- Email: [email protected]
The Alabama Preterm Birth study: polymorphonuclear and mononuclear cell placental infiltrations, other markers of inflammation, and outcomes in 23- to 32-week preterm newborn infants
American Journal of Obstetrics and Gynecology (2006) 195, 803–8
www.ajog.org
The Alabama Preterm Birth study: polymorphonuclear and mononuclear cell placental infiltrations, other markers of inflammation, and outcomes in 23- to 32-week preterm newborn infants William W. Andrews, MD,a Robert L. Goldenberg, MD,a Ona Faye-Petersen, MD,b Suzanne Cliver, BA,a Alice R. Goepfert, MD,a John C. Hauth, MDa Departments of Obstetrics and Gynecologya and Pathology,b University of Alabama at Birmingham, Birmingham, Alabama Received for publication March 8, 2006; revised June 15, 2006; accepted June 29, 2006
KEY WORDS Chorioamnionitis Interleukin-6 Neonatal outcome Preterm birth
Objective: The purpose of this study was to better understand the relationship between placental polymorphonuclear and mononuclear cell infiltrations with bacterial cultures, markers of inflammation, and preterm outcomes. Study design: This was a prospective study in 446 women who were delivered of a singleton infant at !32 weeks of gestational age. Five placental sites were categorized as having polymorphonuclear or mononuclear infiltrations. Results were compared with placental and cord cultures, umbilical cord interleukin-6 levels, and neonatal outcomes. Results: Polymorphonuclear, but not mononuclear, cell infiltrations were more common at the earliest gestational ages and in black women (56.0% vs 39.3%; P ! .01). Polymorphonuclear infiltration was associated with spontaneous preterm birth (73.9% vs 8.0%; P ! .0001), but not with preeclampsia (9.9% vs 34%; P ! .0001). Women with positive cultures, high interleukin6 levels, and clinical chorioamnionitis all had significantly more polymorphonuclear infiltrations than did women without those conditions (all probability values, !.0001). In all sites, polymorphonuclear infiltration was associated with neonatal systemic inflammatory response syndrome (P ! .0001) and in the cord with necrotizing enterocolitis (22.4% vs 13.5%; P = .02). Intraventricular hemorrhage and neonatal death were not associated with polymorphonuclear infiltration. Polymorphonuclear infiltration at all sites was associated with less respiratory distress syndrome (P ! .01). Mononuclear cell infiltration, when present in the decidua basalis, was associated with an increase in neonatal intraventricular hemorrhage (23.8% vs 7.4%; P ! .0004). Plasmacytic infiltrates were associated with increased intraventricular hemorrhage (29.4% vs 8.3%; P = .01) and neonatal death (27.8% vs 9.2%; P = .02).
Presented at the 26th Annual Meeting of the Society for Maternal Fetal Medicine, January 30-February 4, 2006, Miami, Florida. Reprint not available from the authors. Address correspondence to Robert L. Goldenberg, MD, Department of Obstetrics/Gynecology, 1500 6th Ave South, CRWH 379, Birmingham, AL 35233-1602. E-mail: [email protected] 0002-9378/$ - see front matter Ó 2006 Mosby, Inc. All rights reserved. doi:10.1016/j.ajog.2006.06.083
804
Andrews et al Conclusion: Polymorphonuclear infiltrations of the free membranes, chorionic plate, and umbilical cord were associated with positive intrauterine cultures and elevated cord blood interleukin-6. There was also an association with systemic inflammatory response syndrome and necrotizing enterocolitis, but not with intraventricular hemorrhage or death, and with decreased respiratory distress syndrome. Decidual mononuclear cell infiltration was associated with an increased risk of intraventricular hemorrhage and decidual plasma cell infiltration with increased intraventricular hemorrhage and neonatal death. Ó 2006 Mosby, Inc. All rights reserved.
Placental histologic condition has proved valuable in the determination of the cause of some preterm births and the prediction of some short- and long-term outcomes, especially in preterm infants.1-3 Nevertheless, substantial questions remain about the relationship between many maternal characteristics, markers of infection/ inflammation, newborn infant outcomes, and various placental histologic findings. Histologic chorioamnionitis (HCA) is defined generally as polymorphonuclear cell (PMN) infiltration into the free membranes and chorionic plate and may be accompanied by PMN infiltration into the umbilical cord.3,4 HCA often is considered to be a surrogate measure for intra-amniotic bacterial infection; chorionic plate vasculitis and funisitis are considered markers of fetal inflammatory response.5,6 Both conditions are more common in preterm birth, especially early preterm birth.1,7 HCA and especially funisitis have been associated with a number of adverse outcomes that include intraventricular hemorrhage (IVH), periventricular leukomalacia, and necrotizing enterocolitis (NEC),2,8 but in some studies were found to be protective against respiratory distress syndrome (RDS).9,10 Although a great deal has been written about chorioamnionic PMN infiltration, fewer studies have evaluated PMN infiltration into the umbilical cord, the chorionic plate, decidua basalis, or villi or mononuclear cell infiltration into any site.11,12 Thus, we explored the relationship between both PMN and mononuclear cell infiltration at various placental sites with placental bacterial cultures and other markers of inflammation and various newborn infant outcomes in very early gestational age pregnancies.
