fourcriteria: uterine tenderness, malodorous vaginal dis-charge, fetal tachycardia >160 beats/min, and maternal

leukocytosis >15,000 cells/mm3. The diagnosis of histo-logic chorioamnionitis was based on the polymorphonu-clear leukocyte infiltration of the chorionic plate or ofthe extraplacental fetal membranes. Severe neonatalmorbidity was defined as the presence of any of the fol-lowing conditions: respiratory distress syndrome, sus-pected or proved neonatal sepsis, pneumonia, bron-chopulmonary dysplasia, intraventricular hemorrhage,periventricular leukomalacia, and necrotizing enterocol-itis. The diagnosis of respiratory distress syndrome re-quired the presence of respiratory grunting and retract-ing, increased need for oxygen, and diagnosticradiographic and laboratory findings in the absence ofevidence for other causes of respiratory disease.Neonatal sepsis was diagnosed in the presence of a posi-tive culture of blood, urine, or cerebrospinal fluid.Suspected neonatal sepsis was diagnosed in the absenceof a positive culture when two or more of the followingcriteria were present: (1) white blood cell count of <5000cells/mm3; (2) polymorphonuclear leukocyte count of<1800 cells/mm3; and (3) I:T ratio (ratio of bands tototal neutrophils) >0.2. These criteria have been previ-ously used in the pediatric and obstetric literature.17

Pneumonia was diagnosed in the presence of definiteclinical and radiologic findings, with or without a positiveculture from tracheal aspirate, blood, or chest tube spec-imen. Bronchopulmonary dysplasia was diagnosed if theneonate required oxygen and ventilatory therapy for >28days during the first 2 months of life, had typical radi-ographic changes, or had dysplasia of the bronchopul-monary tree at autopsy. Intraventricular hemorrhage wasdiagnosed by ultrasonographic examination of theneonatal head. Periventricular leukomalacia was diag-nosed in the presence of cystic lesions within the periven-tricular white matter or persistent abnormally increasedperiventricular echogenicity. Necrotizing enterocolitiswas diagnosed in the presence of abdominal distentionand feeding intolerance for at least 24 hours (vomitingor increased gastric residual) with clear radiologic evi-dence of intramural air, perforation, meconium plug syn-drome, or definite surgical or autopsy findings of necro-tizing enterocolitis.

Statistical analysis. Two-tailed Mann-Whitney U testwas used to compare continuous nonparametric vari-ables. Comparisons between proportions were per-formed with χ2 or Fisher’s exact test. Receiver operatingcharacteristic curves were constructed to describe the re-lationship between the sensitivity (true-positive rate) andthe false-positive rate (1 – Specificity) for different valuesof fetal plasma IL-6 concentrations in the prediction ofsevere neonatal morbidity. Logistic regression was usedto investigate the regression relationships between theoccurrence of severe neonatal morbidity and gestationalage at delivery, fetal plasma IL-6 concentrations, the

cause of preterm delivery (preterm labor or preterm pre-mature rupture of membranes), amniotic fluid culture,and amniotic fluid IL-6 results. SPSS version 7.5 (SPSS,Inc., Chicago) and True Epistat (Epistat Services,Richardson, Tex.) statistical packages were used foranalysis.

Results

Study population. One hundred five patients withpreterm labor and 52 patients with preterm prematurerupture of membranes were included in this study. Therewere two perinatal deaths. In both cases the infants weredelivered before 24 weeks of gestation and were ex-cluded from analysis. The overall prevalence of pretermdelivery <37 weeks was 80% (124/155). Severe neonatalmorbidity developed in 54 neonates (34.8%). Table I dis-plays the clinical characteristics of patients according tothe presence or absence of severe neonatal morbidity.Fetuses with subsequent development of severe neonatalmorbidity had a lower mean gestational age at admissionand at delivery, shorter procedure-to-delivery intervals,and a higher proportion of diagnoses of preterm prema-ture rupture of membranes, microbial invasion of theamniotic cavity, clinical chorioamnionitis, and histologicchorioamnionitis, as well as a significantly higher con-centration of IL-6 in amniotic fluid and fetal plasma(Table I and Fig 1, A).

