The Preterm Prediction Study: Granulocyte colony-stimulating factor and spontaneous preterm birth

The Preterm Prediction Study: Granulocyte colony-stimulating factor and spontaneous preterm birth R o b e r t L Goldenberg, MD, a W'dliam W. Andrews, MD, PhD, a Brian M. Mercer, MD, b A t e f H. Moawad, MD, c Paul J. Meis, MD,dJay D. Iams, MD, e Anita Das, MS, k Steve N. Caritis, MD, f James M. Roberts, MD, f Menachem Miodovnik, MD,g Kathryn Menazd, MD, MPH, h Gary Thttrnau, MD, i Mitchell P. Dombrowski, MD~ a n d Donald McNellis, MD, 1for the National Institute o f Child Health and H n m a n D e v e l o p m e n t Maternal-Fetal Medicine Units Network Birmingham, Alabama, Cincinnati and Columbus, Ohio, Chicago, Illinois, Oklahoma City, Oklahoma, Pittsburgh, Pennsylvania, Charleston, South Carolina, Memphis, Tenness~, Winston-Salera, North Carolina, Detroit, Michigan, and Bethesda, Maryland OBJECTIVE: Granulocyte colony-stimulating factor is elevated in the amniotic fluid and plasma of women with chorioamnionitis and active preterm labor. We investigated the relationship between plasma granulocyte colonystimulating factor and subsequent spontaneous preterm birth in pregnant women without symptoms. STUDY DESIGN: We performed a nested case-control study involving 194 women who had a singleton spontaneous preterm birth and 194 matched term control subjects from the patient pool (n = 2929) enrolled in the Preterm Prediction Study. Plasma collected at 24 and 28 weeks' gestation was analyzed for granulocyte colony-stimulating factor, and the results were compared with subsequent spontaneous preterm birth. RESULTS: Compared with term control subjects, women who were delivered of their infants spontaneously at <28 weeks' gestation had increased mean granulocyte colony-stimulating factor values at 24 weeks' gestation (84.7 ± 38.4 vs 67.7 ± 28.6 pg/mL; P= .049), and women who were delivered of their infants at <32 weeks' gestation had increased mean plasma granuiocyte colony-stimulating factor values at 28 weeks' gestation (80.4 ± 24.1 vs 55.9 ± 16.5 pg/mL; P= .001). At 24 weeks' gestation a granulocyte colony-stimulating factor value >75th percentile in control subjects (approximately 80 pg/mL) was found in 48.9% (23/47) of all women delivered of their infants at <32 weeks' gestation versus 14.9% (7/47) of the term control subjects (adjusted odds ratio, 6.2; 95% confidence interval, 1.8-20.8). At 28 weeks' gestation a granulocyte colony-stimulating factor value >75th percentile was found in 36.8% (7/19) of women delivered of their infants at <32 weeks' gestation versus 5.30 (1/19) of term control subjects (adjusted odds ratio, 25.7; 95% confidence interval, 1.5-470.4). When measured at 24 or 28 weeks' gestation, granulocyte colony-stimulating factor did not predict spontaneous preterm birth at 32 to 34 weeks' gestation or at 35 to 38 weeks' gestation. CONCLUSION: In pregnant women without symptoms at 24 and 28 weeks' gestation, elevated plasma granulocyte colony-stimulating factor levels are associated with subsequent early (<32 weeks' gestation) spontaneous preterm birth, especially within the next 4 weeks, but not with late spontaneous preterm birth. These data provide further evidence that early spontaneous preterm birth is associated with an inflammatory process that is identifiable by the presence of a cytokine in maternal plasma several weeks before the early spontaneous preterm birth; however, later spontaneous preterm birth is not associated with this process. (Am J Obstet Gyneco12000;182:625-30.)

