American Journal of Obstetrics and Gynecology (2006) 194, 1176–85

www.ajog.org

Are women with recurrent spontaneous preterm births

different from those without such history?
Brian M. Mercer, MD,a Cora A. Macpherson, PhD,b Robert L. Goldenberg, MD,c
Alice R. Goepfert, MD,c Sylvie Haugel-De Mouzon, PhD,a Michael W.
Varner, MD,d Jay D. Iams, MD,e Paul J. Meis, MD,f Atef H. Moawad, MD,g
Menachem Miodovnik, MD,h Steve N. Caritis, MD,i J. Peter Van Dorsten, MD,j
Yoram Sorokin, MD,k Gary R. Thurnau, MD,l Catherine Y. Spong, MD,m for the
NICHD Maternal-Fetal Medicine Units Network, Bethesda, MD

Department of Reproductive Biology, Case Western Reserve University,a Cleveland, OH; Biostatistics Center, George
Washington University,b Washington, DC; Departments of Obstetrics and Gynecology of University of Alabama at
Birmingham,c Birmingham, AL; University of Utah School of Medicine,d Salt Lake City, UT; Ohio State University,e
Columbus, OH; Wake Forest University,f Winston-Salem, NC; University of Chicago,g Chicago, IL; University of
Cincinnati,h Cincinnati, OH; Magee Women’s Hospital-University of Pittsburgh,i Pittsburgh, PA; Medical University
of South Carolina,j Charleston, SC; Wayne State University,k Detroit, MI; University of Oklahoma,l Oklahoma City,
OK; National Institute of Child Health and Human Development,m Bethesda, MD

Received for publication November 9, 2005; revised December 16, 2005; accepted January 20, 2006

KEY WORDS Objective: This study was undertaken to determine whether women with recurrent spontaneous
Preterm birth preterm births (rSPBs) have different clinical characteristics or systemic markers than those with
Recurrence isolated preterm (iSPBs) or recurrent term births (rTBs), when assessed remote from delivery.
Prediction Study design: We compared clinical characteristics and findings (including cervical ultrasound,
Risk factors bacterial vaginosis, fetal fibronectin), maternal plasma markers obtained at 22 to 24 weeks’ ges-
tation (inflammatory cytokines, cortisol, and corticotrophin-releasing hormone), between women
with rSPBs (2 or 3 consecutive SPBs and no TBs), iSPBs (1 SPB and 1 or 2 TBs), and rTBs (2 or 3
consecutive TBs and no SPBs).
Results: A total of 1257 women met our inclusion criteria; 47 rSPBs, 241 iSPBs (80 current and
161 prior iSPBs), and 969 rTBs. Before pregnancy, women with rSPBs had lower weights (P !
.0001) and body mass indexes (BMIs) (P ! .001), and were more likely to be less than 100 lbs
(P = .008) or less than 19.8 kg/m2 BMI (P = .001). At 22 to 24 weeks those with rSPBs remained
lighter and leaner, and had more advanced Bishop scores than iSPBs and rTBs. Ultrasound

Supported by grants from the National Institute of Child Health and Human Development (HD21410, HD21414, HD27860, HD27861,
HD27869, HD27883, HD27889, HD27905, HD27915, HD27917, HD19897, HD36801, HD34208, and HD40544), and the Department of Repro-
ductive Biology, Case Western Reserve University, MetroHealth Medical Center.
Presented at the Twenty-Fourth Annual Meeting of the American Gynecological and Obstetrical Society, September 29 through October 1, 2005,
Victoria, British Columbia, Canada.
Reprints not available from the authors.

0002-9378/$ - see front matter Ó 2006 Mosby, Inc. All rights reserved.
doi:10.1016/j.ajog.2006.01.069
Mercer et al 1177

demonstrated progressive decrease in cervical length for those with rTBs, prior iSPBs, current
iSPBs, and rSPBs, and also progressively more frequent short cervixes with worsening history
(P ! .001). Cervical length was shorter for women of lower pregravid weight and BMI, but
not with shorter height. At 22 to 24 weeks, women with rSPBs had more common uterine con-
tractions and tocolytic agents, but not more infections or antibiotic therapy. Those with an
SPB in the current gestation had higher fetal fibronectin levels and more frequent vaginal bleed-
ing, regardless of prior outcome. Maternal cortisol and corticotrophin-releasing hormone were
higher in women with iSPBs and rSPBs than in rTB controls, (P = .001 and .0027), a finding
more apparent with SPB in the current pregnancy. However, maternal cytokines were not
increased with either iSPBs or rSPBs.
Conclusion: Women with rSPBs are leaner, contract more, have shorter cervixes, and have more
advanced Bishop scores than women with iSPBs or rTBs.
Ó 2006 Mosby, Inc. All rights reserved.

