Pettker, Christian M. MD; Buhimschi, Irina A. MD; Magloire, Lissa K. MD; Sfakianaki, Anna K. MD; Hamar, Benjamin D. MD; Buhimschi, Catalin S. MD
The rate of preterm birth has increased to 12.3% of all pregnancies in the United States.1 Magnifying and exemplifying its significance in contemporary obstetrics, prematurity accounts for nearly three quarters of cases of perinatal morbidity and mortality.2 Intense effort is dedicated to understanding the various causes contributing to preterm birth, and several pathways have been implicated in its pathogenesis.3 Of these, intraamniotic infection and inflammation stand out as leading contributors not only to preterm delivery but also to its adverse consequences.4 Antenatal infections, particularly chorioamnionitis, have been shown to be significantly related to both short- and long-term damage.5,6 Furthermore, neonatal infections, many of which begin at the end of gestation, are leading contributors to long-term adverse neurodevelopmental outcomes.7
Direct analysis of the amniotic fluid is the best method of assessing intraamniotic infection and inflammation, and several diagnostic clinical tests of the amniotic fluid are available to the clinician, including microbiologic culture, Gram stain, l-lactate dehydrogenase activity, glucose, neutrophil count, and most recently, proteomic profiling (Mass Restricted scoring).4,8,9 However, limiting its use, amniocentesis can be technically challenging and its role in the clinical management of patients presenting with preterm labor and preterm premature rupture of the membranes (PROM) continues to be subject to debate. Therefore, surrogate tests that do not require an amniocentesis have been proposed.10,11 For instance, prior studies have proposed that placental histology may correlate with intraamniotic infection.12–14 Moreover, it is common practice, in patients with risk factors for intraamniotic infection or when retrieval of the amniotic fluid is technically impossible (eg, anhydramnios or precipitous preterm delivery), to perform placental evaluation with histologic analysis and microbial cultures to speculate upon the presence or absence of intraamniotic infection or inflammation and thus aid in future counseling. Yet, the scientific basis of this practice is lacking.
We hypothesized that microbiologic and histologic evaluation of the placenta for acute inflammation at the time of delivery accurately reflects the presence and severity of amniotic fluid infection and inflammation. Our objective was to evaluate the ability of microbiologic and histologic placental examination to accurately reflect the microbial and inflammatory status of the amniotic fluid near the time of delivery.
MATERIALS AND METHODS
A total of 183 consecutive women undergoing transabdominal amniocentesis under a prospective research protocol at Yale-New Haven Hospital from September 2004 to January 2006 were eligible for this study (Fig. 1). The Human Investigational Committee of Yale University School of Medicine approved this study, and written informed consent was obtained from all participants before the procedure. One hundred thirty-eight women presented with clinical signs or symptoms of preterm labor or preterm PROM and had a clinically indicated amniocentesis to rule out infection (cases). This represents approximately 20% of all preterm deliveries at our institution during this time. Forty-five women in the third trimester had amniocentesis to determine fetal lung maturity before delivery (controls). Indications for fetal lung maturity testing included previous uterine surgery, history of previous adverse pregnancy outcome, gestational hypertension, and timing of anticoagulation. The decision for amniocentesis in both groups was made independently of our study protocol, and clinical management was based on existing obstetric standards independent of our protocol.
From the time of amniocentesis, each woman was followed up to the point of delivery. Induction of labor or a surgical delivery was performed for clinical indications such as nonreassuring fetal testing, amniotic fluid laboratory results interpreted by the provider indicating intraamniotic inflammation or infection, documented fetal lung maturity, prolapsed umbilical cord, or a gestational age 34 weeks or more. Exclusion criteria for entry into this study included precipitous delivery, human immunodeficiency virus (HIV) or hepatitis infections, or presence of fetal heart rate abnormalities at the time of enrollment (bradycardia, or prolonged variable decelerations).
To maintain a meaningful temporal relationship between the microbiologic results of the amniotic fluid and placental tissues, we applied our analysis in the study group only to the cases delivered within 48 hours relative to the return of the rapid clinical tests or the microbiologic cultures. This time period was chosen to provide a significant duration of time to elapse between amniocentesis and delivery, but also to limit confounding factors influencing the contamination of the placental flora between amniocentesis and delivery, and is a similar method to previous studies.15–17 This resulted in a study group consisting of 56 women. Of these, 26 women had proven positive amniotic fluid cultures, and 30 women had negative amniotic fluid cultures.