Methods Four hundred fifty-seven consecutive singleton infants who had been born between 23 and 32 weeks of gestation were evaluated. A chart review was performed to gather demographic and obstetric characteristics. Spontaneous preterm birth was defined as delivery after either spontaneous preterm labor or spontaneous preterm premature rupture of membranes. Indicated preterm birth was defined as delivery effected for maternal or fetal indications. Clinical chorioamnionitis was
considered to be present if it was diagnosed by the faculty obstetrician and generally was based on the presence of fever, abdominal pain, and an elevated white count. Amniotic fluid infection was diagnosed by a positive amniotic fluid culture. Neonatal outcome data through hospital discharge or death were also recorded. Neonatal systemic inflammatory response syndrome (SIRS) was defined as the presence of negative cerebrospinal fluid and blood cultures plus clinically suspected sepsis or a band (band C PMN cell ratio O0.15). The diagnoses of grade 3 or 4 IVH or cystic periventricular leukomalacia were made with ultrasound criteria.13 NEC stage 2 or greater was considered present if diagnosed clinically by the neonatologist. RDS was defined as the documentation of any 3 of the following criteria: (1) infant oxygen requirement at 6 hours through 24 hours of life, (2) an abnormal chest radiograph that was consistent with RDS within the first 24 hours of life, and (3) the need for surfactant. Bronchopulmonary dysplasia was defined as oxygen requirement at 28 days, and chronic lung disease was defined as oxygen requirement at 36 weeks of life. Placentas for 446 of the neonates (98%) were available. In each case, a minimum of 2 membrane rolls, 2 sections of umbilical cord (1 each from the placental and fetal ends), and 2 to 4 transmural parenchymal sections were submitted for histologic evaluation. All cases were evaluated by 1 pathologist (O.F.-P.), who was blinded to clinical outcome and who used a protocol adapted from Bendon et al.14 Specifically, the free membranes, chorionic plate, umbilical cord, decidua basalis, and chorionic villi were evaluated for the presence of PMN and mononuclear cell infiltrations. Mononuclear cell infiltrates were further subdivided into lymphohistiocytic and lymphoplasmacytic subtypes.12 Placental and umbilical cord cultures were obtained as described previously15,16 and were available for 445 and 343 women, respectively. Samples of cord blood were also collected for interleukin (IL)-6 determination and were available for 309 women. The specimens were centrifuged, aliquoted, and stored at 70 C. IL-6 concentrations in the cord plasma were determined by enzyme-linked immunosorbent assay kits (R&D Systems, Inc, Minneapolis, MN). The lower limit of sensitivity of the assay was 0.7 pg/mL. The intra-assay and interassay coefficients of variation were 2.6% and
Andrews et al 4.5%, respectively. Values of R34.5 pg/mL, which was the cutoff between the 3rd and 4th quartiles of indicated preterm births in our population, were considered elevated. Data analyses were performed with SAS software (version 9.0; SAS Institute, Inc, Cary, NC). Frequencies and means between groups were compared with the chisquare tests and Fisher’s exact tests for discrete variables and t-tests and analysis of variance for continuous variables. Adjusted odds ratios (OR) for the determination of the association between PMN and mononuclear cell infiltration and specific morbidities that were adjusted for other factors were calculated by means of logistic regression analysis. A probability value of %.05 was chosen to define statistical significance. The study was approved by the University of Alabama at Birmingham Institutional Review Board.
Results This study evaluated 446 mother/infant dyads with a singleton delivery at %32 weeks of gestation and placental histologic information available. The mean birthweight was 1140 G 390 g, and the mean gestational age was 28.6 G 2.2 weeks. Table I shows the relationship between several maternal characteristics and PMN or mononuclear cell infiltration at any of 5 placental sites. PMN infiltration was more common at early gestational age and in black women and was significantly less common in women with preeclampsia but was not related to maternal age, smoking, marital status, maternal education, parity, or diabetes mellitus (data not shown.) Mononuclear cell infiltration was not related significantly to any of these characteristics, except marginally, with indicated preterm birth. The distribution of PMN and mononuclear cell infiltrations into various placental sites was evaluated. PMN infiltration was common in the free membranes (48.3%), chorionic plate (39.9%), and umbilical cord (30.5%) but was far less common (!1.0%) in the decidua basalis or chorionic villi. Mononuclear cell infiltration, which included lymphoplasmacytic and lymphohistiocytic subtypes, was present in approximately 10% of the placental free membranes and decidua basalis but was far less common (!1.0%) in the chorionic plate, umbilical cord, and villi. Thus, further analyses that were related to PMN infiltrations were restricted to the free membranes, chorionic plate, and umbilical cord and that were related to mononuclear cell infiltration were restricted to the free membranes and decidua basalis. Only 18 women (3.6%) had plasma cells in any location (in the decidua basalis, 15 women; in the free membranes, 3 women). Table II shows the relationship of PMN infiltration at each site to other markers of infection/inflammation. As can be seen, regardless of which marker is
805 considered, women with these conditions were significantly more likely to have PMN infiltration than women who did not have these conditions. For example, after adjustment for demographic characteristics (gestational age, race, sex), the data showed that women with an IL-6 level of R34.5 pg/mL had an OR of 15.6 (95% CI, 7.8-31.2) for membrane PMN infiltration. Despite this strong relationship, even when these markers were present, many placentas had no PMN infiltrations and, conversely, when the marker was absent, some women still had PMN infiltrations. To quantitate this relationship, we calculated the sensitivity and specificity of 3 of the markers of inflammation to predict PMN infiltrations into the membranes. For an IL-6 level of R34.5 pg/mL, the sensitivity and specificity were 54.0% and 92.0%, respectively; for any positive culture, the sensitivity and specificity were 78.9% and 62.2%, respectively; and for clinical chorioamnionitis, the sensitivity and specificity were 23.0% and 96.8%, respectively. There was therefore a relatively strong, but far from perfect, correlation between PMN infiltration and the other inflammatory markers. Table III shows a similar analysis for mononuclear cell infiltration into the free membranes and descidua basalis. When compared with women without the markers, women with the various markers of inflammation were less likely to have mononuclear infiltration into the free membranes, although the presence of the markers did not influence mononuclear cell infiltration into the decidua basalis. The presence of plasma cells also was not associated significantly with any of the markers (data not shown). Table IV shows the relationship between PMN infiltration in the various sites and neonatal outcome. PMN infiltration at all 3 sites was associated significantly with an increase in neonatal SIRS (for the membranes: OR, 3.5; 95% CI, 2.2-5.5) and in the umbilical cord with NEC (OR, 1.9; 95% CI, 1.1-3.3), both adjusted for gestational age, race, and sex. PMN infiltration at all 3 sites appears to be protective against RDS (for the membranes: OR, 0.4; 95% CI, 0.3-0.7). However, at no site did PMN infiltration have a significant relationship with IVH, neonatal death, sepsis, periventricular leukomalacia, or chronic lung disease. Mononuclear cell infiltrations generally had little relationship to pregnancy outcome. However, mononuclear cell infiltrations into the decidua basalis were associated with a significant increase in grade 3/4 IVH (23.8% vs 7.4%; P = .0004), even after an adjustment for gestational age, race, and infant sex (OR, 4.3; 95% CI, 1.6-11.4). Because the relationship between mononuclear cell infiltration into the decidua basalis and IVH was so strong, we compared the relationship between plasmacytic infiltrations into the decidua and membranes combined with various neonatal outcomes (n = 18 patients). Both IVH (29.4% vs 8.3%; P = .01) and
806 Table I
Andrews et al The association between maternal characteristics and PMN and mononuclear cell infiltration at any placental site
Pregnancy characteristic Gestational age (wk) 23-24 25-28 29-32 Race Black White Hispanic/other Type of preterm birth Spontaneous Indicated Preeclampsia Yes No
Mononuclear infiltrationy
P value
.01
18.2 20.0 13.0
NS
56.0 39.3 76.9
!.01
18.4 13.7 7.7
NS
64.1 35.9
73.9 8.0
!.0001
13.4 20.9
.04
34.1 65.9
9.9 70.4
!.0001
21.1 13.6
NS
With characteristic (%)
PMN infiltration (%)*
7.4 42.6 50.0
72.7 51.0 45.3
56.1 41.0 2.9
P value
NS, Not significant. * N = 222/446 patients. y N = 72/446 patients.