Despite extensive efforts directed towards prevention
of preterm birth, the incidence of prematurity has
increased from 9% to 12% over the past 2 decades and
remains twice as common among black women as white
women.1,2 Long-term sequelae, including neurologic
handicaps, blindness, deafness, and chronic respiratory
disease, are directly linked to preterm birth, particularly
birth occurring before 30 weeks’ gestation.3-5 In some
cases, preterm birth has been associated with intrauter-
ine inflammation6-8 and biochemical evidence of sys-
temic inflammation9,10 or to markers of maternal
stress.11-14 Several lines of evidence have also supported
a genetic predisposition to preterm birth among some
women.15-17
Prior preterm birth is a leading risk factor for preterm
birth.18-22 Women who deliver preterm have a number
of differences when compared with those who deliver
at term. These include clinical findings, demographic
characteristics, cervical length estimated by ultrasound,
cervical-vaginal fetal fibronectin levels, and blood and
cervical markers obtained remote from delivery.23-30
However, it is not known if women with recurrent pre-
term births are different from those with isolated pre-
term births or those with term births only.
The purpose of this study was to determine whether
women with recurrent preterm births demonstrate dif-
ferences in clinical characteristics and physical findings,
markers for preterm birth (systemic, vaginal, sono-
graphic) as compared with those having isolated pre-
term births or recurrent term births (rTBs) only, when
evaluated remote from delivery.
Material and methods
This is a secondary analysis of data collected by 10
clinical centers participating in the NICHD Maternal-
Fetal Medicine Units Network in the Preterm Prediction
Study between 1992 and 1994. This study was per-
formed with institutional review board approval from
each participating center and informed consent of each
patient. The current analysis was approved by the
institutional review board of MetroHealth Medical
Center–Case Western Reserve University.
Specific details regarding the study design, and the
predictive values of clinical risk assessment, fetal fibro-
nectin and bacterial vaginosis screening, and cervical
sonography for spontaneous preterm birth have been
published in the principal analyses of this study.20,23-28
To evaluate clinical characteristics, physical findings,
vaginal and sonographic markers, we studied gravid
women with singleton pregnancies who were initially
evaluated at 22 to 24 weeks’ gestation. Excluded were
women with placenta previa, major fetal malformations,
poly- or oligohydramnios, human immunodeficiency vi-
rus, cervical cerclage in situ, as well those with sympto-
matic preterm labor, premature rupture of membranes
(PROM), cervical dilatation (R2 cm for nulliparous
women, R3 cm for multiparous women) or fetal mem-
brane prolapse to the external os. Gestational age was
determined in a standardized manner based on the
best obstetric estimate using last menstrual period and
the earliest ultrasound. Each participant underwent a
structured interview to ascertain demographic factors
and clinical characteristics, socioeconomic status,
home and work environment, drug or alcohol use, and
prior medical and obstetric history. Medical history,
prior, and current obstetric information were docu-
mented through medical record review where possible.
When a history of preterm birth or low birth weight
was obtained, the gestational age at delivery was docu-
mented and those with a preterm birth (20-366 weeks’)
caused by preterm labor or PROM were defined to
have had a ‘‘spontaneous preterm birth’’ (SPB). Mater-
nal pregravid weight was obtained by patient recall.
Current weight and height were measured, and body
mass index (BMI) was calculated (kg/m2). Cervical
and vaginal fibronectin samples and specimens for eval-
uation of bacterial vaginosis (based on Nugent score
and pH) were obtained by speculum examination.
Bishop score was subsequently evaluated by digital
examination, and cervical length was evaluated by vag-
inal ultrasound. Results of fetal fibronectin, bacterial
1178 Mercer et al

Table I Demographic and clinical characteristics of 1257 multiparous women with or without rSPBs

Term Isolated SPB Recurrent

Deliveries Only Prior pregnancy Current pregnancy SPBs
n = 969 n = 161 n = 80 n = 47 P value
Age (y)* 24.8 (5.1) 23.8 (4.9) 24.6 (5.5) 23.2 (4.9) .01
Race (% black) 62.8 62.1 67.5 74.5 .34
Parity = 2 (%) 67.4 61.5 67.5 68.1 .52
Education (y)* 12.0 (1.9) 11.8 (1.6) 11.8 (1.4) 11.4 (1.8) .14
Married (%) 34.7 34.2 23.8 29.8 .23
Government insurance (%) 89.4 91.3 90 89.4 .90
Income !$800/mo (%) 59.3 47.8 72.5 57.5 .008
Work in pregnancy (%) 42.6 41.0 43.8 44.7 .96
Pelvic inflammatory disease (%) 6.2 4.4 6.3 10.6 .43
Tobacco use in pregnancy (%) 33.2 34.2 38.8 19.2 .15
Alcohol use in pregnancy (%) 4.3 8.0 1.8 6.1 .25
Diabetes (%) 3.1 6.2 1.3 0.0 .10
Chronic hypertension (%) 4.9 1.9 3.8 2.1 .35
Chronic lung disease (%) 1.1 2.5 1.3 0.0 .46
Mullerian abnormalities (%) 0 0.62 0.00 2.1 .02
Pregravid weight (lbs) 155 (43.4) 143 (36.5) 150 (40.9) 128 (25.6) !.0001
Pregravid weight !100 lbs (%) 2.6 2.6 6.7 10.9 .008
Pregravid BMI (kg/m2)* 26.2 (7.0) 24.7 (6.2) 25.3 (6.4) 21.6 (4.2) !.0001
Pregravid BMI !19.8 kg/m2 (%) 16.2 20.0 22.7 37.8 .001
* Mean (SD).

