Vous êtes sur la page 1sur 6

The Preterm Labor Index and Fetal Fibronectin for

Prediction of Preterm Delivery With Intact


Membranes
Masatoshi Sakai, MD, Yasushi Sasaki, MD, Naoko Yamagishi, MD, Kyoko Tanebe, MD,
Satoshi Yoneda, MD, and Shigeru Saito, MD

OBJECTIVE: To compare the preterm labor index with the tered in caring for women with findings suggesting pre-
biochemical marker fetal fibronectin in predicting preterm term labor is predicting risk of preterm delivery. Preterm
delivery. labor has been defined in terms of clinical manifestations
METHODS: In 185 women with preterm labor and intact such as regular contractions, cervical dilatation, cervical
membranes, the preterm labor index was determined effacement, and genital bleeding. However, even in sim-
based on clinical findings of uterine contractions, bleeding, ple combinations (such as regular contraction and cervi-
and cervical dilatation. An immunosorbent assay was used
cal change) these are not adequate criteria for diagnosis
to measure oncofetal fibronectin.
of preterm labor (false-positive rate for preterm delivery,
RESULTS: A significant negative correlation was noted be- 20 – 40%4). The main purpose of predictors of preterm
tween the preterm labor index and the interval from hos- delivery is to determine when one can safely withhold
pitalization to delivery. Based on a receiver operating char-
interventions such as corticosteroids and tocolysis and
acteristic curve, a preterm labor index 4 or higher was
discharge the patient home. In 1973,5 Baumgarten and
defined as positive. Rates of preterm delivery overall and
of delivery within 1 week in the preterm labor index– Gruber proposed a tocolysis index for evaluating likeli-
negative group were 55.2% and 13.8%, respectively. Those hood of preterm delivery, combining the four clinical
in the preterm labor index–positive group were signifi- factors of uterine contraction, premature rupture of
cantly higher (88.4% and 56.5%). With respect to predicting membranes (PROM), genital bleeding, and cervical dila-
when a delivery would occur during the first week of tation. Although the tocolysis index correlated nega-
hospitalization, the preterm labor index and fetal fibronec- tively with the interval from the start of treatment to
tin did not differ in sensitivity (77.7% and 73.8%, respective- delivery,6 this index has not attained wide use as a
ly), specificity (79.5% and 74.2%), positive predictive value predictive marker of preterm delivery. Nonetheless, the
(66.7% and 60.8%), or negative predictive value (84.5% and tocolysis index is a good marker for predicting preterm
83.9%). However, within the preterm labor index-negative
delivery in premature labor cases with and without
group, the rates of preterm delivery (68.4%), delivery
PROM. Premature rupture of membranes is a useful
within 1 week (28.9%), and neonatal treatment in an inten-
sive care unit (36.8%) were significantly higher in fetal marker in itself; when it was present, 75% of patients
fibronectin-positive than in fibronectin-negative cases already were in labor at admission, and another 10%
(46.2%, 9.0%, and 10.3%, respectively). delivered after spontaneous labor within 48 hours.7 Ad-
CONCLUSION: The preterm labor index was similar to the
ditional predictive markers for preterm delivery are re-
fetal fibronectin assay in its ability to predict preterm quired in cases with intact membranes. We used a mod-
delivery. (Obstet Gynecol 2003;101:123– 8. © 2003 by ified version of the tocolysis index, called the preterm
The American College of Obstetricians and Gynecolo- labor index (excluding rupture of membranes and defin-
gists.) ing each parameter precisely), to evaluate the likelihood
of preterm delivery in patients without PROM. Lock-
wood et al8 reported that oncofetal fibronectin content in
Despite significant advances in perinatal medicine, pre-
cervicovaginal secretions reliably predicted preterm de-
mature delivery remains a major cause of neonatal mor-
bidity and mortality.1–3 The primary problem encoun- livery in cases where membranes were intact, and subse-
quent reports have been confirmatory.9,10 The purpose
From the Department of Obstetrics and Gynecology, Toyama Medical and Phar- of this study was to evaluate the predictive value of the
maceutical University, Toyama, Japan. preterm labor index in comparison with fetal fibronectin