Similarly, in the control group we applied our analysis only to those patients delivered with “positive” lung maturity testing, expecting delivery of these patients within 48 hours. At Yale-New Haven Hospital an amniotic fluid lecithin/sphingomyelin (L/S) ratio 2.5 or more is considered to indicate fetal lung maturity.18 Twenty-three women met the inclusion criteria and served as controls. None had symptoms of labor, ruptured membranes, or signs or symptoms of clinical chorioamnionitis. After amniocentesis all of the control patients had their cesarean delivery procedure or labor induction initiated within 24 hours.
Gestational age was established based on an ultrasonographic examination before 20 weeks in all instances. Preterm labor was defined as regular uterine contractions associated with cervical dilation or effacement before 37 weeks gestational age.19 Membrane rupture was confirmed by clinical tests during speculum examination (vaginal pooling, Nitrazine (Bristol-Myers Squibb, Princeton, NJ) positivity, ferning visible on microscopy) before amniocentesis or a positive indigo carmine amniocentesis dye test. In preterm PROM patients, digital exams were not permitted. Patients received corticosteroids for lung maturity and antibiotic therapy (ampicillin and erythromycin or clindamycin) if their gestational age was less than 32 weeks. Clinical chorioamnionitis was diagnosed in the presence of maternal fever (more than 37.8°C), uterine tenderness, foul smelling amniotic fluid or visualization of pus at the time of the speculum examination, and maternal (100 beats per minute or more) or fetal tachycardia (160 beats per minute or more).6
Amniotic fluid was analyzed through currently used or proposed tests of infection and inflammation. Glucose concentration, l-lactate dehydrogenase activity, neutrophil (white blood cell, WBC) count, and Gram stain were performed in our clinical laboratory at Yale-New Haven Hospital. In addition, amniotic fluid was sent for microbial culture for aerobic and anaerobic microbes and Ureaplasma and Mycoplasma species to our clinical microbiologic laboratory. All of these results were available to the primary providers responsible for clinical management.
Enzyme-linked immunosorbent assays for human interleukin (IL)-6 (Pierce-Endogen, Rockford, IL) and matrix metalloproteinase-8 (R&D Systems, Minneapolis, MN) were performed in duplicate according to the manufacturers' instructions by investigators unaware of sample origin. The minimal detectable concentration for IL-6 was 1 pg/mL and less than 0.02 ng/mL for matrix metalloproteinase-8. The interassay and intra-assay coefficients of variation were less than 10% for IL-6 and less than 6% for matrix metalloproteinase-8. An amniotic fluid concentration above 11.4 ng/mL for IL-6 and 23 ng/mL for matrix metalloproteinase-8 were considered indicative of intraamniotic inflammation or infection.15,20
A proteomic fingerprint, the Mass Restricted score, was immediately generated from the fresh amniotic fluid by laboratory personnel blinded to clinical information. The methodology for the generation of the Mass Restricted score by proteomics assessment (surface-enhanced-laser-desorption-ionization time-of-flight) was previously described in detail.4,21,22 The Mass Restricted score provides qualitative information regarding the presence or absence of intraamniotic inflammation. The Mass Restricted score ranges from 0 to 4, depending upon the presence or absence of four protein biomarkers (defensin-1, defensin-2, calgranulin-A, calgranulin-C). A categorical value of 1 is assigned if a particular peak is present and 0 if absent. A score of 3 or 4 indicates the presence of inflammation, whereas a score of 0–2 excludes it. The Mass Restricted score results were not used for clinical management.
At Yale-New Haven Hospital, placental pathologic study is a routine part of the evaluation of a pregnancy complicated by preterm labor or preterm PROM; for the purposes of this study, placentas from the control group were evaluated in the same manner. Immediately after delivery and under sterile conditions, culture specimens were obtained by one of two authors (C.M.P. or C.S.B.) with a placental swab of the fetal side of the placenta, an approximately 1 cm2 portion of amnion-chorion taken at a location away from the site of membrane rupture (amnion-chorion biopsy), and an approximately 1 cm3 partial-thickness section from the fetal side of the placenta (placenta biopsy) (Fig. 2). These were sent for general microbiologic culture for aerobic and anaerobic microbes and Ureaplasma and Mycoplasma species at the same clinical laboratory as the amniotic fluid cultures. Histologic examination of the placenta was performed by a blinded perinatal clinical pathologist according to recognized standards of inflammatory grades and stages. From each placenta, sections of chorionic plate, extraplacental membranes, and umbilical cord were examined systematically for inflammation. Three histologic stages of chorioamnionitis (stage I: intervillositis, stage II: chorionic inflammation, and stage III: full-thickness inflammation of both chorion and amnion) were complemented by the histologic grading system devised by Salafia et al,23,24 which includes four grades of inflammation of the amnion, chorion-decidua, and umbilical cord. The pathologist was unaware of the results of the amniotic fluid clinical laboratory or proteomic analyses.