vaginosis screening, and cervical sonography were with-
held from caregivers and other study personnel, unless
fetal death, fetal membrane prolapse to the external
cervical os, poly- or oligohydramnios, or symptomatic
regular uterine contractions were identified. All parti-
cipants were monitored prospectively to determine
pregnancy outcomes.
From the above cohort, we identified and included
only those women with 2 or 3 completed singleton
pregnancies (including the current gestation) delivering
at 20 or more completed weeks gestation. These gravid
women were divided into 1 of 3 groups: recurrent SPBs
(rSPBs) (2 or 3 consecutive SPBs and no TBs), isolated
SPBs (iSPBs) (1 SPB and 1 or 2 TBs), and recurrent TBs
(rTBs: 2 or 3 consecutive TBs and no SPBs). Those with
iSPBs were further divided into those with an iSPB in a
prior pregnancy or iSPB in the current pregnancy.
To evaluate systemic markers for SPB, we performed
a nested case-control study of inflammatory cytokines
(interleukin 1-beta [IL1-beta], interleukin-6 [IL-6], and
interleukin-10 [IL-10]), and stress markers (corticotro-
phin-releasing hormone [CRH], and cortisol) from
untimed plasma samples obtained by maternal venous
blood draw into EDTA-coated tubes at the 22- to 24-
week visit. Specimens were processed, aliquoted, and
stored at –70(C within 2 hours of collection. For each
rSPB case, we identified 2 iSPB and 2 rTB controls.
Frozen samples were transferred to MetroHealth Med-
ical Center–Case Western Reserve University where
high-sensitivity enzyme-linked immunosorbent assays
(ELISAs) were conducted in duplicate for each marker
in batch fashion. The assays were performed by using
standard high-sensitivity kits and techniques for cyto-
kines (Quantikine-HS, R&D Systems Inc, Minneapolis,
MN); IL1-beta (range: 0.1-8 pg/mL), IL-6 (range: 0.04-
10 pg/mL), IL-10 (range: 0.5-50 pg/mL), cortisol (range:
0.1-60 mg/dL, Diagnostic Systems Laboratories Inc,
Webster, TX), and CRH (range: 25 ng/mL, Penninsula
Laboratories Inc, San Carlos, CA). Mean results of the
duplicate samples for each analyte were compared
between women with rSPBs, iSPBs (current or prior
pregnancy), and rTBs.
Statistical analyses were performed using SAS 8.2
(Cary, NC). Categorical variables were compared by c2
or Fisher exact tests, where appropriate. Continuous
variables were compared using the Wilcoxon rank-sum
test or Kruskal-Wallis test for continuous data. Nomi-
nal 2-sided P-values are reported with no adjustments
made for multiple comparisons. For this analysis a
P-value below .05 was considered statistically significant.
Results
From a cohort of 2929 gravid women, we identified 1257
women (42.9%) meeting our inclusion criteria; 47 rSPB
cases, 241 iSPBs (80 with SPB in the current pregnancy,
and 161 with a prior SPB), and 969 rTB controls.
Clinical and demographic characteristics evaluated at
the 22- to 24-week visit are presented in Table I. Women
with rSPBs were younger; however, the 0.6- to 1.6-year
difference is not likely clinically meaningful. Women
Mercer et al 1179

Table II Clinical symptoms and findings, and ancillary evaluations (cervical sonography, cervicovaginal fetal fibronectin screening,
bacterial vaginosis screening) identified at 22 to 24 weeks’ gestation among 1257 multiparous women with or without rSPBs

Term Isolated SPB Recurrent

Deliveries Only Prior pregnancy Current pregnancy SPBs
n = 969 n = 161 n = 80 n = 47 P value
First or second trimester vaginal 24.5 26.7 38.8 40.4 .005
bleeding (%)
Uterine contractions (%)* 18.7 21.7 20.0 42.6 .001
Prior tocolytic administration (%) 0.4 1.2 0 4.3 .03
Antibiotic Rx in pregnancy (%) 39.3 41.0 47.5 42.6 .52
Trichomonas (%) 3.5 3.7 0 4.3 .30
Bacterial vaginosis (%) 23.9 21.7 28.8 29.8 .51
Nugent score R9 (%) 13.6 13.0 22.5 14.9 .17
pH R5 (%) 30.7 30.4 35 40.4 .46
Maternal weight (lbs)y 172 (44.6) 162 (39.0) 167 (41.7) 144 (26.6) !.0001
Weight !115 lbs (%) 3.6 6.2 7.5 12.8 .009
BMI (kg/m2)y 29.2 (7.1) 27.9 (6.4) 28.3 (6.5) 24.3 (4.4) !.0001
BMI !19.8 kg/m2 (%) 4.0 3.1 5.0 15.2 .01
Digital cervical examination
Cervical length %1 cm (%) 0.7 6.2 1.3 4.3 !.0001
Internal os dilation O1 cm (%) 0.3 0 0 8.0 !.019
Cervical consistency (% soft) 6.6 9.9 15.0 23.4 !.0001
Bishop score R4 (%) 9.3 16.8 12.5 31.9 !.0001
Protruding fetal membranes (%) 4.4 7.5 16.3 21.3 !.0001
Ultrasound cervical length (mm)y 36.7 (7.8) 35.1 (8.7) 33.9 (9.9) 29.4 (8.2) !.0001
Cervical fetal fibronectin (ng/mL)y 50.9 (21.5) 49.4 (5.0) 72.7 (80.9) 63.4 (56.6) !.0001
Vaginal fetal fibronectin (ng/mL)y 53.3 (37.6) 52.4 (41.9) 71.1 (94.7) 81.9 (121.8) !.0001
* Symptomatic contractions within the preceding 2 wks.
y
Mean (SD).