Key words: Cytokines, granulocyte colony-stimulating factor, preterm birth From the Departments of Obstenics and Gynecotog~ of the f o U ~ n g institutions--The University of Alabama at Bi~ningham, a the University of Tennessee,b the University of Chicago/Wake Forest University,'i Ohio State University/the University of Pittsburgh's Magee Women's Center,/ the University of Cincinnati, g the Medical University of South Carolina,h the University of Oklahoma,i and Wayne State Unive~j; The Biostatistics Cent~ George Washington Universityk; and the National Institute of Child Health and Human Development.t A list of participants in the National Institute of Child Health and Human Development Maternal-Fetal Medidne Units Network and their institutional affiliations appears at the end of the article. Supported by National Institute of Child Health and Human Development grants HD21410, HD21414, HD21434, HD27860, HD27861, HD27869, HD27883, HD27889, HD27905, HD27915, HD27917, and HD19897. Receivedfor publication April 16, ] 999; revisedSeptember14, 1999; accepted November Z 1999. Reprints not available from the authors. 6/1/104210 doi:l O.106 7/mob.2000.104 210

In r e c e n t years m a n y reports have suggested a relationship between b o t h vaginal a n d intrauterine infections and spontaneous p r e t e r m birth. Chronic chorioamnionitis, present as early as 16 to 20 weeks' gestation, has been i m p l i c a t e d as a cause o f early s p o n t a n e o u s p r e t e r m birth. I Unfortunately, d e f i n i n g which w o m e n have chronic c h o r i o a m n i o n i t i s before the onset o f labor o r rupture o f m e m b r a n e s has been dilficulL Amniocentesis with m e a s u r e m e n t o f various amniotic fluid cytokines has shown promise, b u t because o f the invasive nature of the procedure, it is n o t a p p r o p r i a t e for routine screening. Vaginal e x a m i n a t i o n for bacterial vaginosis also has shown promise for defining a population of women likely to have c h r o n i c chorioamnionitis. 2 However, bacterial vaginosis has a relatively low correlation with histologic 625

626 Goldenberget

al

chorioamnionitis, as well as a relatively low relative risk a n d sensitivity for p r e d i c t i n g s p o n t a n e o u s p r e t e r m birth.S Fetal fibronectin f o u n d in the vagina o r cervix after 24 weeks' gestation has b e e n l i n k e d to c h r o n i c chorioamnionitis, and it is currently the strongest predictor o f s p o n t a n e o u s p r e t e r m b i r t h yet d e s c r i b e d . 4, 5 W h e t h e r the relationship o f fetal fibronectin to infection occurs sufficiently early in the infectious process that leads to spontaneous preterm birth for a n effective intervention is unknown, but this is now u n d e r g o i n g evaluation. The relationship of various cytokine levels in cervical or vaginal fluid to subsequent preterm birth has been inconsistent, b u t several reports suggest a relationship between elevated levels and subsequent p r e t e r m birth. 6, 7 Measurement of various cytokines in maternal blood has also been proposed as a useful predictor o f spontaneous preterm birth, but to date predictive relationships have been described only in women already in labor. 8-10 O t h e r studies have described a relationship between elevated cytokine levels in labor and histologic chorioamnionitis found on delivery of the placenta. 9, 10 T h e r e are few, if any, examples o f cytoldnes m e a s u r e d in the s e r u m o r plasma o f pregnant women without symptoms and n o t in l a b o r t h a t p r e d i c t subsequent s p o n t a n e o u s p r e t e r m birth. Granulocyte colony-stimulating factor is a cytokine p r o d u c e d by monocytes, which, a m o n g o t h e r characteristics, attracts circulating leukocytes into tissue. In a previous study p e r f o r m e d by the National Institute o f Child H e a l t h a n d H u m a n D e v e l o p m e n t Maternal-Fetal Medicine Units Network in women with p r e t e r m rupture o f m e m b r a n e s , plasma granulocyte colony-stimulating factor levels increased 10-fold over a week's time, likely indicating intrauterine infection. 11 Because o f o u r interest in discovering predictors of subsequent spontaneous preterm birth in women without symptoms, we studied the relationship between granulocyte colony-stimulating factor and preterm birth related to infection in women u n d e r g o i n g routine prenatal care at 24 a n d 28 weeks' gestation. Methods