We subjected all data sets to normality testing using the Kolmogorov-Smirnov test. Statistical analyses were performed with Sigma Stat v.23 (SPSS Inc., Chicago, IL) and MedCalc (Broekstraat, Belgium) statistical software. Data were compared with Student t test, one-way analysis of variance followed by Student-Newman-Keuls (parametric) or Kruskal-Wallis on ranks followed by Dunn's tests (nonparametric), to adjust for multiple comparisons as appropriate. Comparisons between proportions were done with χ2 testing. Estimates of the accuracy, sensitivity, and specificity and the positive and negative predictive values were computed for each test. Two-by-two contingency tables were constructed and χ2 analysis of independence used to identify significant differences among test performances. Variables (glucose, L-lactate dehydrogenase, WBC, IL-6, matrix metalloproteinase-8) were dichotomized based on previously defined clinical thresholds (less than 10 mg/dL, more than 419 units/L, more than 100 cells/mm3, more than 11.4 ng/mL, and more than 23 ng/mL, respectively) (Buhimschi CS, Bhandari V, Hamar BD, Bahtiyar MO, Zhao G, Sfakianaki A, et al. Proteomic profiling of the amniotic fluid to detect inflammation, infection and neonatal sepsis. PLoS Med 2007; in press).8,15,20,21,25 We assigned placental inflammation as “present” or “absent” based on the presence or absence of inflammatory cells in any of the three analyzed placental compartments. Patients with missing values for a test were not included in the analysis for that particular test. We performed Cohen's κ as a measure of agreement, and the degree of concordance was appreciated after the scale suggested by Landis and Koch (0.21–0.40 “fair”; 0.40–0.60 “moderate”; 0.61–0.80 “substantial”; 0.81–1.00 “almost perfect”).26 A P value of less than .05 was used to indicate significance. Stepwise logistic regression was used for multivariate analysis using dependent and independent variables as specified with a P<.05 used for variable entry and a P≥.1 for variable removal. The primary outcome for this study was the accuracy of a placental culture to reflect the amniotic fluid culture. Assuming that a valuable test will be able to predict correctly the presence or absence of infection for 80% of measurements (compared with 50% accuracy for chance alone) we estimated a minimal required sample size of 45, given 80% power and a confidence coefficient of 95% (one-sample test of hypothesis for proportion; PASS 2005, NCSS Statistical Software, Kaysville, UT).20,27,28
We present the demographics, clinical, and outcome characteristics of the patients in Table 1. There were no significant differences in these characteristics between the two preterm study (positive amniotic fluid culture and negative amniotic culture) groups. The third trimester control group differed from each of the study groups with respect to gestational age, the incidences of ruptured membranes, and symptomatic uterine contractions. The three groups did not significantly differ in rates of clinical symptoms of chorioamnionitis, although this may be due to sample size. The median amniocentesis-to-delivery intervals for both positive amniotic fluid culture and negative amniotic culture groups were similar, each significantly different from the third trimester controls. The third trimester control group had a higher incidence of cesarean delivery compared with the positive amniotic fluid culture and negative amniotic culture groups. Compared with the negative amniotic culture patients and third trimester controls, positive amniotic fluid culture women had significantly lower glucose concentrations, higher median levels of l-lactate dehydrogenase, WBC, IL-6, and matrix metalloproteinase-8 and an increased frequency of a positive Gram stain and Mass Restricted scores of 3–4 (Kruskal-Wallis analysis of variance, P<.05) (Table 2). None of the third trimester control patients tested with an Mass Restricted score of 3 or 4.