Table II The percentage of women who have PMN cell infiltrations at various placental sites when other indicators of infection/ inflammation are present or absent Indicator present PMN in membranes
PMN in chorionic plate
PMN in umbilical cord
Yes (%)
No (%)
P value
Yes (%)
No (%)
P value
29.7 !.0001 32.2 !.0001
84.9 68.5
21.6 25.1
!.0001 !.0001
73.0 48.7
13.9 21.0
!.0001 !.0001
76.9
37.7 !.0001
74.4
30.4
!.0001
51.3
25.1
!.0001
421
77.9
30.2 !.0001
66.7
23.1
!.0001
48.2
19.3
!.0001
443 177 446
66.1 66.4 87.3
24.2 !.0001 12.1 !.0001 42.6 !.0001
55.4 48.6 87.3
18.5 6.8 33.1
!.0001 !.0001 !.0001
40.3 33.3 76.4
16.9 6.9 24.0
!.0001 .0001 !.0001
Infection/inflammation indicator
Indicators Yes tested (n) (%)
No (%)
IL-6 R34.5 pg/mL Ureaplasma urealyticum/ Mycoplasma hominis C Cord blood C for U urealyticum or M hominis U urealyticum/M hominis C anywhere Any placenta culture C Amniotic fluid culture C Clinical chorioamnionitis present
309 421
85.1 80.6
343
P value
C, Culture positive.
neonatal death (27.8% vs 9.2%; P = .02) were associated significantly with lymphoplasmacytic infiltrations. After an adjustment for gestational age, race, and infant sex, the ORs for both relationships remained significant (IVH: OR, 8.9; 95% CI, 2.1-37.9; infant death: OR, 3.7; 95% CI, 1.0-12.7).
Comment In a large population of early preterm deliveries, we examined the relationship between placental histologic condition, other indicators of intrauterine infection/ inflammation, and newborn infant outcomes. The results confirm that, within a population of early gestational age births, deliveries that occur at the earliest gestational ages
were more likely to have HCA.1,7 Black women were also more likely to have HCA, as previously reported.17 The extent to which HCA is a marker for bacterial infection has been questioned.5,18 In this study, PMN infiltrations into the free membranes, chorionic plate, and umbilical cord were related strongly to positive placental, umbilical cord, and amniotic fluid cultures, which confirmed that HCA is a relatively strong surrogate for placental bacterial infection. However, because some women with negative cultures had HCA and because some women with positive cultures did not have HCA, these data suggest that explanations for HCA, other than bacterial infection, may exist. It is also very possible that, as in other areas of the body, unrecognized or very fastidious bacteria are responsible for the HCA.19,20
Andrews et al
807
Table III The percentage of women who have mononuclear cell infiltrations at various placental sites when other indicators of infection/inflammation are present or absent (n = 446 women) Indicator present Infection/inflammation indicator IL-6 R34.5 pg/mL Ureaplasma urealyticum/M hominis C Cord blood C for U urealyticum or M hominis U urealyticum/M hominis Canywhere Any placenta culture C Amniotic fluid culture C Clinical chorioamnionitis present
Mononuclear cell infiltration in membranes
Mononuclear cell infiltration in decidua (%)
Tested (n)
Yes (%)
No (%)
P value
Yes (%)
No (%)
P value
309 421 343 421 443 177 446
0.0 6.3 5.1 6.8 7.3 8.6 1.8
12.7 11.4 10.3 11.8 12.9 15.3 11.0
.001 NS NS NS .05 NS .03
6.7 9.4 11.3 11.3 9.5 15.6 5.4
12.2 10.4 9.4 9.6 10.2 13.8 10.9
NS NS NS NS NS NS NS
C, Culture positive.
Table IV The percentage of women who have various newborn infant outcomes, compared with whether they have or do not have PMN infiltration at various sites (n = 446 women) PMN in membranes
PMN in fetal plate
PMN in umbilical cord
Neonatal outcome
Yes (%)
No (%)
P value
Yes (%)
No (%)
P value
Yes (%)
No (%)
P value
RDS Chronic lung disease Bronchopulmonary dysplasia IVH 3/4 Periventricular leukomalacia SIRS Sepsis NEC Death*
60.8 6.7 17.7 9.9 4.5 43.9 1.4 18.2 10.1
72.8 8.5 17.0 8.4 2.3 18.3 0.9 13.8 9.3
.008 NS NS NS NS !.0001 NS NS NS
58.6 5.8 15.5 11.2 2.4 50.6 1.7 17.8 10.3
72.4 8.4 18.0 8.0 4.0 18.0 0.8 15.7 9.5
.003 NS NS NS NS !.0001 NS NS NS
57.5 6.8 14.9 10.7 3.8 55.0 1.5 22.4 10.5
72.0 7.9 18.4 8.6 3.5 20.1 1.0 13.5 9.8
.003 NS NS NS NS !.0001 NS .02 NS
NS, Not significant. * Newborn infant death before discharge from hospital.