Overall, microbial invasion of the amniotic cavity waspresent in 26.5% (41/155) of patients. The prevalence ofmicrobial invasion was 10.7% (11/103) in patients withpreterm labor and intact membranes and 57.7% (30/52)in patients with preterm premature rupture of mem-branes (P < .005). The most common microbial isolateswere Ureaplasma urealyticum (n = 29), Mycoplasma hominis(n = 10), Fusobacterium species (n = 4), and Streptococcusviridans (n = 4). Other microorganisms includedStreptococcus agalactiae, Gardnerella vaginalis, Haemophilusinfluenzae, Peptostreptococcus, Prevotella, and Candida albi-cans. Histologic evidence of chorioamnionitis was pre-sent in 35.8% (19/53) and 65.9% (27/41) of availableplacentas from patients with preterm labor and pretermpremature rupture of membranes, respectively (P < .01).Fifteen patients had a positive amniotic fluid Gram stain,and the infants of all of these patients were deliveredwithin 25 hours of amniocentesis. Nine deliveries werespontaneous and 6 were either induced or augmented orthe patients underwent cesarean section.

Receiver operating characteristic curve analysis for therelationship between fetal plasma IL-6 and severe neona-tal morbidity. To preserve a meaningful temporal rela-tionship between the results of fetal plasma IL-6 concen-trations and the occurrence of severe neonatalmorbidity, the analysis was restricted to 73 fetuses deliv-ered within 7 days of the cordocentesis who survived.

196 Gomez et al July 1998Am J Obstet Gynecol

The prevalence of severe neonatal morbid events in thissubset was 53.4% (39/73). Receiver operating character-istic curve analysis demonstrated that a value of 11pg/mL was located close to the knee of the curve andtherefore was used to define the presence of a systemicinflammatory response in the fetus (Fig 2). A separatepopulation of 29 fetuses undergoing diagnostic cordo-centesis who subsequently had normal pregnancy out-comes was used to determine the reference range forfetal plasma IL-6 throughout gestation. Fetal plasma IL-6concentrations were mean 4.4 pg/mL, median 3.4pg/mL, and SD 3 pg/mL. The 2 SD above the mean inthis control population coincided with the cutoff of 11pg/mL obtained with receiver operating characteristiccurve analysis in our study population.

The prevalence of a fetal plasma IL-6 value >11 pg/mLwas 49.3% (36/73). Specifically, this condition occurredwith a frequency of 50% (21/42) in patients withpreterm premature rupture of membranes and 48.4%(15/31) in patients with preterm labor and intact mem-branes. Fetuses who had subsequent development ofneonatal morbidity had higher concentrations of plasmaIL-6 than did fetuses who did not have complicationsduring the neonatal period (Fig 1, B). Table II shows theincidence of severe neonatal complications according tothe presence or absence of a fetal inflammatory re-sponse. Overall, fetuses with fetal plasma IL-6 concentra-tions >11 pg/mL had a higher rate of severe neonatalmorbidity than did those with fetal plasma IL-6 ≤11pg/mL (77.8% [28/36] vs 29.7% [11/37]; P < .001).

Significant differences were found for respiratory distresssyndrome (63.9% vs 24.3%) and neonatal sepsis (33.3%vs 8.1%; see Table II). In the 2 cases of perinatal deaththat were excluded from analysis in this study, the fetalplasma IL-6 concentrations were above 100 pg/mL.

Multivariate analysis of the relationship between fetalplasma IL-6 concentrations and the occurrence of severeneonatal morbidity. Stepwise logistic regression was usedto adjust for the effect of several explanatory variables onthe occurrence of adverse neonatal outcome in 59 pa-tients with available amniotic fluid results (ie, IL-6 con-centrations). Because the effect of gestational age at de-livery and fetal plasma IL-6 concentrations on the oddsof development of severe neonatal morbidity was not lin-ear but quadratic, both variables were dichotomized. Forgestational age at delivery, the mean (30.5 weeks) wasused as a cutoff value whereas a fetal plasma IL-6 concen-tration of 11 pg/mL was selected. The diagnosis at ad-mission of the mother (preterm labor or preterm prema-ture rupture of membranes), presence of clinicalchorioamnionitis, amniotic fluid culture, amniotic fluidIL-6 results, procedure-to-delivery interval, and gesta-tional age at delivery ≥30.5 or <30.5 weeks were enteredalong with fetal plasma IL-6 results in a stepwise selectionprocedure that was based on the maximum partial likeli-hood estimates (–2 log likelihood). Gestational age at de-livery and fetal plasma IL-6 concentrations were the onlycovariates that remained significantly associated with thesubsequent development of severe neonatal morbidity(odds ratio 32.8, 95% confidence interval 6.0 to 180.5