There were no differences notable in placental culture results between the positive amniotic fluid culture and negative amniotic culture groups. However, we determined that the incidence of recovering positive cultures from the placenta, in all categories, was significantly higher in both the positive amniotic fluid culture and negative amniotic culture groups when compared with third trimester controls (Table 3). Women with positive amniotic fluid cultures had higher severities (median grades) of funisitis and acute histologic chorioamnionitis compared with those with negative amniotic cultures and controls (P<.05). Patients with symptoms of preterm birth but negative cultures had higher rates of placental inflammation than third trimester controls as assessed by funisitis and acute histologic chorioamnionitis (P<.05).
When broken down into groups, 92% of women with positive amniotic fluid cultures tested with at least one positive placenta culture (24 of 26) (Fig. 3). Eighty percent of those women who had negative amniotic fluid cultures also tested with positive placenta cultures (24 of 30) and 52% of the third trimester control patients, none of which had a positive amniotic fluid culture, tested with at least one positive placenta culture test result (χ2 11.2, P=.004).
We further evaluated the performance of placenta cultures to predict amniotic fluid microbiologic results in women with positive amniotic fluid culture and negative amniotic culture for each independent culture type as well as all three combined (Table 4). The accuracy of predicting a positive amniotic fluid culture result varied from 44% to 57% when each test was analyzed separately or combined, and no confidence interval (CI) crossed 80%. Of all tests, the one that performed best for correlation was the placental biopsy, which achieved only the level of “fair concordance” as assessed by Cohen κ analysis. In addition, the presence of any grade of acute histologic chorioamnionitis had 100% sensitivity (95% CI 83–100%), 100% negative predictive value (95% CI 56–100%) but only 23% specificity (95% CI 11–43%), 52% positive predictive value (95% CI 37–66%), and 58% accuracy (95% CI 44–71%) in identifying women with positive amniotic fluid culture. In summary, this suggests that microbiologic and histologic studies of amnion-chorion biopsy, placenta biopsy, and placenta swab cultures perform poorly and have low accuracy in predicting amniotic fluid infection.
Table 5 presents the seven most frequent isolates from amniotic fluid cultures and placental cultures. Our observation demonstrated that Ureaplasma urealyticum was the most common isolate from both amniotic fluid and placenta cultures and that the spectrum of these organisms was consistent with other published studies.12–14,29,30 To assess for the concordance between the amniotic fluid and placental microbiologic organisms, we calculated the rate of concordance between cultures, restricting the analysis to patients with proven amniotic fluid infection (positive amniotic fluid culture). Overall, there was only a 58% rate of concordance of species between placenta and amniotic fluid cultures (69% for amnion-chorion biopsy, 58% for placenta biopsy, and 48% for placenta swab), with no significant differences seen among three culture types (χ2 0.891, P=.641). Of the positive amniotic fluid culture cases, only 50% (13 of 26) had all three placenta cultures positive with concordant species, and of these, 69% (9 of 13) agreed with the amniotic fluid organism isolate.
We performed multivariate analysis with placenta biopsy culture as the dependent variable in a stepwise logistic model. In the positive amniotic fluid culture and negative amniotic culture groups, steroid use was significantly associated with a positive culture independent of gestational age, amniotic fluid culture, mode of delivery, and antibiotic use (overall model fit P=.04, steroid use odds ratio 4.8, 95% CI 1.1–20.3). When multivariate analysis was applied to the third trimester control group with placenta biopsy culture as the dependent variable, mode of delivery was significantly inversely associated with a positive culture, independent of gestational age, antibiotic use, and maternal age (overall model fit P=.02, cesarean delivery odds ratio 0.08, 95% CI 0.01–0.91).
Overall, we determined that patients with intraamniotic inflammation (as assessed by a Mass Restricted score 3–4) had a higher incidence of acute histologic chorioamnionitis (25 of 26) (96%) than both women without amniotic fluid inflammation (24 of 30) (80%) and third trimester controls (7 of 23) (30.4%) (χ2 27.5, P<.001) (Fig. 4). Yet, our data confirmed that the frequency of acute histologic chorioamnionitis was significantly higher in preterm labor women without evidence of intraamniotic inflammation compared with our third trimester controls (χ2 11.2, P<.001).