with iSPB in the current gestation were most likely to
have a low family income. Women with rSPBs were
more likely to have known mullerian abnormalities
than those with isolated or no SPBs (P = .02). Before
pregnancy, women with rSPBs were lighter (P !
.0001), had lower BMIs (P ! .001), and were more
likely to have maternal weight less than 100 lbs (P =
.008) or BMI 19.8 kg/m2 (P = .01). No differences in
other clinical characteristics, including responses to
structured questionnaires regarding anxiety, self-esteem,
mastery, or perceived stress were seen between groups
(data not shown).
At 22 to 24 weeks, those with rSPBs were more likely
to have had symptomatic uterine contractions within the
preceding 2 weeks, and to have been treated with
tocolytic agents (Table II); however, no differences in
vaginal infections or need for antibiotic therapy were
seen. The finding of lower maternal weight and BMI
among those with rSPBs persisted at the 22- to 24-
week visit. Women with rSPBs had worse findings on
digital cervical examination, including total Bishop
scores and individual components. Women with iSPBs
in the current and past pregnancy demonstrated findings
intermediate between those with rSPBs and rTB con-
trols. On transvaginal ultrasound mean cervical length
was progressively shorter (Table II), and the frequency
of short cervix (!30, 25, 20, and 15 mm) was progres-
sively more common for those with prior iSPB, current
iSPB, and rSPB (P ! .001 for each cutoff, Figure 1)
when compared with rTB controls. Alternatively, first-
or second-trimester vaginal bleeding and cervical or vag-
inal fetal fibronectin levels were higher for those with an
SPB in the current gestation regardless of prior outcome
(Table II, Figure 2).
Comparison of those with iSPBs in the current
pregnancy and those with rSPBs revealed women with
rSPBs to be twice as likely to deliver before 35 weeks’
(63.8% vs 30.0%, P!.001), and before 32 weeks’ gesta-
tion (25.5% vs 12.5%, P = .06). However, the frequen-
cies of SPB caused by preterm labor (61.7% vs 60.0%)
and PROM (38.3% vs 40.0%) were similar.
The 230 maternal plasma samples collected at 22 to
24 weeks included: 46 rSPBs, 92 iSPBs (60 with prior
and 32 with current pregnancy SPBs), and 92 rTB
controls. Two were subsequently determined to have
been collected outside the study window (21 weeks) and
were excluded. Significant overlap was seen between
study groups for each parameter studied; however,
maternal cortisol and CRH levels were significantly
different between groups (Table III, Figure 3), with
1180
Figure 1 Frequency of sonographic cervical length below
various cutoffs at 22 to 24 weeks gestation for 1257 multipa-
rous women with: rSPBs (closed circles), an iSPB in the current
gestation (open diamonds), an iSPB in a prior gestation (closed
diamonds), and rTB only controls (open triangles).
median levels increasing with worsening pregnancy out-
comes, particularly with SPB the current gestation. Ma-
ternal interleukins were not increased among those with
iSPBs or rSPBs. IL-6 levels were significantly different
between groups, but appeared lower among those with
rSPBs (Table III).
Because women with rSPBs were lighter and leaner,
had shorter cervical lengths, and also appeared to have
lower IL-6 levels, we evaluated correlations between
these findings. Those with lower pregravid weights and
BMIs had shorter cervixes on transvaginal ultrasound
(P = .003 and .007). We found no correlation between
maternal height and sonographic cervical length (P =
.45). Women with a pregravid weight less than 100 lbs
were more likely to have an ultrasound cervical length
less than 25 mm (10.8% vs 3.2%, P = .04) and 35 mm
or less (67.6% vs 40.9%, P = .002). Among the 219
women with maternal plasma cytokine values and
pregravid weights available, IL-6 levels were lower
among women who were less than 115 lbs before preg-
nancy (median: 0.92 vs 1.42 pg/mL, P = .001), and
among those with a pregravid BMI less than 19.8 kg/m2
(median: 0.88 vs 1.47 pg/mL, P = .0001) as compared
with larger and less lean women.
Comment
We have previously reported that clinical risk factors, a
short cervix on ultrasound, bacterial vaginosis, and a
positive fetal fibronectin screen, evaluated at 22 to 24
weeks’ gestation identify those at increased risk for SPB
in the current pregnancy.20,23-27 Neither systemic
markers of inflammation (interleukins) nor stress (corti-
sol, CRH) were increased among those delivering before
Mercer et al
Figure 2 Results of cervical and vaginal fetal fibronectin
screening (ng/mL) at 22 to 24 weeks gestation among 1257
multiparous women with: rSPBs (black), an iSPB in the current
gestation (dark grey), an iSPB in a prior gestation (light grey),
and rTB only controls (white) (*P ! .0001, **P = .002).
35 or 32 weeks’ gestation.28,30 It is not known if women
with repeated SPBs are inherently different from those
without such a history. The objective of the current
study was to determine whether women with rSPBs
demonstrate differences in clinical characteristics, bio-
logic markers, from those without rSPBs. To address
this issue we focused on those with consecutive SPBs
and those with consecutive TBs rather than focusing
only on the current pregnancy outcome. Because acute
clinical findings at the time of preterm delivery may
mask underlying chronic characteristics, we assessed
those characteristics available remote from delivery
(22-24 weeks’ evaluation), and before conception (his-
torical information). By selecting those extreme out-
comes (only preterm deliveries or only term deliveries)
we anticipated that any major biologic differences
between these groups should be evident. In addition,
through inclusion of women with isolated preterm births
in the current or in a prior pregnancy, we intended
to determine whether there were intermediate risks in
this population, reflecting acute pregnancy events or a
chronic predisposition to preterm birth.
We have found highly significant differences, with the
rSPB group demonstrating lower maternal weights and
BMIs before pregnancy and at 22 to 24 weeks as
compared with those having iSPBs or rTBs. Physical
findings at 22 to 24 weeks were different between groups,
with those having rSPBs having cervixes that are
shorter, more dilated, softer, and with a higher Bishop
score. These physical findings are mirrored by transvag-
inal ultrasound findings of shorter cervixes, and more
frequent short cervices for each cutoff value studied.
These findings suggest that women with rSPBs are
different in regard to in body composition, possibly
through suboptimal nutrition or through constitutional
differences, remote from delivery, and even before
Mercer et al 1181