Table I. Clinical characteristics of patients according to occurrence of severe neonatal morbidity*

Severe neonatal morbidity

Characteristic Present (n = 54) Absent (n = 101) Statistical significance

Maternal age (y, mean ± SD) 24.1 ± 5.9 24.6 ± 5.5 NSGestational age at admission (wk, mean ± SD) 27.9 ± 3.6 31.8 ± 2.2 P < .05Procedure-to-delivery interval (d) P < .005†

Median 1.6 21Range 0-141 0-86

Gestational age at delivery (wk, mean ± SD) 29.3 ± 3.7 35.1 ± 2.6 P < .005Preterm delivery (<37 wk) 53 (98.1%) 71 (70.3%) P < .05Diagnosis at admission

Preterm labor 24 (44.4%) 79 (78.2%)Preterm premature rupture of membranes 30 (55.6%) 22 (21.8%) P < .05

Positive amniotic fluid culture 29 (53.7%) 12 (11.9%) P < .005Clinical chorioamnionitis 10 (18.5%) 5 (4.9%) P < .05Histologic chorioamnionitis 32/45 (71%) 14/49 (28.6%) P < .005Amniotic fluid IL-6 concentration (ng/mL)‡

Median 11.9 0.9 P < .005Range 0.005-99.5 0.06-64.3

Fetal plasma IL-6 concentration (pg/mL)Median 14.0 5.2 P < .005Range 0.5-900 0.3-900

*Two perinatal deaths were excluded from analysis.†Wilcoxon test for censored data.‡Analysis restricted to 129 patients with fluid available.

and odds ratio 4.3, 95% confidence interval 1 to 18.5, re-spectively). When amniotic fluid IL-6 is excluded and theanalysis is expanded to all 73 patients delivered within 7days of cordocentesis, logistic regression analysis con-firmed that only gestational age at delivery and fetalplasma IL-6 remained significantly associated with thesubsequent development of severe neonatal morbidity(odds ratio 21.0, 95% confidence interval 4.9 to 89.8 andodds ratio 6.0, 95% confidence interval 1.7 to 21.7, re-

spectively). Similarly, when the analysis is restricted to thepatients delivered within 6, 5, 4, 3, 2, or 1 day of amnio-centesis or cordocentesis, the contribution of fetalplasma IL-6 to the occurrence of neonatal morbid eventsremained significant. A plot of the predicted probabilitythat neonatal morbidity will develop according to gesta-tional age at delivery for fetuses with and without a sys-temic inflammatory response is displayed in Fig 3.

Table III shows the incidence of severe neonatal com-


Fig 1. Fetuses with subsequent development of severe neonatal morbidity had higher plasma IL-6 concentrations thanfetuses without severe neonatal morbidity. A, Whole study population: median 14.0 pg/mL, range 0.5 to 900 vs median5.2 pg/mL, range 0.3 to 900, respectively; P < .005; B, patients delivered within 7 days of cordocentesis: median 20.3pg/mL [0.5 to 900] vs median 5.3 pg/mL [1.7 to 900], respectively; P < .005).

plications according to amniotic fluid culture results andfetal plasma IL-6 concentrations in patients deliveredwithin 1 week of cordocentesis. The mere presence of mi-crobial invasion of the amniotic cavity was not associatedwith an increase in neonatal complications (25.9% vs40%, groups 1 and 2, Table III). Moreover, among pa-tients with microbial invasion of the amniotic cavity(groups 2 and 4), the presence of a fetal inflammatory re-sponse (group 4) was associated with a significant in-crease in the rate of neonatal morbidity (from 40%[4/10] to 84.6% [22/26]; P < .01). Although gestationalage at delivery in group 4 was lower than in group 2 (29.3vs 31.9 weeks), this difference was not significant.

Table IV displays the different prenatal interventionsperformed on patients delivered within 7 days of cordo-centesis, according to diagnosis at admission and fetalplasma IL-6 concentrations. No significant differenceswere observed in the frequency of prenatal interventionsincluding tocolysis, antibiotics, corticosteroids, and indi-cated delivery for maternal or fetal reasons between pa-tients with and those without fetal inflammatory re-


sponse in both subgroups, patients with preterm prema-ture rupture of membranes and patients with pretermlabor and intact membranes (P > .05 for all comparisonsbetween patients with and without a fetal inflammatoryresponse; see Table IV).