Mononuclear cell infiltration into the free membranes was less prevalent in women who had various markers of intrauterine infection/inflammation compared with women who did not have these markers. However, similar results were not seen with mononuclear cell infiltrations into the decidua basalis. The reason for this difference is not clear but may be related to the observation that most intrauterine infections are caused by bacteria that are ascending from the vagina; therefore, the infection is adjacent to the free membranes.1 The subsequent acute inflammatory response is predominantly neutrophilic and may also obscure a mononuclear response.11 In our data, PMN infiltration into various placental sites is associated with an increase in neonatal SIRS and into the umbilical cord is associated with NEC. Because neonatal SIRS is defined as an infant having the appearance of sepsis, but with negative cultures, the association is not surprising.21 We surmise that many infants attain high levels of circulating cytokines through their exposure to cytokines in the amniotic fluid but
do not become infected, in part because in approximately one half of the cases of intrauterine infection, the bacteria are confined to the membranes and do not reach the amniotic fluid.1,15 Whether the infant’s inflammatory response without concomitant infection is also explained by fetal exposure to nonviable bacteria, bacterial fragments, or unrecognized bacteria is unknown.19,20 The cause of NEC is also not clear but at times is thought to be related to bacterial infection of the infant’s gastrointestinal tract.22 Infants with funisitis are more likely to have organisms in the cord blood. Therefore, this relationship also is biologically plausible. Finally, several authors have suggested that, by enhancing the maturation of type II pneumocytes, HCA and the associated inflammatory response might be protective against RDS.9,10 Our results confirm that HCA is associated with a reduction in RDS. In this study, mononuclear cell infiltrations into various placental sites generally had little relationship to infant outcomes. However, the relationship between mononuclear cell infiltration into the decidua basalis
808 and IVH was relatively strong. The added finding of both IVH and neonatal death being associated with plasma cell infiltrates provides further support that something more than chance is responsible for the observed relationship. Mononuclear cell infiltration usually indicates a chronic bacterial or viral infection or an immune response to noninfectious antigens.10,11 Mononuclear cell infiltration at the maternal-fetal interface might also suggest poor placental function rather than exposure to high inflammatory cytokine levels as the cause of IVH in these infants. In summary, this study shows a strong but not perfect relationship between HCA, placental cultures, and other markers of infection/inflammation, which indicates that most, but not all, HCA is caused by bacterial infection. Placental inflammation was associated strongly with an increase in SIRS, and funisitis was associated strongly with NEC. HCA was associated with less RDS. The association between mononuclear and especially plasmacytic infiltration of the decidua basalis with IVH and neonatal death was unexpected but, if confirmed in other studies, may help to provide information about possible mechanisms that lead to these adverse neonatal outcomes.
References 1. Goldenberg RL, Hauth JC, Andrews WW. Intrauterine infection and preterm delivery. N Engl J Med 2000;342:1500-7. 2. O’Shea TM, Klinepeter KL, Meis PJ, Dillard RG. Intrauterine infection and the risk of cerebral palsy in very-low birthweight infants. Paediatr Perinat Epidemiol 1998;12:72-83. 3. Redline RW. Placental inflammation. Semin Neonatol 2004;9: 265-74. 4. Zhang J, Kraus FT, Aquino TI. Chorioamnionitis: a comparative histologic, bacteriologic, and clinical study. Int J Gynecol Pathol 1985;4:1-10. 5. Smulian JC, Shen-Schwarz S, Vintzileos AM, Lake MF, Ananth CV. Clinical chorioamnionitis and the histology of placental inflammation. Obstet Gynecol 1999;94:1000-5. 6. Redline RW, Heller D, Keating S, Kingdom J. Placental diagnostic criteria and clinical correlation: a workshop report. Placenta 2005;26(suppl):S114-7.
Andrews et al 7. Russell P. Inflammatory lesions of the human placenta: I, clinical significance of acute chorioamnionitis. Am J Diagn Gynecol Obstet 1979;1:127-37. 8. Redline RW. Severe fetal placental vascular lesions in term infants with neurologic impairment. Am J Obstet Gynecol 2005;192:452-7. 9. Watterberg KL, Demers L, Scott SM, Murphy S. Chorioamnionitis and early lung inflammation in infants in whom bronchopulmonary dysplasia develops. Pediatrics 1996;97:210-5. 10. Jobe AH, Ikegami M. Mechanisms initiating lung injury in the preterm. Early Hum Dev 1998;53:81-94. 11. Redline RW, Faye-Petersen O, Heller D, Qureshi F, Savell V, Vogler C, et al. Amniotic infection syndrome: nosiology and reproducibility of placental reaction patterns. Pediat Dev Pathol 2003;6:435-8. 12. Khong TY, Bendon RW, Qureshi F, Redline RW, Gould S, Stallmach T, et al. Chronic deciduitis in the placental basal plate: definition and interobserver reliability. Hum Pathol 2000;31: 292-5. 13. Volpe JJ. Hypoxic-ischemic encephalopathy: clinical aspects. In: Volpe JJ, editor. Neurology of the newborn. Philadelphia: Saunders; 2001. p. 331-94. 14. Bendon RW, Faye-Petersen O, Pavlova Z, Qureshi F, Elder N, Das A, et al. Histologic features of chorioamnion membrane rupture: development of methodology. Pediatr Pathol Lab Med 1997; 17:27-42. 15. Andrews WW, Hauth JC, Goldenberg RL, Gomez R, Romero R, Cassell GH. Amniotic fluid interleukin-6: correlation with upper genital tract microbial colonization and gestational age in women delivered after spontaneous labor versus indicated delivery. Am J Obstet Gynecol 1995;173:606-12. 16. Waites KB, Rikihisa Y, Taylor-Robinson D. Mycoplasma and ureaplasma. In: Murray PR, Baron EJ, Pfaller MA, Tenover FC, Yolken RH, editors. Manual of clinical microbiology. 8th ed. Washington, DC: ASM Press; 2003. p. 972-90. 17. Fiscella K. Racial disparities in preterm births: the role of urogenital infections. Public Health Rep 1996;111:104-13. 18. Pankuch GA, Applebaum PC, Lorenz RP, Botti JJ, Schachter J, Naeye RL. Placental microbiology and histology and the pathogenesis of chorioamnionitis. Obstet Gynecol 1984;64:802-6. 19. Relman DA. The search for unrecognized pathogens. Science 1999; 284:1308-10. 20. Fredricks DN, Fiedler TL, Marrazzo JM. Molecular identification of bacteria associated with bacterial vaginosis. N Engl J Med 2005; 353:1899-911. 21. Gomez R, Romero R, Ghezzi F, Yoon BH, Mazor M, Berry SM. The fetal inflammatory response syndrome. Am J Obstet Gynecol 1998;179:194-202. 22. Patole S. Prevention of necrotizing enterocolitis: year 2004 and beyond. J Matern Fetal Neonatal Med 2005;17:69-80.