The original study in which these patients participated was the National Institute of Child Health and H u m a n D e v e l o p m e n t Maternal-Fetal Medicine Units Network Preterm Prediction Study, which was carried o u t from 1992_1994.3-5, 12, is The overall population a n d m e t h o d s for this study have been described in detail. Before the study the primary outcome was defined as a spontaneous p r e t e r m birth (after premature rupture o f m e m b r a n e s o r spontaneous labor) at <35 weeks' gestation. Gestational age was based on the last menstrual p e r i o d if the last menstrual p e r i o d and the earliest ultrasonographically d e t e r m i n e d biparietal diameter agreed within 10 days. If not, the biparietal d i a m e t e r was used to d e f i n e gesta-

March 2000 AmJ Obstet Gynecol

tional age. The initial study visit o c c u r r e d at 22 to 24 weeks' gestation. This r e p o r t is based on the analysis of plasma samples collected at 22 to 24 weeks' and 28 to 29 weeks' gestation. Granulocyte colony-stimulating factor concentrations in plasma were measured by means o f enzyme-linked imm u n o s o r b e n t assay ( I m m u n o t e c h , Marseille, France). This assay uses a specific m o n o c l o n a l antibody against the granulocyte colony-stimulating factor as the capture antibody and a second biotinylated monoclonal antibody as the detection antibody. T h e second antibody reacts with streptavidin peroxidase. After n o n b o u n d enzyme conjugate is removed by washing, substrate solution is added. The concentration of cytokine is directly proportional to the measured absorbance. The inter'assay coeiticient of variation was 2.94%, and the intra-assay coefficient of variation was 5.52%. O f the original population of 2929 women o f average risk for spontaneous preterm birth who were screened at 22 to 24 weeks' gestation, 207 women were delivered o f their infants spontaneously after either r u p t u r e o f the membranes or spontaneous preterm labor at <37 weeks' gestation. Of these 207 women, 194 had 24-week plasma samples available for granulocyte colony-stimulating factor analysis. These 194 w o m e n c o m p o s e d o u r study group and were matched for race, parity, and center with 194 women who were delivered o f their infants at >37 weeks' gestation, who served as the control group for this analysis. O n e h u n d r e d thirty-two of these w o m e n returned for a visit at about 28 weeks' gestation, and they and their control subjects also had a plasma sample available for analysis. Samples for fetal fibronectin were obtained from the cervix with a Dacron polyester swab before the performance of any other portion o f the pelvic examination. 4, 5 These samples were analyzed for fetal f i b r o n e c t i n by Adeza Biomedical Company (Sunnyvale, CatL0. A positive fetal fibronectin result was defined as any value ->50 n g / m L . The presence o f bacterial vaginosis was defined by Gram stain according to the criteria o f Nugent et al. 14 The methods for the cervical length determination have been described previously,is A cervical length o f 25 m m was approxa'mately the 10th percentile and by receiver operating characteristic curves appeared to be the best cutoff for the diagnosis of a short cervix. In addition to these data, which were obtained in all women, those involved in this nested case-control study also had urine analyzed for the presence of chlamydia with the ligase chain reaction method 15 and had the cervical fluid analyzed for levels of interleukin 6 (IL,6) 7 and sialidase 16 and had serum o r plasma analyzed for II-6, interleukin 10 (IL-10), C-reactive protein, [3~-microglobulin, ct-fetoprotein, and alkaline phosphatase. 17 All laboratories performing analyses were masked as to pregnancy outcome, with the results collated by the data center after completion of the tests.

Volume182,Number3 AmJ ObstetGynecol

This study was approved by each site's institutional review board, and each woman provided i n f o r m e d consent- Statistical analysis was performed with the SAS system. The Wilcoxon rank-sum test, the Z2 test, and the Fisher exact test were used to determine differences between groups. Correlations were d e t e r m i n e d by Spearman correlation analyses, Crude odds ratios a n d 95% confidence intervals were calculated. Logistic regression analyses were also conducted, and all of the following were controlled for: previous s p o n t a n e o u s preterm birth, acute pulmonary disease, urinary tract infection, vaginal bleeding, body mass index, race, contractions, pelvic infections, bacterial vaginosis, cervical length, and fetal fibronectin.