Comment

Our results demonstrate that a fetal plasma IL-6 levelabove 11 pg/mL is a major independent risk factor forthe subsequent development of severe neonatal morbid-ity. The prevalence of an elevated fetal plasma IL-6 valuein patients with preterm labor or premature rupture ofmembranes delivered within 1 week of cordocentesis wasabout 50% for either condition.

Microbial invasion of the amniotic cavity was associ-ated wiith a fetal plasma IL-6 value >11 pg/mL (63.4%[26/41] vs 14.9% [17/114], for fetuses with and with-out microbial invasion of the amniotic fluid, respec-tively; P < .001), suggesting that fetal exposure to mi-croorganisms and their products may result in theelevation of plasma IL-6 in the fetus. Although 63.4% of

Fig 2. Receiver operating characteristic curve analysis for rela-tionship between severe neonatal morbidity and concentrationsof fetal plasma IL-6.

Fig 3. Plot of predicted probability of severe neonatal morbidityas a function of gestational age at delivery and presence or ab-sence of fetal plasma IL-6 >11 pg/mL.

Table II. Severe neonatal morbidity according to fetal plasma IL-6 concentrations in 73 patients delivered within 7 daysof amniocentesis or cordocentesis

Fetal plasma IL-6Statistical

Morbidity >11 pg/mL (n = 36) ≤11 pg/mL (n = 37) significance

Respiratory distress syndrome 23 (63.9%) 9 (24.3%) P < .005Proved or suspected sepsis 12 (33.3%) 3 (8.1%) P < .01Intraventricular hemorrhage (grades III and IV) 1 (2.8%) 4 (10.8%) NSPeriventricular leukomalacia 2 (5.6%) 1 (2.7%) NSNecrotizing enterocolitis 1 (2.8%) 2 (5.4%) NSBronchopulmonary dysplasia 4 (11%) 2 (5.4%) NSOverall severe neonatal morbidity 28 (77.8%) 11 (29.7%) P < .001


fetuses with proved microbial invasion of the amnioticcavity had an elevated fetal plasma IL-6 value, subsequentmicrobial cultures obtained during the early neonatal pe-riod were positive in only one case. A potential explana-tion for the failure to recover microorganisms is that pre-natal administration of antibiotics occurred in 59.7% ofpatients delivered within 1 week of admission (43/72, seeTable IV). Another explanation is the limitation of stan-dard microbiologic techniques in the isolation of mi-croorganisms from neonates. Neonatal cultures for my-coplasmas were not performed in this study. Thelikelihood of microbial recovery is a function of the vol-ume and timing of collection of neonatal blood used forculture. Inasmuch as neonatal bacteremia is unlikely tobe continuous and that chills and fever (clinical indica-tors of bacteremia used to time blood drawing in adults)rarely develop in neonates, a random blood culture witha small blood volume may not be an adequate method todetect congenital sepsis. Finally, microbial products re-leased after bacterial death (i.e., endotoxin) may be re-sponsible for fetal immune activation but bacteria may

not be recoverable with standard microbiologic tech-niques.

The hypothesis that exposure of fetal cells to microor-ganisms may account for the elevated fetal plasma IL-6 inour study is supported by the observations that cells col-lected from the umbilical cord blood can produce proin-flammatory cytokines in vitro in response to microbialproducts.

It is noteworthy that a fetal plasma IL-6 >11 pg/mL wasdetected in the absence of microbial invasion of the am-niotic cavity in 10 cases (see Table III, group 3). Thesepatients had a higher rate of neonatal morbidity (al-though not significantly, 60% vs 25.9%, power [1 – β] =.49) than patients with fetal plasma IL-6 <11 pg/mL(group 1 in Table III). What is the etiology of the ele-vated fetal plasma IL-6 in these patients? Other than anincrease in amniotic fluid IL-6 concentrations, evidenceof infection could not be found in most cases; amnioticfluid cultures were negative, none of the mothers hadclinical chorioamnionitis, and only 1 newborn had sus-pected sepsis. Of 9 placentas examined, 4 had histologic

Table III. Severe neonatal morbidity according to presence of microbial invasion of amniotic cavity and fetal plasma IL-6 concentration in 73 patients delivered within 7 days of amniocentesis or cordocentesis

AmnioticSevere Gestational fluid IL-6

Depiction neonatal age at delivery (ng/mL, median,Group of group No. morbidity (%) (wk, mean ± SD) range) (n = 59)