www.ajog.org
The Alabama Preterm Birth study: polymorphonuclear and mononuclear cell placental infiltrations, other markers of inflammation, and outcomes in 23- to 32-week preterm newborn infants William W. Andrews, MD,a Robert L. Goldenberg, MD,a Ona Faye-Petersen, MD,b Suzanne Cliver, BA,a Alice R. Goepfert, MD,a John C. Hauth, MDa Departments of Obstetrics and Gynecologya and Pathology,b University of Alabama at Birmingham, Birmingham, Alabama Received for publication March 8, 2006; revised June 15, 2006; accepted June 29, 2006
KEY WORDS Chorioamnionitis Interleukin-6 Neonatal outcome Preterm birth
Objective: The purpose of this study was to better understand the relationship between placental polymorphonuclear and mononuclear cell infiltrations with bacterial cultures, markers of inflammation, and preterm outcomes. Study design: This was a prospective study in 446 women who were delivered of a singleton infant at !32 weeks of gestational age. Five placental sites were categorized as having polymorphonuclear or mononuclear infiltrations. Results were compared with placental and cord cultures, umbilical cord interleukin-6 levels, and neonatal outcomes. Results: Polymorphonuclear, but not mononuclear, cell infiltrations were more common at the earliest gestational ages and in black women (56.0% vs 39.3%; P ! .01). Polymorphonuclear infiltration was associated with spontaneous preterm birth (73.9% vs 8.0%; P ! .0001), but not with preeclampsia (9.9% vs 34%; P ! .0001). Women with positive cultures, high interleukin6 levels, and clinical chorioamnionitis all had significantly more polymorphonuclear infiltrations than did women without those conditions (all probability values, !.0001). In all sites, polymorphonuclear infiltration was associated with neonatal systemic inflammatory response syndrome (P ! .0001) and in the cord with necrotizing enterocolitis (22.4% vs 13.5%; P = .02). Intraventricular hemorrhage and neonatal death were not associated with polymorphonuclear infiltration. Polymorphonuclear infiltration at all sites was associated with less respiratory distress syndrome (P ! .01). Mononuclear cell infiltration, when present in the decidua basalis, was associated with an increase in neonatal intraventricular hemorrhage (23.8% vs 7.4%; P ! .0004). Plasmacytic infiltrates were associated with increased intraventricular hemorrhage (29.4% vs 8.3%; P = .01) and neonatal death (27.8% vs 9.2%; P = .02).
Presented at the 26th Annual Meeting of the Society for Maternal Fetal Medicine, January 30-February 4, 2006, Miami, Florida. Reprint not available from the authors. Address correspondence to Robert L. Goldenberg, MD, Department of Obstetrics/Gynecology, 1500 6th Ave South, CRWH 379, Birmingham, AL 35233-1602. E-mail: [email protected] 0002-9378/$ - see front matter Ó 2006 Mosby, Inc. All rights reserved. doi:10.1016/j.ajog.2006.06.083
804
Andrews et al Conclusion: Polymorphonuclear infiltrations of the free membranes, chorionic plate, and umbilical cord were associated with positive intrauterine cultures and elevated cord blood interleukin-6. There was also an association with systemic inflammatory response syndrome and necrotizing enterocolitis, but not with intraventricular hemorrhage or death, and with decreased respiratory distress syndrome. Decidual mononuclear cell infiltration was associated with an increased risk of intraventricular hemorrhage and decidual plasma cell infiltration with increased intraventricular hemorrhage and neonatal death. Ó 2006 Mosby, Inc. All rights reserved.
Placental histologic condition has proved valuable in the determination of the cause of some preterm births and the prediction of some short- and long-term outcomes, especially in preterm infants.1-3 Nevertheless, substantial questions remain about the relationship between many maternal characteristics, markers of infection/ inflammation, newborn infant outcomes, and various placental histologic findings. Histologic chorioamnionitis (HCA) is defined generally as polymorphonuclear cell (PMN) infiltration into the free membranes and chorionic plate and may be accompanied by PMN infiltration into the umbilical cord.3,4 HCA often is considered to be a surrogate measure for intra-amniotic bacterial infection; chorionic plate vasculitis and funisitis are considered markers of fetal inflammatory response.5,6 Both conditions are more common in preterm birth, especially early preterm birth.1,7 HCA and especially funisitis have been associated with a number of adverse outcomes that include intraventricular hemorrhage (IVH), periventricular leukomalacia, and necrotizing enterocolitis (NEC),2,8 but in some studies were found to be protective against respiratory distress syndrome (RDS).9,10 Although a great deal has been written about chorioamnionic PMN infiltration, fewer studies have evaluated PMN infiltration into the umbilical cord, the chorionic plate, decidua basalis, or villi or mononuclear cell infiltration into any site.11,12 Thus, we explored the relationship between both PMN and mononuclear cell infiltration at various placental sites with placental bacterial cultures and other markers of inflammation and various newborn infant outcomes in very early gestational age pregnancies.