The performance of placental histology to predict intraamniotic inflammation was further analyzed. Acute histologic chorioamnionitis showed sensitivity of 100% (95% CI 83–100%) and negative predictive value of 100% (77–100%), but specificity of only 36% (95% CI 23–52%), positive predictive value of 45% (95% CI 32–59%) and overall accuracy of 58% (95% CI 44–71%) in diagnosing intraamniotic inflammation. This suggests that histologic evaluation of the placenta is not very accurate in predicting intraamniotic inflammation in women who deliver prematurely.
Our research was motivated by the fact that when intraamniotic infection or inflammation cannot be assessed by direct analysis of the amniotic fluid, clinicians believe that placental evaluation with histologic analysis and microbial cultures can provide guidance both for a retrospective diagnosis and in future counseling. This common practice has not yet been validated. We show that microbiologic studies of the placenta correlate poorly and show poor accuracy to diagnose amniotic fluid cultures and that the presence of histologic chorioamnionitis is a sensitive, but not specific test to diagnose intraamniotic inflammation. Most diagnostic tests involve a careful balance of sensitivity and specificity and often sacrifice one to optimize another. For tests such as those used in this study, it is more important not to mislabel a nondiseased patient as diseased, and falsely believe that infection played a role in the pregnancy outcome. Thus, in addition to demonstrating a high level of accuracy, specificity should be optimized at the expense of high sensitivity. In fact, the opposite is true and the specificity for placental cultures as reported in our study perform poorly, with confidence intervals that are all well below 80%.
Overall there seems to be a high enough level of colonization or contamination of the placenta to complicate adequate evaluation by postpartum bacteriologic studies. Moreover, a high level of acute inflammation of the placenta on histology is also noted in patients with negative amniotic fluid cultures and in patients who do not show inflammation according to proteomics analysis and other accepted markers of inflammation. Therefore, our hypothesis that placental cultures after delivery are a valuable tool in evaluating or even speculating on intra-amniotic infection and inflammation was rejected. In the multivariate analysis, antenatal corticosteroid use was the only factor that significantly associated with a positive placenta culture result. These findings should come as no surprise, because this is consistent with existing practice patterns of giving steroids to the patients at highest risk for delivery. We do not believe that steroids affected our results, because meta-analysis of the literature on antenatal corticosteroids shows that they do not increase the risk of chorioamnionitis and in fact decrease the risk of early neonatal systemic infections.31
Previous studies have shown that inflammation of the chorionic plate is a sensitive (96.6%) indicator of positive amniotic fluid culture, with marginal specificity (65.9%), and that funisitis has a higher specificity (85.7%) at the expense of sensitivity (74.2%).14 In this study there does not seem to be a consistent pattern of difference in histologic inflammation between positive amniotic fluid culture and negative amniotic culture patients to serve as a guide to interpretation of pathology examination in predicting amniotic fluid culture results. A strength of this study is that we used a direct comparison of placental cultures and infection of the amniotic cavity, as well as an analysis of the relationship between inflammatory markers of the placenta and inflammation of the amniotic fluid using proteomics technology, which is proposed as the most accurate test to date in predicting inflammation of the amniotic fluid (Buhimschi CS, Bhandari V, Hamar BD, Bahtiyar MO, Zhao G, Sfakianaki A, et al. Proteomic profiling of the amniotic fluid to detect inflammation, infection and neonatal sepsis. PLoS Med 2007; in press). Furthermore, we used a control population of patients not symptomatic for preterm labor or preterm PROM. For these reasons we believe this study and its findings are clinically relevant and add a comprehensive and rigorous analysis to the complicated issue of evaluating infection and inflammation in pregnancy. One half of our third trimester controls had elective cesarean delivery and had no evidence of labor or rupture of membranes, suggesting that this group was at low risk for infection or vaginal contamination. There was no pattern of culture isolates or histology within this subgroup and no difference in the rate of positive cultures or placenta inflammation when compared with those groups at risk for infection (positive amniotic fluid culture and negative amniotic culture groups) or the other subgroup of controls.
The source of the bacteria in the placentas of our negative amniotic culture and third trimester control patients could come from colonization or contamination. While colonization would indicate the presence of bacteria without pathologic consequences, tissue contamination would come from factors that followed the amniocentesis, either during labor and delivery or in the acquisition of the cultures themselves. The high level of positive cultures in our third trimester control group, particularly in the patients who had elective cesarean deliveries (in the absence of labor or ruptured membranes) allows us to speculate that there may be a low level of bacterial colonization of the membranes or placenta and that changes in the maternal, fetal, or intraamniotic milieu may influence a conversion from colonization to pathologic infection. That cesarean delivery was inversely associated with—and thus protective against—a positive placental culture result in this group provides further evidence for this concept.