Table III Nested case control analysis of plasma cytokines and stress markers between multiparous women with rSPBs, iSPB in a prior
or current pregnancy, and rTBs

Recurrent Term Isolated SPB Recurrent

Deliveries Prior pregnancy Current pregnancy SPBs
n = 92 n = 60 n = 30 n = 46 P value*
IL1-b (pg/mL)
25th percentile 0.10 0.08 0.00 0.09 .63
50th percentile 0.17 0.15 0.15 0.14
75th percentile 0.39 0.32 0.29 0.26
IL-6 (pg/mL)
25th percentile 0.94 1.00 0.69 0.81 .03
50th percentile 1.43 1.53 1.09 1.07
75th percentile 2.27 2.30 2.10 1.63
IL-10 (pg/mL)
25th percentile 3.92 3.68 3.71 3.45 .56
50th percentile 4.52 4.52 4.62 4.25
75th percentile 5.28 5.05 5.24 5.15
Cortisol (mg/mL)
25th percentile 14.2 13.4 17.2 16.8 .001
50th percentile 17.0 16.2 19.3 23.8
75th percentile 21.5 22.5 21.5 27.5
CRH (ng/mL)
25th percentile 0.17 0.18 0.19 0.20 .0027
50th percentile 0.21 0.23 0.28 0.29
75th percentile 0.27 0.28 0.32 0.32
CRH, Corticotrophin releasing hormone.
* Kruskal-Wallis test.