VOL. 101, NO. 1, JANUARY 2003 0029-7844/03/$30.00 123


© 2003 by The American College of Obstetricians and Gynecologists. Published by Elsevier Science Inc. PII S0029-7844(02)02463-8
Table 1. Scoring of Obstetric Status by the Preterm Labor Index*
No. of points
Signs 0 1 2 3 4
Uterine contractions None ⬍6/h ⱖ6/h — —
Bleeding (g) None ⬍10 ⱖ10 — —
Cervical dilatation (cm) None 1 to ⬍2 2 to ⬍3 3 to ⬍4 ⱖ4
* The score is computed by addition of the number of points for each sign.

determinations in women in preterm labor with intact Fetal fibronectin was determined quantitatively in
membranes. vaginal fluid with an enzyme-linked immunosorbent as-
say (fetal fibronectin enzyme immunoassay; Adeza Bio-
medical, Sunnyvale, CA). Specimens were obtained as
MATERIALS AND METHODS
sterile Dacron high vaginal swabs from the posterior
The study included 185 pregnant women brought to our fornix upon admission, before any digital examination.
hospital between 1997 and 2001 because of preterm A fetal fibronectin exceeding 50 ng/mL was defined as
labor. Women with preterm PROM were excluded from fibronectin positive, whereas fibronectin below 50
this study. Preterm labor was defined according to the ng/mL was fibronectin negative, as reported by Lock-
Canadian Preterm Labor Investigators Group6 as the wood et al.8
presence of regular uterine contractions (six per 60 min- Results are presented as the mean ⫾ standard devia-
utes documented by external tocography) or any uterine tion. Patient demographic data representing continuous
activity associated with a cervix effaced by at least 50% variables were analyzed by the Student t test; a ␹2 test
or dilated by 2 cm or more. Patients with multiple was used for discrete variables. Correlations were calcu-
pregnancies and those with early deliveries performed lated using simple regression and the Spearman rank
therapeutically because of fetal asphyxia, preeclampsia, correlation coefficient test. We used multivariate logistic
placenta previa, abruptio placentae, or such maternal regression analysis to evaluate the association between
medical complications as diabetes mellitus, hyperthy- cervicovaginal fetal fibronectin status and outcome of
roidism, or asthma were excluded from the study. pregnancy. Differences associated with P values less than
To treat preterm labor, women with regular uterine .05 were regarded as significant. The sample size was
contractions as defined above were treated with an intra- predetermined using a power analysis. We calculated
venous infusion of the ␤-2 stimulant ritodrine hydro- that 180 patients would be required to demonstrate a
chloride (33 ␮g per minute). When administration of significant association between fetal fibronectin and out-
ritodrine hydrochloride at the maximum dose (100 ␮g come with a positive predictive value of 50% and a
per minute) was not effective, intravenous administra- negative predictive value of 80%.
tion of magnesium sulfate was added (4 g per 30 minute),
and then continued at 1 to 2 g per hour.
At admission, three clinical factors (uterine contrac- RESULTS
tion, genital bleeding, and cervical dilatation) were eval- Gestational age on admission was 28.6 ⫾ 4.0 weeks.
uated to determine the preterm labor index for all pa- Total numbers and percentages of patients who had
tients (Table 1). All initial assessments were performed preterm delivery before 37 weeks and delivery within 1
by skilled obstetricians. week after admission were 125 (67.6%) and 55 (29.7%),

Table 2. Characteristics of the Study Population


Preterm labor index Fetal fibronectin*
ⱕ3 ⱖ4 Positive Negative
(n ⫽ 116) (n ⫽ 69) (n ⫽ 89) (n ⫽ 96)
Maternal age (y) 25.5 ⫾ 5.2 25.3 ⫾ 5.8 25.4 ⫾ 5.7 25.3 ⫾ 5.2
Primiparous (%) 47.6 46.9 46.6 47.1
Previous preterm delivery (%) 9.1 8.7 9.3 9.0
Education ⬍12 y (%) 1.2 1.0 1.1 1.1
Smoker (%) 2.6 4.3 2.2 3.1
Gestational age at hospitalization (wk) 29.3 ⫾ 8.7 30.5 ⫾ 9.2 29.6 ⫾ 8.9 28.9 ⫾ 8.2
Plus-minus values are mean ⫾ standard deviation.
* Fetal fibronectin positivity was defined as fetal fibronectin concentrations in cervicovaginal secretions of 50 ng /mL or more.