Results Table I shows the characteristics of women who were delivered of their infants at <37 weeks' gestation a n d their matched control subjects. Women who were delivered of their infants preterm were more likely to be thin or to have had pulmonary disease, a preterm birth, vaginal bleeding, or a urinary tract infection. Significant correlations were f o u n d between the 24and 28-week plasma granulocyte colony-stimulating factor values (r --- 0.44; P -- .0001). Plasma granulocyte colony-stimulating factor c o n c e n t r a t i o n s were significantly correlated with IL-10 at both 24 (r= 0.24; P = .001) and 28 (r= 0.17; P = .006) weeks' gestation b u t n o t consistently or significantlywith cervical or plasma ILAS,sialidase, fetal fibronectin, cervical length, C-reactive protein, [~2-microgiobulin, a-fetoprotein, alkaline phosphatase, the presence of chlamydia, or the bacterial vaginosis Gram stain score. Table II shows the mean +_ SD of plasma granulocyte colony-stimulating factor values at 24 and 28 weeks' gestation in women whose infants were delivered at 24 to 28, 29 to 31, 32 to 34, and 35 to 36 weeks' gestation a n d their matched control subjects. Women sampled at 24 weeks' gestation who were delivered of their infants at -<28 weeks' gestation had significantly higher m e a n plasma granulocyte colony-stimulating factor values compared with their matched control subjects, and women sampled at 28 weeks' gestation who were delivered of their infants at <32 weeks' gestation also had significantly higher m e a n plasma granulocyte colony-stimulating factor values than their m a t c h e d control subjects. However, at both 24 and 28 weeks' gestation, women who were delivered of their infants >4 weeks from sampling had plasma granulocyte colony-stimulating factor values that were n o t significantly different from those of their matched control subjects. Table 1II shows the percentage of case a n d control subjects at both 24 and 28 weeks' gestation who had plasma granulocyte colony-stimulating factor values >75th percentile. In this analysis, when tested at 24 weeks' gesta-

Goldenberg et al 627

Table L Maternal characteristics Characteristic

African American Age <16 y Age >35 y Education <12 y Nulliparous Lives alone Smoker Alcohol use Illicit drug use Pulmonary disease Body mass index <19.8 kg/m 9 Prior sponumeous preterm birth Vaginal bleeding Urinary tract infection

Cases (%) Control subjects (%) Statistical (n = 194) (n = 194) significance

72.7 1.0 3.6 43.3 33.5 25.8 99.9 9.3 5.7 11.3 30.5

72.7 2.6 3.1 35.6 33.5 20.6 25.8 15.5 7.9 4.6 19.2

P = 1.000 P= .449 P= .778 P= .119 P= 1.000 P= .239 P= .365 P= .064 P= .535 P= .015 P= .011

39.5

12.4

P= .001

35.1 17.5

95.3 9.3

P= .036 P = .017

tion, women who were delivered of their infants at 24 to 28 a n d at 29 to 31 weeks' gestation but n o t later had a significantly increased likelihood of having an elevated plasma granulocyte colony~stimulating factor concentration. Similarly, when tested at 28 weeks' gestation, women who were delivered of their infants at 28 to 31 weeks' gestation b u t not later had a significantly increased risk of having an elevated granulocyte colony-stimulatingfactor concentration. Both crude and adjusted odds ratios were determined for plasma granulocyte colony-stimulating factor values >75th percentile at both 24 and 28 weeks' gestation for spontaneous preterm birth at <32, 32 to 34, and 35 to 36 weeks' gestation. T h e crude odds ratios are shown in Table l~I. For testing performed at 24 weeks' gestation the adjusted odds ratio with 95% confidence interval for spontaneous preterm birth at 24 to 31 weeks' gestation was 6.2 (1.8-20.8), and for testing performed at 28 weeks' gestation the adjusted odds ratio with 95% confidence interval for spontaneous preterm birth at 28 to 31 weeks' gestation was 25.7 (1.5-470.4). Plasma granulocyte colony-stimulatingfactor values at both 24 and 28 weeks' gestation were n o t significantly associated with spontaneous preterm births at >-_32weeks' gestation. At 24 weeks' gestation the sensitivity of an elevated plasma granulocyte colony-stimulating factor value for predicting spontaneous preterm birth at 24 to 28 weeks' gestation was 54%, with a specificity of 79%. The positive and negative predictive values were 2.2% and 99%, respectively. At 28 weeks' gestation an elevated plasma granulocyte colony-stimulating factor value had a sensitivity of 37% and a specificity of 95% for spontaneous preterm birth at 28 to 31 weeks' gestation. The positive a n d negative predictive values were 7% a n d 99%, respectively.