pregnancy. The evident correlation between low mater-
nal weight and body mass with short cervical length,
combined with a lack of correlation between height and
cervical length, supports the assertion that these women
are constitutionally different rather than simply smaller.
Maternal body habitus (weight, BMI) and cervical
length findings on transvaginal ultrasound demon-
strated progressive worsening with worsening obstetric
history. We hypothesize that some women with isolated
preterm births have characteristics intermediate between
those with recurrent preterm births and recurrent term
birth controls, and that acute or evolving subacute
factors could precipitate preterm delivery in these
women.
Alternatively first- and second-trimester bleeding,
protruding fetal membranes on digital examination,
elevated cervical and vaginal fetal fibronectin levels,
elevated cortisol and CRH values appeared to reflect
SPB in the current pregnancy regardless of prior out-
come. These characteristics suggest more acute local or
environmental factors such as decidual/placental disrup-
tion and underlying maternal stress to be more closely
linked to current pregnancy outcomes rather than
inherent patient characteristics.
We anticipated that women with rSPBs might dem-
onstrate elevations of circulating maternal cytokines
when compared with women having iSPBs or rTBs.
Although more subtle differences may have been
missed, it is not completely surprising that these
maternal systemic inflammatory markers were not ele-
vated as prior case-control study of women with or
without SPBs from this population revealed no signif-
icant associations with preterm births before 35 or 32
weeks.28,30 The finding of decreased maternal IL-6 with
rSPBs is unexpected, and contrary to current knowl-
edge suggesting a link between increased systemic in-
flammation and subsequent preterm birth. Further
evaluation revealed those with the lowest pregravid
weights and BMIs to have lower IL-6 levels. Although
it is plausible that such women are undernourished and
are less able to mount an inflammatory response to
ascending infection, this speculation requires confir-
mation. It should be noted that the cytokine, cortisol,
and CRH analyses performed on this cohort were
conducted after prolonged freezing at –70(C. As plasma
levels can decline with extended storage, it is plausible
that differences between groups in this analysis were
less apparent. Alternatively, plasma IL-6, CRH, and
cortisol were significantly elevated despite prolonged
storage. We do not anticipate that differences between
groups would be overestimated in this circumstance.
Plasma cortisol levels are subject to significant
1182
Figure 3 Box plots for plasma cortisol and CRH obtained at 22 to 24 weeks’ gestation from 230 multigravid women with: rSPBs
(black), an iSPB in the current gestation (dark grey), an iSPB in a prior gestation (light grey), and rTB only controls (white).
circadian variation. This could obscure more significant
differences, but could also result in a spuriously positive
finding should women with rSPBs have been more com-
monly sampled at the time of peak levels. Although no
systemic collection bias is anticipated, timed sampling
or assessment of urinary-free cortisol would be more
appropriate in future studies.
Preterm birth is multifactorial in nature. In many
cases, the ultimate cause of preterm labor or PROM
leading to SPB is unknown. However, genital tract and
intrauterine infection and inflammation have been
highly linked to preterm birth at the time of delivery,
and preterm birth has been associated with systemic
inflammation in some cases. Preterm birth has also been
associated with abnormal placentation, uterine over-
distension, maternal stress, maternal race and socio-
economic status, work during pregnancy, underlying
medical conditions, fetal abnormalities, maternal to-
bacco exposure, and also maternal undernutrition as
exemplified by maternal weight and BMI. The strongest
clinical marker for preterm birth is a prior history of 1
or more preterm deliveries, particularly if these occurred
early in pregnancy. Approximately, 1 in 4 pregnant
women with a prior delivery will have a history of prior
preterm birth,20 and in this evaluation we have found
those who have repeated preterm births are to be nearly
Mercer et al
twice as likely to have an early preterm birth before
32 or 35 weeks’ gestation.
Finally, it may be that 22 to 24 weeks is too late to
identify characteristics related to the risk of recurrent
preterm birth but unrelated to ongoing changes in a
pregnancy destined to deliver preterm. Although only
1 patient delivered within 2 weeks of assessment, eval-
uation earlier in pregnancy might allow evaluation of
the maternal response to the pregnancy more remote
from delivery, and preconceptual evaluation could help
discern chronic differences unrelated to pregnancy. It
must be reiterated that this is an exploratory, hypothesis
generating secondary analysis, and that the identified
associations warrant further study for confirmation and
to identify potential mechanisms. That being said, the
information presented herein offers direction for further
study toward a better understanding of the biology of
women who have had recurrent preterm births.
Acknowledgments
We recognize Cora MacPherson, PhD, for her assis-
tance in conduct of the statistical analyses related to this
manuscript, Judy Minium, MSc, and Patricia Brandt,
BSc (MetroHealth Medical Center-Case Western Re-
serve University) for the performance of plasma
Mercer et al
cytokine, cortisol, and CRH evaluations. In addition to
the authors, other members of the National Institute of
Child Health and Human Development Maternal-Fetal
Medicine Units Network are as follows:
University of Alabama, Birmingham, AL–J. C. Hauth,
A. Northen, and C. Neely; Wake Forest University, Win-
ston-Salem, NC–E. Mueller-Heubach, M. Swain, and
A. Frye; University of Chicago, Chicago, IL–M. Lind-
heimer, P. Jones, and M. E. Lewis Brown; University
of Cincinnati, Cincinnati, OH–T. A. Siddiqi, N. Elder,