The exact route of bacteria travel and placental colonization is heterogeneous, and organisms may originate in the decidua, the cervix, the fallopian tubes, the intrauterine space, or hematogenously.32 As a result, many different techniques for placental cultures have been proposed, although there is no published evidence validating any of the culture techniques with comparison to amniotic fluid analysis. Although previous studies have used culture swabs of the fetal side of the placenta, swabs of the potential space between the amnion and chorion, and subchorionic fibrin cultures33 to diagnose infection on placental specimens, it is not inherently obvious which culture technique is superior. Moreover, some of the proposed methods, particularly the separation of the amnion and chorion, require technical skill and can be challenging in many clinical settings. This study evaluated three simple and reproducible types of culture assessment of the placenta and did not find differences in their sensitivities or specificities in predicting intraamniotic infection. Thus we cannot conclude that any placenta culture technique performs better than any other. In fact, all of the evaluated techniques performed equally poorly.
There are several limitations that we readily identify in this study. The ideal control group for this study would be viable, preterm patients undergoing amniocentesis, without clinical suspicion of infection, and having a high probability of delivering within a reasonable amount of time after amniocentesis. This group does not exist for obvious ethical issues and thus our control group represents the best possible alternative. Furthermore, our small sample size may have been unable to detect some differences in background characteristics that may have confounded the results.
In the end, we believe our results have important medicolegal consequences. The placenta is often used in the courtroom as evidence to demonstrate that a poor neonatal outcome may have been attributable to the clinical management or scenario of a patient.34,35 In fact, speculation of amniotic fluid infection or inflammation based purely on analysis of the placenta is not straightforward. Given the findings presented here, this approach may not be as valid as previously suggested.
2. Slattery M, Morrison J. Preterm delivery. Lancet 2002;360:1489–97.
3. Lockwood C, Kuczynski E. Risk stratification and pathological mechanisms in preterm delivery. Paediatr Perinat Epidemiol 2001;15 supp:78–89.
4. Buhimschi C, Weiner C, Buhimschi I. Proteomics, Part II: The emerging role of proteomics over genomics in spontaneous preterm labor/birth. Obstet Gynecol Surv 2006;61:543–53.
5. Wu Y, Colford J Jr. Chorioamnionitis as a risk factor for cerebral palsy: A meta-analysis. JAMA 2000;284:1417–24.
6. Newton E. Preterm labor, preterm premature rupture of membranes, and chorioamnionitis. Clin Perinatol 2005;32:571–600.
7. Stoll B, Hansen N, Adams-Chapman I, Fanaroff A, Hintz S, Vohr B, et al. Neurodevelopmental and growth impairment among extremely low-birth-weight infants with neonatal infection. JAMA 2004;19:2357–65.
8. Edwards R, Clark P, Locksmith G, Duff P. Performance characteristics of putative tests for subclinical chorioamnionitis. Infect Dis Obstet Gynecol 2001;9:209–24.
9. ACOG Committee on Practice Bulletins. American College of Obstetricians and Gynecologist. ACOG Practice Bulletin. Clinical management guidelines for obstetrician-gynecologist. Number 43, May 2003. Management of preterm labor. Obstet Gynecol 2003;101:1039–47.
10. Buhimschi C, Sfakianaki A, Hamar B, Pettker C, Bahtiyar M, Funai E, et al. A low vaginal “pool” amniotic fluid glucose measurement is a predictive but not a sensitive marker for infection in women with preterm premature rupture of membranes. Am J Obstet Gynecol 2006;194:309–16.
11. Shim S, Romero R, Jun J, Moon K, Kim G, Yoon B. C-reactive protein concentration in vaginal fluid as a marker for intra-amniotic inflammation/infection in preterm premature rupture of membranes. J Matern Fetal Neonat Med 2005;18:417–22.
12. Hillier S, Martius J, Krohn M, Kiviat N, Holmes K, Eschenbach D. A case-control study of chorioamnionic infection and histologic chorioamnionitis in prematurity. N Engl J Med 1988;319:972–8.
13. Dong Y, St. Clair P, Ramzy I, Kagan-Hallet K, Gibbs R. A microbiologic and clinical study of placental inflammation at term. Obstet Gynecol 1987;70:175–82.