628 Goldenberg et al

March~ooo AmJ Obstet Gynecol

Table I L Mean plasma granulocyte colony-stimulating factor values at 24 and 28 weeks' gestation in women who were delivered o f their infants spontaneously at <37 weeks' gestation and their matched control subjects

Plasma granulocyte colony stiraulating factor values (pg,/mL) Gestational age at delivery for cases (wk) 24-wk samples 24-28 29-31 32-34 35-36 28-wk samples 28-31 32-34 35-36

No. of cases/ No. of control subjects

Cases

Control subjects

Statistical significance

77/77

84.7 ± 38.4 69.8 ± 31.7 61.0 ± 25.8 65.6 ± 29.7

67.6 ± 28.6 56.5 ± 20.1 61.1 ± 42.0 73.1 ± 31.6

P= .049 P--- .084 P= .486 P= .143

19/19 53/53 64/64

80.4 ± 24.1 65.3 ± 29.5 66.4 ± 30.3

55.9 ± 16.5 68.8 ± 27.4 68.6 ± 27.5

P= .001 P= .323 P = .550

24/24 23/23 70/70

Table I l l . Elevated plasma granulocyte colony-stimulating factor levels and subsequent spontaneous preterm birth

>75th percentile Gestational age at deliveryfor cases (zok) 24-wk samples 24-28 29-31 32-34 35-36 28-wk samples 28-31 32-34 35-36

No. of cases/ No. of control subjects

Cases(%)

Controlsubjects ( %)

24/24 23/23 70/70 77/77

54.2 43.5 18.6 26.0

20.8 8.7 12.9 39.0

P= P= P= P=

19/19 53/53 64/64

36.8 22.6 21.9

5.3 30.2 26.6

P= .042 P= .378 P= .536

Comment To date, we have not been successfial in reducing the rate o f spontaneous preterm birth.18 Until o u r ability to p r e d i c t which women will likely have a s p o n tan eo u s pret erm birth is improved, allowing us to target specific interventions toward those women, it is unlikely that we will be successful in this endeavor. 19 R e d u c i n g late pret erm birth, the 50% that occur at 35 and 36 weeks' gestation, is n o t likely to have a major effect on perinatal outcomes. Instead, the 1% to 2% of infants born before 30 to 32 weeks' gestation account for much o f the mortality and both short- and long-term neurologic morbidity. Recent evidence suggests that the majority o f very early spontaneous p r et erm births are associated with, and probably caused by, a chronic bacterial infection within the chorion-decidual spaceA This infection initiates an inflammatory response modulated by various cytokines, one o f which is granuiocyte colony-stimulating factor. 20 This cytokine has been f o u n d elevated in association with many types of infections. During pregnancy, Mercer et al tl showed, in women with premature rupture of membranes at <32 weeks' gestation, that elevated plasma levels of granulocyte colony-stimulating factor were associated with a shorter time to delivery and more perinatal infections. Stallmach et a121 also found that in the presence o f

Statistical signif~noe .017 .007 .353 .085

Unadjusted odds ratio and 95% confidence interual 4.5 (1.3.-16.0) 8.1 (1.5-42.8) 1.5 (0.6--3:9) 0.6 (0.3--1.1) 10.5 (1.1-499) 0.7 (0.3-1.6) 0.8 (0.3-1.7)