Methods Four hundred fifty-seven consecutive singleton infants who had been born between 23 and 32 weeks of gestation were evaluated. A chart review was performed to gather demographic and obstetric characteristics. Spontaneous preterm birth was defined as delivery after either spontaneous preterm labor or spontaneous preterm premature rupture of membranes. Indicated preterm birth was defined as delivery effected for maternal or fetal indications. Clinical chorioamnionitis was
considered to be present if it was diagnosed by the faculty obstetrician and generally was based on the presence of fever, abdominal pain, and an elevated white count. Amniotic fluid infection was diagnosed by a positive amniotic fluid culture. Neonatal outcome data through hospital discharge or death were also recorded. Neonatal systemic inflammatory response syndrome (SIRS) was defined as the presence of negative cerebrospinal fluid and blood cultures plus clinically suspected sepsis or a band (band C PMN cell ratio O0.15). The diagnoses of grade 3 or 4 IVH or cystic periventricular leukomalacia were made with ultrasound criteria.13 NEC stage 2 or greater was considered present if diagnosed clinically by the neonatologist. RDS was defined as the documentation of any 3 of the following criteria: (1) infant oxygen requirement at 6 hours through 24 hours of life, (2) an abnormal chest radiograph that was consistent with RDS within the first 24 hours of life, and (3) the need for surfactant. Bronchopulmonary dysplasia was defined as oxygen requirement at 28 days, and chronic lung disease was defined as oxygen requirement at 36 weeks of life. Placentas for 446 of the neonates (98%) were available. In each case, a minimum of 2 membrane rolls, 2 sections of umbilical cord (1 each from the placental and fetal ends), and 2 to 4 transmural parenchymal sections were submitted for histologic evaluation. All cases were evaluated by 1 pathologist (O.F.-P.), who was blinded to clinical outcome and who used a protocol adapted from Bendon et al.14 Specifically, the free membranes, chorionic plate, umbilical cord, decidua basalis, and chorionic villi were evaluated for the presence of PMN and mononuclear cell infiltrations. Mononuclear cell infiltrates were further subdivided into lymphohistiocytic and lymphoplasmacytic subtypes.12 Placental and umbilical cord cultures were obtained as described previously15,16 and were available for 445 and 343 women, respectively. Samples of cord blood were also collected for interleukin (IL)-6 determination and were available for 309 women. The specimens were centrifuged, aliquoted, and stored at 70 C. IL-6 concentrations in the cord plasma were determined by enzyme-linked immunosorbent assay kits (R&D Systems, Inc, Minneapolis, MN). The lower limit of sensitivity of the assay was 0.7 pg/mL. The intra-assay and interassay coefficients of variation were 2.6% and
Andrews et al 4.5%, respectively. Values of R34.5 pg/mL, which was the cutoff between the 3rd and 4th quartiles of indicated preterm births in our population, were considered elevated. Data analyses were performed with SAS software (version 9.0; SAS Institute, Inc, Cary, NC). Frequencies and means between groups were compared with the chisquare tests and Fisher’s exact tests for discrete variables and t-tests and analysis of variance for continuous variables. Adjusted odds ratios (OR) for the determination of the association between PMN and mononuclear cell infiltration and specific morbidities that were adjusted for other factors were calculated by means of logistic regression analysis. A probability value of %.05 was chosen to define statistical significance. The study was approved by the University of Alabama at Birmingham Institutional Review Board.
Results This study evaluated 446 mother/infant dyads with a singleton delivery at %32 weeks of gestation and placental histologic information available. The mean birthweight was 1140 G 390 g, and the mean gestational age was 28.6 G 2.2 weeks. Table I shows the relationship between several maternal characteristics and PMN or mononuclear cell infiltration at any of 5 placental sites. PMN infiltration was more common at early gestational age and in black women and was significantly less common in women with preeclampsia but was not related to maternal age, smoking, marital status, maternal education, parity, or diabetes mellitus (data not shown.) Mononuclear cell infiltration was not related significantly to any of these characteristics, except marginally, with indicated preterm birth. The distribution of PMN and mononuclear cell infiltrations into various placental sites was evaluated. PMN infiltration was common in the free membranes (48.3%), chorionic plate (39.9%), and umbilical cord (30.5%) but was far less common (!1.0%) in the decidua basalis or chorionic villi. Mononuclear cell infiltration, which included lymphoplasmacytic and lymphohistiocytic subtypes, was present in approximately 10% of the placental free membranes and decidua basalis but was far less common (!1.0%) in the chorionic plate, umbilical cord, and villi. Thus, further analyses that were related to PMN infiltrations were restricted to the free membranes, chorionic plate, and umbilical cord and that were related to mononuclear cell infiltration were restricted to the free membranes and decidua basalis. Only 18 women (3.6%) had plasma cells in any location (in the decidua basalis, 15 women; in the free membranes, 3 women). Table II shows the relationship of PMN infiltration at each site to other markers of infection/inflammation. As can be seen, regardless of which marker is
805 considered, women with these conditions were significantly more likely to have PMN infiltration than women who did not have these conditions. For example, after adjustment for demographic characteristics (gestational age, race, sex), the data showed that women with an IL-6 level of R34.5 pg/mL had an OR of 15.6 (95% CI, 7.8-31.2) for membrane PMN infiltration. Despite this strong relationship, even when these markers were present, many placentas had no PMN infiltrations and, conversely, when the marker was absent, some women still had PMN infiltrations. To quantitate this relationship, we calculated the sensitivity and specificity of 3 of the markers of inflammation to predict PMN infiltrations into the membranes. For an IL-6 level of R34.5 pg/mL, the sensitivity and specificity were 54.0% and 92.0%, respectively; for any positive culture, the sensitivity and specificity were 78.9% and 62.2%, respectively; and for clinical chorioamnionitis, the sensitivity and specificity were 23.0% and 96.8%, respectively. There was therefore a relatively strong, but far from perfect, correlation between PMN infiltration and the other inflammatory markers. Table III shows a similar analysis for mononuclear cell infiltration into the free membranes and descidua basalis. When compared with women without the markers, women with the various markers of inflammation were less likely to have mononuclear infiltration into the free membranes, although the presence of the markers did not influence mononuclear cell infiltration into the decidua basalis. The presence of plasma cells also was not associated significantly with any of the markers (data not shown). Table IV shows the relationship between PMN infiltration in the various sites and neonatal outcome. PMN infiltration at all 3 sites was associated significantly with an increase in neonatal SIRS (for the membranes: OR, 3.5; 95% CI, 2.2-5.5) and in the umbilical cord with NEC (OR, 1.9; 95% CI, 1.1-3.3), both adjusted for gestational age, race, and sex. PMN infiltration at all 3 sites appears to be protective against RDS (for the membranes: OR, 0.4; 95% CI, 0.3-0.7). However, at no site did PMN infiltration have a significant relationship with IVH, neonatal death, sepsis, periventricular leukomalacia, or chronic lung disease. Mononuclear cell infiltrations generally had little relationship to pregnancy outcome. However, mononuclear cell infiltrations into the decidua basalis were associated with a significant increase in grade 3/4 IVH (23.8% vs 7.4%; P = .0004), even after an adjustment for gestational age, race, and infant sex (OR, 4.3; 95% CI, 1.6-11.4). Because the relationship between mononuclear cell infiltration into the decidua basalis and IVH was so strong, we compared the relationship between plasmacytic infiltrations into the decidua and membranes combined with various neonatal outcomes (n = 18 patients). Both IVH (29.4% vs 8.3%; P = .01) and
806 Table I
Andrews et al The association between maternal characteristics and PMN and mononuclear cell infiltration at any placental site
Pregnancy characteristic Gestational age (wk) 23-24 25-28 29-32 Race Black White Hispanic/other Type of preterm birth Spontaneous Indicated Preeclampsia Yes No
Mononuclear infiltrationy
P value
.01
18.2 20.0 13.0
NS
56.0 39.3 76.9
!.01
18.4 13.7 7.7
NS
64.1 35.9
73.9 8.0
!.0001
13.4 20.9
.04
34.1 65.9
9.9 70.4
!.0001
21.1 13.6
NS
With characteristic (%)
PMN infiltration (%)*
7.4 42.6 50.0
72.7 51.0 45.3
56.1 41.0 2.9
P value
NS, Not significant. * N = 222/446 patients. y N = 72/446 patients.