Group 1 27 7 (25.9%) 32.1 ± 3.1 1.2 (0.1-60.7)Negative amniotic fluid cultureFetal plasma IL-6 ≤11 pg/mL

Group 2 10 4 (40%) 31.9 ± 2.3 2.0 (0.08-19.8)Positive amniotic fluid cultureFetal plasma IL-6 ≤11 pg/mL

Group 3 10 6 (60%) 30.1 ± 4.9 22.0 (0.5-99.5)Negative amniotic fluid cultureFetal plasma IL-6 >11 pg/mL

Group 4 26 22 (84.6%) 29.3 ± 2.9 36.7 (0.5-92.8)Positive amniotic fluid cultureFetal plasma IL-6 >11 pg/mL

White color in fetal or amniotic fluid compartment represents low fetal plasma IL-6 or negative amniotic fluid culture, respectively. Blackarea in fetal or amniotic fluid compartment denotes elevated fetal plasma IL-6 or positive amniotic fluid culture, respectively.

Table IV. Prenatal interventions in 73 patients delivered within 7 days of amniocentesis or cordocentesis

Fetal inflammatory response Fetal inflammatory response syndrome in patients syndrome in patients with preterm

with preterm labor (n = 31) premature rupture of membranes (n = 42)

Intervention Positive Negative Positive Negative

Tocolytics 14/15 (93%) 16/16 (100%) 1/21 (5%) 1/21 (5%)Antibiotics 9/15 (60%) 7/16 (44%) 14/20* (70%) 13/21 (62%)Corticosteroids 10/15 (67%) 11/16 (69%) 7/21 (33%) 9/21 (43%)Censored observations 2/15 (13%) 1/16 (6%) 7/21 (33%) 8/21 (38%)

*Information about antibiotic administration could not be ascertained in 1 case.

evidence of acute chorioamnionitis. In view of these find-ings, two explanations must be considered as potentialetiologies: infection that escaped detection (caused byviruses, Chlamydia, or other fastidious microorganisms)or a non-infection-related process. The absence of histo-logic chorioamnionitis in 56% of cases argues in favor ofa noninfectious process in these patients. In adults, othercauses of the systemic inflammatory response syndromeinclude trauma, burns, and pancreatitis. Further studiesare required to establish the pathologic process responsi-ble for fetal systemic inflammatory response in the ab-sence of infection. Because fetuses affected by this condi-tion were delivered preterm and had significant neonatalcomplications, an elevated fetal plasma IL-6 cannot beconsidered to be a benign state.

A major finding of our study is that an elevated fetalplasma IL-6 was an independent predictor of neonatalmorbid events even after we corrected for gestational ageat birth and other confounders (odds ratio 4.3, 95% con-fidence interval 1 to 18.5). These findings suggest thatthe pathophysiologic derangements responsible forsome neonatal complications traditionally attributed toimmaturity begin before birth. Such an interpretation isconsistent with several recent observations which suggestthat an inflammatory process already present at birthmay mediate short- and long-term complications of pre-maturity.18, 19 Indeed, neonates born with elevated um-bilical cord levels of IL-6 are at risk for the developmentof brain white matter lesions20 and congenital neonatalsepsis.21 Moreover, neonates born to women with ele-vated amniotic fluid IL-6 and interleukin-8 and histologicchorioamnionitis are at risk for bronchopulmonary dys-plasia.22, 23 Collectively, these findings suggest that fetalplasma IL-6 elevation is an indicator that the fetus hasbeen the subject of an insult severe enough to elicit a sys-temic inflammatory response.

We have proposed that the onset of preterm labor hassurvival value and that it is initiated when the intrauter-ine environment is so hostile that it threatens the survivalof the maternal-fetal pair.24, 25 The evidence presentedherein supports the concept that a systemic fetal inflam-matory response may be detected in a population ofwomen in preterm labor and preterm premature ruptureof membranes. This condition is frequently associatedwith microbial invasion of the amniotic cavity, but thereare probably other pathologic conditions yet to be eluci-dated that are potential causes of this state. An elevatedfetal plasma IL-6, a marker of a systemic fetal inflamma-tory response, should probably be considered a syn-drome (the fetal inflammatory response syndrome) in amanner similar to hyperglycemia and hypercholes-terolemia. Previous studies in which serial ultrasono-graphic examination of the fetus with preterm prema-ture rupture of membranes was performed suggest thatdecreased biophysical activities (fetal breathing move-

ments and gross body movements) were associated withinfection and the onset of spontaneous labor.17 Similarly,a decrease in breathing movements has been reported infetuses during preterm labor and delivery.26