chorioamnionitis the granulocyte colony-stimulating factor levels were increased 10- to 100-fold in amniotic fluid and about 10-fold in maternal blood. Immunolocalization of granulocyte colony-stimulating factor was f o u n d in placental macrophages, as well as decidual cells.~ Several other cytokines, such as the interleukins IL-6 and II_,8, have been found to be elevated in the amniotic fluid or ptasma of women in p r et er m labor who subsequently were delivered o f their infants prematurely compared with those o f co n t r o l w o m e n , often d e f i n e d as those who had contractions but who went on to be delivered of their infants at term. 8-10, 23 Similarly, w o m en with histologic chorioamnionitis t e n d e d to have elevated plasma cytokines. 9, 10, 94 O u r study is unique because the women tested at 24 and 28 weeks' gestation were receiving r o u t i n e prenatal care, were n o t c o n s i d e r e d to be high-risk status, and were n o t in labor or otherwise having symptoms at the time o f testing. The Preterm Prediction Study evaluated women at 24 and 28 weeks' gestation to determine predictors o f subseq u e n t p r e t e r m birth.4. 5 In previous analyses we have shown that a n u m b e r o f markers have been associated with spontaneous preterm birth, with several o f these factors being better predictors of early, compared with later, preterm birth. 19, 25 For example, serum or plasma alka-

Volume 182,Number 3

Goldenberg et al 629

AmJ Obstet Gynecol

line phosphatase and a-fetoprotein levels, a short cervix, vaginal and cervical fibronectin levels, cervical IL-6 level, and bacterial vaginosis were all more dosely associated with early, compared with late, preterm birth.12, 95 In this same population plasma granulocyte colony~stimulating factor was also associated with early, c o m p a r e d with late, preterm birth. However, plasma granulocyte Colony-stimulating factor levels were generally n o t correlated with these o t h e r predictors of early preterm birth. This lack of correlation suggests that determining the levels of several predictors may ultimately be useful in predicting early spontaneous preterm birth. Plasma granulocyte colony-stimulating factor appears to be a powerful predictor of subsequent early preterm birth, with adjusted odds ratios ranging from 6 to 25. To date, o f the putative predictors of early p r e t e r m birth, only fetal fibronectin has h a d consistently h i g h e r reported odds ratios. F o r example, in the overall population studied, at 24 weeks' gestation a positive fetal fib r o n e c t i n test result h a d an odds ratio o f 40 for p r e d i c t i n g a s p o n t a n e o u s p r e t e r m birth at <32 weeks' gestation, whereas the n e x t m o s t p o t e n t predictor, a short cervix d e t e r m i n e d by ultrasonography, had an odds ratio o f approximately 8. 4, is Bacterial vaginosis had an odds ratio o f approximately 2.3 for spontaneous preterm birth at <32 weeks' gestation, and most o t h e r putative predictors, if significant at all, had odds ratios of _<9.a Granulocyte colony-stimulating factor, measurable in plasma at 24 to 28 weeks' gestation, t h e r e f o r e appears to be a m o n g the most p o t e n t predictors o f early preterm birth yet described. Granulocyte colony-stimulating factor is a marker for inflammation.10, 11 The fact that elevated plasma granulocyte colony-stimulating factor at 24 and 28 weeks' gestation predicts spontaneous p r e t e r m birth, especially pret e r m b i r t h within 4 weeks, confirms a developing literature l i n k i n g early p r e t e r m birth a n d infection. 1 These results raise the issue o f the best gestational age for testing. 2~ T h e Preterm Prediction Study was begun at 24 weeks' gestation to predict p r e t e r m births at significant risk for morbidity and mortality. Most of the perinatal mortality occurs at <32 weeks' gestation, and intracerebral hemorrhage, periventricular leukomalacia, and, ultimately, cerebral palsy are m u c h more commonly found in these infants as well. Elevated p l a s m a granulocyte colony-stimulating factor at 24 and 28 weeks' gestation therefore is clearly associated with an increased risk of spontaneous p r e t e r m birth and delivery of infants most at risk for morbidity a n d mortality. W h e t h e r the test itself will ultimately prove useful in preventing p r e t e r m births will d e p e n d on a n u m b e r o f factors, including sensitivity a n d positive predictive values in various populations. Most important, whether a treatment strategy related to a positive test response will result in a reduction in preterm births remains to be d e t e r m i n e d .