Table II The percentage of women who have PMN cell infiltrations at various placental sites when other indicators of infection/ inflammation are present or absent Indicator present PMN in membranes
PMN in chorionic plate
PMN in umbilical cord
Yes (%)
No (%)
P value
Yes (%)
No (%)
P value
29.7 !.0001 32.2 !.0001
84.9 68.5
21.6 25.1
!.0001 !.0001
73.0 48.7
13.9 21.0
!.0001 !.0001
76.9
37.7 !.0001
74.4
30.4
!.0001
51.3
25.1
!.0001
421
77.9
30.2 !.0001
66.7
23.1
!.0001
48.2
19.3
!.0001
443 177 446
66.1 66.4 87.3
24.2 !.0001 12.1 !.0001 42.6 !.0001
55.4 48.6 87.3
18.5 6.8 33.1
!.0001 !.0001 !.0001
40.3 33.3 76.4
16.9 6.9 24.0
!.0001 .0001 !.0001
Infection/inflammation indicator
Indicators Yes tested (n) (%)
No (%)
IL-6 R34.5 pg/mL Ureaplasma urealyticum/ Mycoplasma hominis C Cord blood C for U urealyticum or M hominis U urealyticum/M hominis C anywhere Any placenta culture C Amniotic fluid culture C Clinical chorioamnionitis present
309 421
85.1 80.6
343
P value
C, Culture positive.
neonatal death (27.8% vs 9.2%; P = .02) were associated significantly with lymphoplasmacytic infiltrations. After an adjustment for gestational age, race, and infant sex, the ORs for both relationships remained significant (IVH: OR, 8.9; 95% CI, 2.1-37.9; infant death: OR, 3.7; 95% CI, 1.0-12.7).
Comment In a large population of early preterm deliveries, we examined the relationship between placental histologic condition, other indicators of intrauterine infection/ inflammation, and newborn infant outcomes. The results confirm that, within a population of early gestational age births, deliveries that occur at the earliest gestational ages
were more likely to have HCA.1,7 Black women were also more likely to have HCA, as previously reported.17 The extent to which HCA is a marker for bacterial infection has been questioned.5,18 In this study, PMN infiltrations into the free membranes, chorionic plate, and umbilical cord were related strongly to positive placental, umbilical cord, and amniotic fluid cultures, which confirmed that HCA is a relatively strong surrogate for placental bacterial infection. However, because some women with negative cultures had HCA and because some women with positive cultures did not have HCA, these data suggest that explanations for HCA, other than bacterial infection, may exist. It is also very possible that, as in other areas of the body, unrecognized or very fastidious bacteria are responsible for the HCA.19,20
Andrews et al
807
Table III The percentage of women who have mononuclear cell infiltrations at various placental sites when other indicators of infection/inflammation are present or absent (n = 446 women) Indicator present Infection/inflammation indicator IL-6 R34.5 pg/mL Ureaplasma urealyticum/M hominis C Cord blood C for U urealyticum or M hominis U urealyticum/M hominis Canywhere Any placenta culture C Amniotic fluid culture C Clinical chorioamnionitis present
Mononuclear cell infiltration in membranes
Mononuclear cell infiltration in decidua (%)
Tested (n)
Yes (%)
No (%)
P value
Yes (%)
No (%)
P value
309 421 343 421 443 177 446
0.0 6.3 5.1 6.8 7.3 8.6 1.8
12.7 11.4 10.3 11.8 12.9 15.3 11.0
.001 NS NS NS .05 NS .03
6.7 9.4 11.3 11.3 9.5 15.6 5.4
12.2 10.4 9.4 9.6 10.2 13.8 10.9
NS NS NS NS NS NS NS
C, Culture positive.
Table IV The percentage of women who have various newborn infant outcomes, compared with whether they have or do not have PMN infiltration at various sites (n = 446 women) PMN in membranes
PMN in fetal plate
PMN in umbilical cord
Neonatal outcome
Yes (%)
No (%)
P value
Yes (%)
No (%)
P value
Yes (%)
No (%)
P value
RDS Chronic lung disease Bronchopulmonary dysplasia IVH 3/4 Periventricular leukomalacia SIRS Sepsis NEC Death*
60.8 6.7 17.7 9.9 4.5 43.9 1.4 18.2 10.1
72.8 8.5 17.0 8.4 2.3 18.3 0.9 13.8 9.3
.008 NS NS NS NS !.0001 NS NS NS
58.6 5.8 15.5 11.2 2.4 50.6 1.7 17.8 10.3
72.4 8.4 18.0 8.0 4.0 18.0 0.8 15.7 9.5
.003 NS NS NS NS !.0001 NS NS NS
57.5 6.8 14.9 10.7 3.8 55.0 1.5 22.4 10.5
72.0 7.9 18.4 8.6 3.5 20.1 1.0 13.5 9.8
.003 NS NS NS NS !.0001 NS .02 NS
NS, Not significant. * Newborn infant death before discharge from hospital.