14. Romero R, Salafia C, Athanassiadis A, Hanaoka S, Mazor M, Sepulveda W, et al. The relationship between acute inflammatory lesions of the preterm placenta and amniotic fluid microbiology. Am J Obstet Gynecol 1992;166:1382–8.
15. Park J, Romero R, Yoon B, Moon J, Oh S, Han S, et al. The relationship between amniotic fluid matrix metalloproteinase-8 and funisitis. Am J Obstet Gynecol 2001;185:1156–61.
16. Odibo AO, Rodis JF, Sanders MM, Borgida AF, Wilson M, Egan JF, et al. Relationship of amniotic fluid markers of intra-amniotic infection with histopathology in cases of preterm labor with intact membranes. J Perinatol 1999;19:407–12.
17. Ghidini A, Salafia CM, Kirn V, Doria V, Spong CY. Biophysical profile in predicting acute ascending infection in preterm rupture of membranes before 32 weeks. Obstet Gynecol 2000;96:201–6.
18. Kuhnert P, Erhard P, Kuhnert B, Sokol R, Gross T. A modified lecithin/sphingomyelin ratio test for fetal lung maturity. Am J Obstet Gynecol 1979;135:331–6.
19. Lewis D, Bergstedt S, Edwards M, Burlison S, Gallaspy J, Brooks G, et al. Successful magnesium sulfate tocolysis: is “weaning” the drug necessary? Am J Obstet Gynecol 1997;177:742–5.
20. Harirah H, Donia S, Hsu C. Amniotic fluid matrix metalloproteinase-9 and interleukin-6 in predicting intra-amniotic infection. Obstet Gynecol 2002;99:80–4.
21. Buhimschi I, Christner R, Buhimschi C. Proteomic biomarker analysis of amniotic for identification of intra-amniotic inflammation. BJOG 2005;112:173–81.
22. Buhimschi I, Buhimschi C, Christner R, Weiner C. Proteomics technology for the accurate diagnosis of inflammation in twin pregnancies. BJOG 2005;112:250–5.
23. Naeye R Disorders of the placenta and decidua. In: Disorders of the Placenta, Fetus, and Neonate: Diagnosis and Clinical Significance. St. Louis (MO): Mosby; 1992. p. 118–247.
24. Salafia C, Weigl C, Silberman L. The prevalence and distribution of acute placental inflammation in uncomplicated term pregnancies. Obstet Gynecol 1989;73:383–9.
25. Garry D, Figueroa R, Aguero-Rosenfield M, Martinez E, Visintainer P, Tejani N. A comparison of rapid amniotic fluid markers in the prediction of microbial invasion of the uterine cavity and preterm delivery. Am J Obstet Gynecol 1996;175:1336–41.
26. Landis J, Koch G. The measurement of observer agreement for categorical data. Biometrics 1977;33:159–74.
27. Locksmith GJ, Clark P, Duff P, Schultz GS. Amniotic fluid matrix metalloproteinase-9 levels in women with preterm labor and suspected intra-amniotic infection. Obstet Gynecol 1999;94:1–6.
28. Eng J. Sample size estimation: how many individuals should be studied? Radiology 2003;227:309–13.
29. Sperling R, Newton E, Gibbs R. Intraamniotic infection in low-birth-weight infants. J Infect Dis 1988;157:113–7.
30. Pankuch G, Appelbaum P, Lorenz R, Botti J, Schachter J, Naeye R. Placental microbiology and histology and the pathogenesis of chorioamnionitis Obstet Gynecol 1984;64:802–6.
31. Roberts D, Dalziel S, Antenatal corticosteroids for accelerating fetal lung maturation for women at risk of preterm birth. Cochrane Database Syst Rev 2006;3:CD004454.
32. Redline R. Placental inflammation. Semin Neonatol 2004;9:265–74.
33. Aquino T, Zhang J, Kraus F, Knefel R, Taff T. Subchorionic fibrin cultures for bacteriologic study of the placenta. Am J Clin Pathol 1984;81:482–6.
34. Altshuler G. Placenta within the medicolegal imperative. Arch Pathol Lab Med 1991;115:688–95.
35. Redline R, O'Riordan M. Placental lesions associated with cerebral palsy and neurologic impairment following term birth. Arch Pathol Lab Med 2000;124:1785–91.