Participating members a n d institutions f r o m the National Institute of Child Health and H u m a n D e v e l o p m e n t Maternal-Fetal Medicine Units Network are as follows: The University o f A l a b a m a at Birmingham--Rachel L. Copper, MSN, CRNP, J o h n C. Hauth, MD, and Allison N o r t h e n , R_N; Bowman Gray School of M e d i c i n e - - E b e r h a r d Mueller-Heubach, MD, Melissa Swain, RN, and Allison Frye, RN; University o f Chicago--Marshall Lindheimer, MD, and Phyllis Jones, MPH, RN; University o f CincinnatimTariq A. Siddiqi, MD, and Nancy Elder, MSN, RN; The Biostatistics Center, George Washington University--Elizabeth Thorn, Phi), Lucy Leuchtenburg, and Molly Fischer, MPH, CRNP; Magee Women's H o s p i t a l - - J a m e s H. Harger, MD, M a r g a r e t Cotroneo, RN, a n d Cynthia Stallings, MSN; National Institute of Child Health and H u m a n Development--SumnerJ. Yaffe, MD, Charlotte Catz, MD, and Mark Klebanoff, ME); T h e Ohio State University-Francee Johnson, RN, BSN, and Mark B. Landon, MD; University of Oklahoma--J. Christopher Carey, MD, and A r l e n e Meier, RN; Medical University o f S o u t h Carolina--J. Peter VanDorsten, MD, Roger B. Newman, MD, Beth A. Collins, PhD, RNC, and Faye LeBoeuf, MSN, CNM; University of Tennessee---Baha M. Sibai, MD, Risa Ramsey, BSN, RN, and Joyce Fricke, RNC; Wayne State University---Sydney F. Bottoms, MD (deceased), a n d Gwendolyn S. Norman, MPH, RN. REFERENCES

1. Andrews WW,Goldenberg RL HauthJC. Preterm labor, emerging role of genkal tract infections. Infect Agents Dis 1995;4:196211. 2. Krohn MA, Hillier SL, Nugent RP, Cotch MF, Carey JC, Gibbs RS, et al, for the Vaginal Infection and Prematurity Study Group. The genital flora of women with intmmnniotic infection. J Infect Dis 1995;171:1475-80. 3. Meis PJ, Goldenberg RL, Mercer B, MoawadA, Das A, McNellis D, et al. The Preterm Prediction Study: significance of vaginal infections. AmJ Obstet Gynecol 1995;173:1231-5. 4. Goldenberg RL, Mercer BM, Meis PJ, Copper RL, Das A, McNe]lisD. The Premrm Prediction Study:fetal fibronecdn testing and spontaneous preterm birth. Obstet Gynecol 1996;87:643-8.

5. Goldenberg RL, Thorn E, Moawad AH, Johnson F, Roberts J, Caritis SN. The Preterm Prediction Study:fetal fibronecfin, bacterial vaginosis and peripartum infection. Obstet Gynecol 1996;87:656-60. 6. Rizzo G, Capponi A, Rinaldo D, Tedeschi D, Arduini D, Romanini C. Interleuldn-6 concentrations in cervical secretions identify microbial invasion of the anmiotic cavity in patients with preterm labor and intact membranes. AmJ Obstet Gynecol 1996;175:819-7. 7. Goepfert AR. The Preterm Prediction Study: quantitative fetal fibronectin (FFN) values and the prediction of spontaneous preterm birth (SPTB) [abstract]. AmJ Obstet Gynecol 1997;176 (I Pt 2):$52. 8. Murtha AP, Greig PC, Jimmerson CE, Herbert WNP. Maternal serum interleukin-6 concentration as a marker for impending preterm delivery.Obstet Gyneco11998;91:161-4. 9. Maeda K, Matsuzaki N, Fuke S, Mitsuda N, Shimoya K, Nakayama M, et al. Value of the maternal interleukin 6 level for determination of histologic chorioanmionitis in preterm dclively. Gynecol Obstet Invest 1997;43:295-31. 10. Hillier SL, Witkin SS, Krohn M_A,Watts DH, Kiviat NB, Eschcnbach DA. The relationship of amniotic fluid cytokines and preterm delivery, amniotic fluid infection, histologic