Proinflammatory cytokines such as interleukin-1 andtumor necrosis factor have direct effects on the centralnervous system, induce somnolence, and thus may de-crease gross body movements. Tumor necrosis factor andinterleukin-1–induced prostaglandin E biosynthesis candepress total breathing movements. Therefore a de-crease in biophysical activity may be the “fetal clinicalmanifestation” of the fetal inflammatory response syn-drome, whereas preterm labor or preterm prematurerupture of membranes (myometrial contractions, cervi-cal dilatation, and membrane and decidual activationand rupture) would be the maternal counterpart.Further study is required to determine the natural his-tory, optimal diagnostic workup, and treatment of thiscondition. The concept that the fetus of the mother withpreterm labor or preterm premature rupture of mem-branes may be critically ill calls for a reappraisal of thecurrent clinical management of these conditions. It re-mains to be determined whether efforts to prolong in-trauterine stay with tocolysis or expectant managementin fetuses with severe subclinical disease is a wise courseof action.

We acknowledge the contribution of Dr Robert Sokol, DrDavid Cotton, Dr Ruben Quintero, Dr Mark Evans, DrMitchell Dombrowski, Dr Yoram Sorokin, Dr MaurizioGalasso, Dr Karoline Puder, Sandy Field, RN, Carolyn Sudz,RN, Adriana Soto, RDMS, Mary King, RDMS, Tess Tagle,RN, and Elisa Walsh, RN, of Wayne State University/theDetroit Medical Center; Dr John C. Hobbins, Dr E. AlbertReece, Dr Alan H. DeCherney, Dr Maurice H. Mahoney, DrFrederick Naftolin, and Virginia Sabo, RN, of YaleUniversity; Dr Ivan Rojas, Dr Ernesto Behnke, and DrEnrique Oyarzun of the Sotero del Rio Hospital andCatholic University in Santiago, Chile; Dr Vaclav Insler ofthe Ben Gurion University of Israel; and Dr Neil Athayde,Dr Eli Maymon, and Dr Percy Pacora of the PerinatologyResearch Branch of the National Institute of Child Healthand Human Development for their support and advice.

REFERENCES

1. Arias F, Rodriquez L, Rayne S, Kraus F. Maternal placental vas-culopathy and injection: two distinct subgroups among patientswith preterm labor and preterm ruptured membranes. Am JObstet Gynecol 1993;168:585-91.

2. Duff P, Kopelman J. Subclinical intra-amniotic infection inasymptomatic patients with refractory preterm labor. ObstetGynecol 1987;69:756-9.

3. Romero R, Sirtori M, Oyarzun E, Avila C, Mazor M, Callahan R,et al. Infection and labor. V. Prevalence, microbiology, and clin-ical significance of intraamniotic infection in women withpreterm labor and intact membranes. Am J Obstet Gynecol1989;161:817-24.

4. Romero R, Yoon BH, Mazor M, Gomez R, Diamond MP, KenneyJS, et al. The diagnostic and prognostic value of amniotic fluidwhite blood cell count, glucose, interleukin-6, and Gram stain in



patients with preterm labor and intact membranes. Am J ObstetGynecol 1993;169:805-16.

5. Garite T, Freeman R. Chorioamnionitis in the preterm gesta-tion. Obstet Gynecol 1982;59:539-45.

6. Romero R, Quintero R, Oyarzun E, Wu YK, Sabo V, Mazor M,Hobbins JC. lntra-amniotic infection and the onset of labor inpreterm premature rupture of membranes. Am J ObstetGynecol 1988;159:661-6.

7. Feinstein S, Vintzileos A, Lodeiro J, Campbell W, Weinbaum P,Nochimson DJ. Amniocentesis with premature rupture of mem-branes. Obstet Gynecol 1986;68:147-52.

8. Romero R, Yoon BH, Mazor M, Gomez R, Gonzales R, DiamondMP, et al. A comparative study of the diagnostic performance ofamniotic fluid glucose, white blood cell count, interleukin-6,and Gram stain in the detection of microbial invasion in pa-tients with preterm premature rupture of membranes. Am JObstet Gynecol 1993;169:839-51.