T. Coombs, and J. VanHorn; University of Pittsburgh,
Pittsburgh, PA–J. M. Roberts, J. H. Harger, M. Cotro-
neo, and C. Stallings; Ohio State University, Columbus,
OH–M. B. Landon, J. Schneider, and C. Mueller; Uni-
versity of Oklahoma, Oklahoma City, OK–J. C. Carey,
A. Meier, and E. Liles; Medical University of South Car-
olina, Charleston, SC–M. K. Menard, R. B. Newman, B.
A. Collins, T. Metcalf, and V. Odell; University of Ten-
nessee, Memphis, TN–B. Sibai, R. Ramsey, and J. L.
Fricke; Wayne State University, Detroit, MI–S. F. Bot-
toms (deceased), M. Treadwell, G. S. Norman; Univer-
sity of Utah, Salt Lake City, UT–D. Dudley and L.
Reynolds; The George Washington University Biostatis-
tics Center, Bethesda, MD–E. Thom, A. Das, L. Leuch-
tenburg, and M. Fischer; National Institute of Child
Health and Human Development, Bethesda, MD–D.
McNellis, S. Yaffe, C. Catz, and M. Klebanoff.
References
1. Martin JA, Hamilton BE, Sutton PD, Ventura SJ, Menacker F,
Munson ML. Births: final data for 2002. Natl Vital Stat Rep
2003;52:1-114.
2. Demissie K, Rhoads GG, Ananth CV, Alexander GR, Kramer
MS, Kogan MD, et al. Trends in preterm birth and neonatal mor-
tality among blacks and whites in the United States from 1989 to
1997. Am J Epidemiol 2001;154:307-15.
3. Stevenson DK, Wright LL, Lemons JA, Oh W, Korones SB, Pap-
ile LA, et al. Very low birth weight outcomes of the National
Institute of Child Health and Human Development Neonatal
Research Network, January 1993 through December 1994. Am J
Obstet Gynecol 1998;179:1632-9.
4. Bottoms SF, Paul RH, Mercer BM, MacPherson CA, Roberts JM,
Caritis SN, et al, for the National Institutes of Child Health and
Human Development Network of Maternal Fetal Medicine
Research Units. Obstetrical determinants of neonatal survival: an-
tenatal predictors of neonatal survival and morbidity in extremely
low birth weight infants. Am J Obstet Gynecol 1999;180:665-9.
5. Perinatal Care at the Threshold of Viability. Clinical Management
Guidelines for Obstetrician-Gynecologists. September 2002. Num-
ber 38.
6. Romero R, Sirtori M, Oyarzun E, Avila C, Mazor M, Callahan R,
et al. Infection and labor, V: prevalence, microbiology, and clinical
significance of intraamniotic infection in women with preterm la-
bor and intact membranes. Am J Obstet Gynecol 1989;161:817-24.
7. Salafia CM, Vogel CA, Vintzileos AM, Bantham KF, Pezzullo J,
Silberman L. Placental pathologic findings in preterm birth. Am
J Obstet Gynecol 1991;165:934-8.
8. Goldenberg RL, Hauth JC, Andrews WW. Intrauterine infection
and preterm delivery. N Engl J Med 2000;342:1500-7.
1183
9. Offenbacher S, Lieff S, Boggess KA, Murtha AP, Madianos PN,
Champagne CM, et al. Maternal periodontitis and prematurity:
part I, obstetric outcome of prematurity and growth restriction.
Ann Periodontol 2001;6:164-74.
10. Goepfert AR, Jeffcoat MK, Andrews WW, Faye-Petersen O,
Cliver SP, Goldenberg RL, et al. Periodontal disease and upper
genital tract inflammation in early spontaneous preterm birth.
Obstet Gynecol 2004;104:777-83.
11. Mazor M, Chaim W, Hershkowitz R, Levy J, Leiberman JR,
Glezerman M. Association between preterm birth and increased
maternal plasma cortisol concentrations. Obstet Gynecol 1994;
84:521-4.
12. Copper RL, Goldenberg RL, Das A, Elder N, Swain M, Norman
G, et al. The preterm prediction study: maternal stress is associated
with spontaneous preterm birth at less than thirty-five weeks’ ges-
tation. National Institute of Child Health and Human Develop-
ment Maternal-Fetal Medicine Units Network. Am J Obstet
Gynecol 1996;175:1286-92.
13. Hobel CJ, Dunkel-Schetter C, Roesch SC, Castro LC, Arora CP.
Maternal plasma corticotropin-releasing hormone associated with
stress at 20 weeks’ gestation in pregnancies ending in preterm
delivery. Am J Obstet Gynecol 1999;180:S257-63.
14. Mancuso RA, Schetter CD, Rini CM, Roesch SC, Hobel CJ.
Maternal prenatal anxiety and corticotropin-releasing hormone
associated with timing of delivery. Psychosom Med 2004;66:
762-9.
15. Annells MF, Hart PH, Mullighan CG, Heatley SL, Robinson JS,
Bardy P, et al. Interleukins-1, -4, -6, -10, tumor necrosis factor,
transforming growth factor-beta, FAS, and mannose-binding pro-
tein C gene polymorphisms in Australian women: risk of preterm
birth. Am J Obstet Gynecol 2004;191:2056-67.
16. Hartel Ch, Finas D, Ahrens P, Kattner E, Schaible T, Muller D,
et al, for the Genetic Factors in Neonatology Study Group. Poly-
morphisms of genes involved in innate immunity: association with
preterm delivery. Mol Hum Reprod 2004;10:911-5.
17. Moore S, Ide M, Randhawa M, Walker JJ, Reid JG, Simpson NA.
An investigation into the association among preterm birth, cyto-
kine gene polymorphisms and periodontal disease. BJOG 2004;
111:125-32.
18. Carr-Hill RA, Hall MH. The repetition of spontaneous preterm
labour. BJOG 1985;92:921-8.
19. Bakketeig LS, Hoffman HJ. Epidemiology of preterm birth: results
from a longitudinal study of births in Norway. In: Elder LS,
Hendricks CH, editors. Preterm labor. London: Butterworth;
1981. p. 17-46.
20. Mercer BM, Goldenberg RL, Das A, Moawad AH, Iams JD, Meis
PJ, et al, for the NICHD MFMU Network. The preterm predic-
tion study: a clinical risk assessment system. Am J Obstet and Gy-
necol 1996;174:1885-95.
21. Mercer BM, Goldenberg RL, Meis PJ, Moawad AH, Shellhaas C,
Das A, et al, for the NICHD-MFMU Network. The preterm
prediction study: prediction of preterm premature rupture of the
membranes using clinical findings and ancillary testing. Am J
Obstet Gynecol 2000;183:738-45.
22. Mercer BM, Goldenberg RL, Moawad AH, Meis PJ, Iams JD,
Das AF, et al, for the NICHD-MFM Network. The preterm pre-
diction study: effect of gestational age and cause of preterm birth
on subsequent obstetric outcome. Am J Obstet Gynecol 1999;
181:1216-21.
23. Meis P, Goldenberg R, Iams J, Mercer B, Moawad A, McNellis D,
et al, for the NICHD Maternal Fetal Medicine Unit Network. The
preterm prediction study: significance of vaginal infections. Am J
Obstet Gynecol 1995;173:1231-5.
24. Iams J, Goldenberg R, Meis P, Mercer BM, Moawad A, Das A,