630 Goldenberg et al

chorioamnionitis, and chorioamnion infection. Obstet Gynccol 1993;81:941-8. II. Mercer B for the N I C H D M F M U Network. Maternal and neonatal plasma cytokinc levels and pcrinatal morbidity [abstract]. J Soc Gynecol Invcstig 1997;4:154A. 12. Goldenberg R.L, laresJD, Mercer BM, Mcis PJ, Moawad AH, Copper RL, et al. The Prctcrm Prediction Study: the value of new vs. standard risk factors in predicting early and all spontancous pretcrrn birth.ArnJ Public Health 1998;88:233-8. 13. Iams JD, Goldcnbcrg RL, Meis PJ, Mercer BM, Moawad A, D~s A, ct al. The length of the cervix and the risk of spontaneous premature delivery.N EnglJ Mcd 1996;334:567-72. 14. Nugent RP, Krohn M.A, HillicrSL. Reliabilityof diagnosing bacterialvaginosisisimproved by standard method of gram staininterpretalion.J Clin Microbiol 1991;29:297-301. 15. Andrews W'W, Goldcnberg RL, Hauth JC, Mercer B, lares J, Meis P, et al. The Preterm Prediction Study: association of midtrimester genital chlamydia infection and subsequent spontaneous preterm birth.A m J Obstet Gynccol. In press 2000. 16. Andrews W W , Tsao J, Goldcnbcrg RL, Hauth JC, Mercer B, Iams J, et al. The Preterm Prediction Study: failure of midtrimester cervicalsialidasclevelelevation to predict subsequent spontaneous prctcrm birth. A m J Obstet Gynccol 1999; 180:1151-4. 17. Moawad AH. The Prcterm Prediction Study:.the value of serum

alpha-fetoprotein (AFP) and alkaline phosphatase (ALKPHOS)

March 2ooo AmJ ObstetGynecol

in prediction of spontaneous preterm birth (SPTB) [abstract]. AmJ Obstet Gyneco11998;178 (1 Pt 2):$194. 18. Goldenberg RL, Rouse DJ. An evaluation of interventions used to prevent or improve the outcome of pregnancy associated with preterm birth. N EnglJ Med 1998;339:31H20. 19. Goldenberg RL, Andrews WW. Intrauterine infection and why preterm prevention programs have failed. Am J Public Health 1996;174:1618-21. 20. Luster AD. Chemokines--chemotactic cytokines that mediate inflammation. N EnglJ Med 1998;338:436-45. 21. Stallmach T, Hebisch G,Joller-Jemelka HI, Orban P, SchwallerJ, Engelmann M. Cytokine production and visl,sliTed effects in the feto-matemal unit. Lab Invest 1995;73:384-92. 22. Shorter SC, Vince GS, Starkey PM. Production of granulocyte colony-stimulating factor at the materno-foetal interface in human pregnancy. Immunology 1992;75:468-74. 23. Salto S, Kasahara T, Kato Y, IshiharaY, Ichijo M. Elevation of amniotic fluid interleukin 6 (IL-6), IL-8 and granulocyte colony stimulating factor (G-CSF) in term and preterm parturition. Cytokine 1993;5:81-8. 24: Shimoya K, Matsusaki N, Taniguchi T, Kameda T, Koyama M, Neki R, et al. Human placenta constitutively produces interleukin-8 during pregnancy and enhances its production in intrauterine infection. Biol Reprod 1992;47:220-6. 25. Goldcnberg RL, Andrews WW, Hauth JC. Markers of preterm birth. Prenat Neonat Med 1998;3:43-6.

Availability of Journal back issues As a service t o o u r subscribers, copies o f back issues o f the A m e r i c a n J o u r n a l o f Obstetrics and Gynecology for the p r e c e d i n g 5 years are m a i n t a i n e d and are available for purchase f r o m Mosby until inventory is depleted. T h e following quantity discounts are available: o n e f o u r t h o f f on quantities o f 12 to 23 a n d o n e third off o n quantities o f 24 o r more. Please write to Mosby, Inc, S u b s c r i p t i o n Services, 11830 Westline Industrial Dr, St. Louis, M O 83146-3318, t e l e p h o n e (800)453-4351 o r (314)453-4351, for information o n availability o f particular issues and prices. If back issues are unavailable f r o m the publisher, p h o t o c o p i e s o f c o m p l e t e issues may be p u r c h a s e d f r o m Bell & H o w e l l I n f o r m a t i o n a n d Learning, 300 N Z e e b Rd, A n n Arbor, MI 48106-1346. T e l e p h o n e (734)761-4700 o r (800)521-0600.