9. Markenson G, Martin R, Foley K, Yancey M. The use of poly-merase chain reaction to detect bacteria in amniotic fluid inpregnancies complicated by preterm labor [abstract 100]. Am JObstet Gynecol 1997;176:S39.

10. Jalava J, Mantymaa ML, Ekblad U. Bacterial 165 rDNA poly-merase chain reaction in the detection of intra-amniotic infec-tion. Br J Obstet Gynaecol 1996;103:664-9.

11. Gomez R, Romero R, Edwin S, David C. Pathogenesis ofpreterm labor and preterm premature rupture of membranesassociated with intraamniotic infection. Infect Dis Clin NorthAm 1997;11:135-76.

12. Rangel-Frausto M, Pittet D, Costigan M, Hwang T, Davis CS,Wenzel RP. The natural history of the systemic inflammatory re-sponse syndrome (SIRS): A prospective study. JAMA1995;273:117-23.

13. Romero R, Brody DT, Oyarzun E, Mazor M, Wu YK, Hobbins JC,et al. Infection and labor. III. Interleukin 1: a signal for theonset of parturition. Am J Obstet Gynecol 1989;160:1117-23.

14. Romero R, Gomez R, Ghezzi F, Yoon BH, Mazor M, Edwin SS,Berry SM. The onset of spontaneous preterm parturition is pre-ceded by an intense pro-inflammatory cytokine response inhuman fetus [abstract 8]. Am J Obstet Gynecol 1997;176:S3.

15. Romero R, Gomez R, Ghezzi F, Yoon BH, Mazor M, Edwin SS, etal. A fetal systemic inflammatory response is followed by the

spontaneous onset of preterm parturition. Am J Obstet Gynecol1998;179:186-93.

16. Nicolaides K, Soothill P, Rodeck C, Campbell S. Ultrasoundguided sampling of umbilical blood and placental blood to as-sess fetal well being. Lancet 1986;1:1065-7.

17. Vintzileos A, Campbell W, Nochimson D, Connolly ME, FuenjerMM, Hoehn GJ. The fetal biophysical profile in patients withpremature rupture of membranes—an early predictor of fetalinfection. Am J Obstet Gynecol 1985;152:510-6.

18. Verma U, Tejani N, Klein S, Reale M, Beneck D, Figueroa D, etal. Obstetric antecedents of intraventricular hemorrhage andperiventricular leukomalacia in the low-birth-weight neonate.Am J Obstet Gynecol 1997;176:275-81.

19. Rojas M, Gonzalez A, Bancalari E, Claure N, Poole C, Silva-NetoG. Changing trends in the epidemiology and pathogenesis ofneonatal chronic lung disease. J Pediatr 1995;126:605-10.

20. Yoon BH, Romero R, Yang SH, Jun JK, Kim IO, Choi JH, et al.Interleukin-6 concentrations in umbilical cord plasma are ele-vated in neonates with white matter lesions associated withperiventricular leukomalacia. Am J Obstet Gynecol1996;174:1433-40.

21. Buck C, Bundschu J, Gallati H, Bartmann P, Pohlandt F.Interleukin-6: a sensitive parameter for the early diagnosis ofneonatal bacterial infection. Pediatrics 1994;93:54-8.

22. Ghezzi F, Gomez R, Romero R, David C, Field S, Edwin SS, et al.Evidence that the injury responsible for the development ofbronchopulmonary dysplasia may begin in utero [abstract 131].Am J Obstet Gynecol 1997;176:S47.

23. Yoon BH, Romero R, Kim CJ, Koo JN, Choe G, Syn HC, et al.High expression of tumor necrosis factor-α and interleukin-6 inperiventricular leukomalacia. Am J Obstet Gynecol1997;177:406-11.

24. Romero R, Sepulveda W, Baumann P, Yoon BH, Brandt F,Gomez R, et al. The preterm labor syndrome: biochemical, cy-tologic, immunologic, pathologic, microbiologic, and clinicalevidence that preterm labor is a heterogeneous disease [abstract9]. Am J Obstet Gynecol 1993;168:288.

25. Romero R, Mazor M, Munoz H, Gomez R, Galasso M, Sherer D.The preterm labor syndrome. Ann N Y Acad Sci 1994;734:414-29.

26. Vintzileos A, Campbell W, Rodis J. Fetal biophysical profile scor-ing: current status. Clin Perinatol 1989;16:661-89.

the fetal inflammatory response syndrome

Documents