et al, the NICHD MFMU Network. The length of the cervix
and the risk of spontaneous premature delivery. N Engl J Med
1996;334:567-72.
1184
25. Goldenberg RL, Mercer BM, Meis PJ, Copper RL, Das A,
McNellis D. The preterm prediction study: fetal fibronectin testing
and spontaneous preterm birth. NICHD Maternal Fetal Medicine
Units Network. Obstet Gynecol 1996;87:643-8.
26. Goldenberg R, Iams J, Mercer B, Meis P, Moawad A, Copper R,
et al, for the NICHD Maternal Fetal Medicine Units Network.
The preterm prediction study: the value of new versus standard
risk factors in predicting early and all spontaneous preterm birth.
Am J Public Health 1998;88:233-8.
27. Goldenberg RL, Iams JD, Mercer BM, Meis PJ, Moawad A, Das
A, et al. The preterm prediction study: toward a multiple-marker
test for spontaneous preterm birth. Am J Obstet Gynecol 2001;
185:643-51.
28. Moawad AH, Goldenberg RL, Mercer B, Meis PJ, Iams JD, Das
A, et al. The preterm prediction study: the value of serum alkaline
phosphatase, alpha-fetoprotein, plasma corticotropin-releasing
hormone, and other serum markers for the prediction of spontane-
ous preterm birth. Am J Obstet Gynecol 2002;186:990-6.
29. Goepfert AR, Goldenberg RL, Andrews WW, Hauth JC, Mercer
B, Iams J, et al. The preterm prediction study: association between
cervical interleukin 6 concentration and spontaneous preterm
birth. Am J Obstet Gynecol 2001;184:483-8.
30. Goldenberg RL, Iams JD, Mercer BM, Meis P, Moawad A,
Das A, et al, for the National Institute of Child Health and Human
Development Maternal-Fetal Medicine Units Network. What we
have learned about the predictors of preterm birth. Semin Perina-
tol 2003;27:185-93.
Discussion
GEORGE MACONES, MD. I would like to thank the Soci-
ety for inviting me to discuss this article and congratu-
late Dr Mercer for this research.
Spontaneous preterm birth continues to be a critical
issue in Perinatal medicine, contributing significantly to
neonatal morbidity and mortality in the United States
and worldwide. Much research on the identification of
risk factors for spontaneous preterm birth has been
performed, with much of it being led by the NICHD-
sponsored Maternal-Fetal Medicine Units Network,
especially with the Preterm Prediction Study. Over the
past 10 to 15 years, a major focus has been on the
association between a short cervix and preterm birth.
More recently, attention has turned to the association
between subclinical infection, inflammation, and pre-
term birth, and whether there may be underlying genetic
susceptibilities for its occurrence.
Despite this research, the incidence of spontaneous
preterm birth is largely unchanged. Many interventions
to prevent or reduce preterm birth have been tested, and
to date, the only one that holds promise is treatment
with progesterone, though even this is still controversial.
Dr Mercer, today, presented results from a thought-
ful secondary analysis of data collected as part of the
Preterm Prediction study, an observational study per-
formed in the early 1990s. The Preterm Prediction study
has given us great insights into the relationship between
cervicovaginal fetal fibronectin, cervical length, uterine
contractions, maternal infections, and spontaneous pre-
term birth. The overall goal of the study presented today
Mercer et al
was to assess whether those with recurrent spontaneous
preterm births have different profiles of inflammatory
markers, clinical characteristics, and cervical character-
istics compared with those with a single spontaneous
preterm birth or those with recurrent term deliveries.
Those with recurrent preterm births are an especially
interesting group to study, and though still limited by a
fairly small sample size, represents one of the larger
studies to date on this group of women.
I do have several questions for Dr Mercer.
First, how was history of term versus spontaneous
preterm birth assessed? Was this based on patient self-
report? If so, are there data on the reliability of patient
self-report for spontaneous versus indicated preterm
birth?
Second, can Dr Mercer discuss the variation in serum
cytokine levels within individuals? How useful is a single
measurement in serum?
Third, in the 2 nested case-control studies, the chosen
control: case ration is 2:1. Can Dr Mercer comment on
this choice?
Fourth, from an analytical perspective, it seems that
only bivariate statistics were used. Was there any con-
sideration given to a series of multivariable models that
would allow for adjustment of covariates?
Fifth, is there any evidence in the data that those with
recurrent spontaneous preterm births were subjected to
more screening/interventions as a result of their history
compared with the other groups?
DR MERCER (Closing statement). I thank Dr Macones
for his detailed review and discussion of this article.
Our impression was that the patient’s recollection of in-
dication for prior preterm birth would not be reliable.
When a history of prior preterm birth was obtained,
we requested permission to retrieve the prior medical
records and made a determination of the gestational
age and indication for that delivery by chart review,
where possible. These women were not specifically asked
whether they had screening for or prophylactic interven-
tions to prevent preterm delivery. However, this study
was conducted before fetal fibronectin and sonographic
cervical length screening were commonly used. Four
percent of women with recurrent preterm birth did re-
ceive tocolytic therapies by 22 to 24 weeks’ gestation.
Though these women also had more frequent contrac-
tions, it is plausible that there was a higher index of sus-
picion, and that interventions were more liberally used.
Had tocolytic therapy been more liberal based on prior
history alone, however, we would have also expected
more frequent treatment among those with a prior
iSPB. Ultimately, those with recurrent preterm births
did deliver preterm in the current pregnancy. It is un-
likely that screening or treatments significantly affected
the parameters we evaluated in this group.