Mononuclear cell infiltration into the free membranes was less prevalent in women who had various markers of intrauterine infection/inflammation compared with women who did not have these markers. However, similar results were not seen with mononuclear cell infiltrations into the decidua basalis. The reason for this difference is not clear but may be related to the observation that most intrauterine infections are caused by bacteria that are ascending from the vagina; therefore, the infection is adjacent to the free membranes.1 The subsequent acute inflammatory response is predominantly neutrophilic and may also obscure a mononuclear response.11 In our data, PMN infiltration into various placental sites is associated with an increase in neonatal SIRS and into the umbilical cord is associated with NEC. Because neonatal SIRS is defined as an infant having the appearance of sepsis, but with negative cultures, the association is not surprising.21 We surmise that many infants attain high levels of circulating cytokines through their exposure to cytokines in the amniotic fluid but
do not become infected, in part because in approximately one half of the cases of intrauterine infection, the bacteria are confined to the membranes and do not reach the amniotic fluid.1,15 Whether the infant’s inflammatory response without concomitant infection is also explained by fetal exposure to nonviable bacteria, bacterial fragments, or unrecognized bacteria is unknown.19,20 The cause of NEC is also not clear but at times is thought to be related to bacterial infection of the infant’s gastrointestinal tract.22 Infants with funisitis are more likely to have organisms in the cord blood. Therefore, this relationship also is biologically plausible. Finally, several authors have suggested that, by enhancing the maturation of type II pneumocytes, HCA and the associated inflammatory response might be protective against RDS.9,10 Our results confirm that HCA is associated with a reduction in RDS. In this study, mononuclear cell infiltrations into various placental sites generally had little relationship to infant outcomes. However, the relationship between mononuclear cell infiltration into the decidua basalis
808 and IVH was relatively strong. The added finding of both IVH and neonatal death being associated with plasma cell infiltrates provides further support that something more than chance is responsible for the observed relationship. Mononuclear cell infiltration usually indicates a chronic bacterial or viral infection or an immune response to noninfectious antigens.10,11 Mononuclear cell infiltration at the maternal-fetal interface might also suggest poor placental function rather than exposure to high inflammatory cytokine levels as the cause of IVH in these infants. In summary, this study shows a strong but not perfect relationship between HCA, placental cultures, and other markers of infection/inflammation, which indicates that most, but not all, HCA is caused by bacterial infection. Placental inflammation was associated strongly with an increase in SIRS, and funisitis was associated strongly with NEC. HCA was associated with less RDS. The association between mononuclear and especially plasmacytic infiltration of the decidua basalis with IVH and neonatal death was unexpected but, if confirmed in other studies, may help to provide information about possible mechanisms that lead to these adverse neonatal outcomes.
References 1. Goldenberg RL, Hauth JC, Andrews WW. Intrauterine infection and preterm delivery. N Engl J Med 2000;342:1500-7. 2. O’Shea TM, Klinepeter KL, Meis PJ, Dillard RG. Intrauterine infection and the risk of cerebral palsy in very-low birthweight infants. Paediatr Perinat Epidemiol 1998;12:72-83. 3. Redline RW. Placental inflammation. Semin Neonatol 2004;9: 265-74. 4. Zhang J, Kraus FT, Aquino TI. Chorioamnionitis: a comparative histologic, bacteriologic, and clinical study. Int J Gynecol Pathol 1985;4:1-10. 5. Smulian JC, Shen-Schwarz S, Vintzileos AM, Lake MF, Ananth CV. Clinical chorioamnionitis and the histology of placental inflammation. Obstet Gynecol 1999;94:1000-5. 6. Redline RW, Heller D, Keating S, Kingdom J. Placental diagnostic criteria and clinical correlation: a workshop report. Placenta 2005;26(suppl):S114-7.
Andrews et al 7. Russell P. Inflammatory lesions of the human placenta: I, clinical significance of acute chorioamnionitis. Am J Diagn Gynecol Obstet 1979;1:127-37. 8. Redline RW. Severe fetal placental vascular lesions in term infants with neurologic impairment. Am J Obstet Gynecol 2005;192:452-7. 9. Watterberg KL, Demers L, Scott SM, Murphy S. Chorioamnionitis and early lung inflammation in infants in whom bronchopulmonary dysplasia develops. Pediatrics 1996;97:210-5. 10. Jobe AH, Ikegami M. Mechanisms initiating lung injury in the preterm. Early Hum Dev 1998;53:81-94. 11. Redline RW, Faye-Petersen O, Heller D, Qureshi F, Savell V, Vogler C, et al. Amniotic infection syndrome: nosiology and reproducibility of placental reaction patterns. Pediat Dev Pathol 2003;6:435-8. 12. Khong TY, Bendon RW, Qureshi F, Redline RW, Gould S, Stallmach T, et al. Chronic deciduitis in the placental basal plate: definition and interobserver reliability. Hum Pathol 2000;31: 292-5. 13. Volpe JJ. Hypoxic-ischemic encephalopathy: clinical aspects. In: Volpe JJ, editor. Neurology of the newborn. Philadelphia: Saunders; 2001. p. 331-94. 14. Bendon RW, Faye-Petersen O, Pavlova Z, Qureshi F, Elder N, Das A, et al. Histologic features of chorioamnion membrane rupture: development of methodology. Pediatr Pathol Lab Med 1997; 17:27-42. 15. Andrews WW, Hauth JC, Goldenberg RL, Gomez R, Romero R, Cassell GH. Amniotic fluid interleukin-6: correlation with upper genital tract microbial colonization and gestational age in women delivered after spontaneous labor versus indicated delivery. Am J Obstet Gynecol 1995;173:606-12. 16. Waites KB, Rikihisa Y, Taylor-Robinson D. Mycoplasma and ureaplasma. In: Murray PR, Baron EJ, Pfaller MA, Tenover FC, Yolken RH, editors. Manual of clinical microbiology. 8th ed. Washington, DC: ASM Press; 2003. p. 972-90. 17. Fiscella K. Racial disparities in preterm births: the role of urogenital infections. Public Health Rep 1996;111:104-13. 18. Pankuch GA, Applebaum PC, Lorenz RP, Botti JJ, Schachter J, Naeye RL. Placental microbiology and histology and the pathogenesis of chorioamnionitis. Obstet Gynecol 1984;64:802-6. 19. Relman DA. The search for unrecognized pathogens. Science 1999; 284:1308-10. 20. Fredricks DN, Fiedler TL, Marrazzo JM. Molecular identification of bacteria associated with bacterial vaginosis. N Engl J Med 2005; 353:1899-911. 21. Gomez R, Romero R, Ghezzi F, Yoon BH, Mazor M, Berry SM. The fetal inflammatory response syndrome. Am J Obstet Gynecol 1998;179:194-202. 22. Patole S. Prevention of necrotizing enterocolitis: year 2004 and beyond. J Matern Fetal Neonatal Med 2005;17:69-80.