124 Sakai et al Predictors of Preterm Delivery OBSTETRICS & GYNECOLOGY


Figure 1. Correlation between the preterm labor index on
Figure 3. Receiver operating characteristic curves of the
admission and interval from hospitalization to delivery (n ⫽
preterm labor index for prediction of preterm delivery (open
185).
circles) and delivery within 1 wk (filled circles). A cutoff
Sakai. Predictors of Preterm Delivery. Obstet Gynecol 2003.
value of 4 points or more best minimized both false-
positive and false-negative results.
respectively. Maternal characteristics in this study are Sakai. Predictors of Preterm Delivery. Obstet Gynecol 2003.
shown in Table 2. All patients in this study were mar-
ried.
A significant negative correlation (r ⫽ ⫺.689, P ⬍ We prepared receiver operating characteristic (ROC)
.001) was noted between preterm labor index and the curves to evaluate optimal cutoff points for selecting
interval from hospitalization to delivery (Figure 1). A preterm labor index scores for predicting which women
significant positive correlation (r ⫽ .973, P ⬍ .001) was were likely to undergo either preterm delivery overall or
noted between the preterm labor index score and rate of delivery within 1 week (Figure 3). A threshold preterm
delivery within 1 week (Figure 2). labor index of 4 points appeared clinically useful for
predicting either event, so scores of 4 or more were
considered positive and scores of 3 or less were consid-
ered negative. Characteristics of the study population
did not differ significantly between the preterm labor
index–positive and preterm labor index–negative groups
(Table 2). Rates of preterm delivery before 37 weeks and
of delivery within 1 week in the preterm labor index–
negative group (n ⫽ 116) were 55.2% and 13.8%, respec-
tively; those in the preterm labor index–positive group (n
⫽ 69) were significantly higher (88.4% and 56.5%, re-
spectively; P ⬍ .001 and P ⬍ .001 by the ␹2 test).
A significant negative correlation (r ⫽ ⫺.391, P ⬍
.001) was evident between fetal fibronectin and the inter-
val from hospitalization to delivery (Figure 4). As for
ability to predict when delivery would occur during the
first week of hospitalization, preterm labor index scores
and fibronectin did not differ significantly in sensitivity,
specificity, positive predictive value, or negative predic-
Figure 2. Correlation between the preterm labor index on tive value (Table 3). Using the two tests together showed
admission and the rate of delivery within 1 week (n ⫽ 185). no improvement in predictive ability for delivery within
Sakai. Predictors of Preterm Delivery. Obstet Gynecol 2003. 1 week (Table 3).

VOL. 101, NO. 1, JANUARY 2003 Sakai et al Predictors of Preterm Delivery 125
ture of membranes, genital bleeding, and cervical dilata-
tion), was reported in 1973 by Baumgarten and Gruber.5
Evaluating the tocolysis index with 202 preterm labor
patients treated with ritodrine, they reported that when
the tocolysis index was 3 or less, the rate of preterm
delivery was only 10%, whereas a tocolysis index of 4 or
more was associated with a rate of preterm delivery of
85%.5 Richter6 evaluated the correlation between the
tocolysis index and the period from initiation of treat-
ment to delivery in 125 patients believed to have preterm
labor (including those with intact membranes and those
with PROM), finding that the tocolysis index had a
significant negative correlation (r ⫽ ⫺.380, P ⬍ .001)
with this interval. However, because PROM is associ-
ated with greater chances of progression in labor,11 we
compared the modified tocolysis index (the preterm
labor index) and fetal fibronectin as predictors for pre-
Figure 4. Correlation between fetal fibronectin in cervico- term delivery in patients without PROM. Patients who
vaginal secretions and the interval from hospitalization to present without evidence of ruptured membranes but
delivery (n ⫽ 185). show regular uterine activity or genital bleeding pose the
Sakai. Predictors of Preterm Delivery. Obstet Gynecol 2003.
most uncertainty as to likelihood of preterm delivery. In
Although preterm labor index–negative patients this study we evaluated correlations between the pre-
showed good overall likelihood of avoiding preterm term labor index and overall rate of preterm delivery, as
labor, some patients still delivered preterm even within 1 well as the rate of delivery within 1 week of hospitaliza-
week after admission. We therefore evaluated whether tion by analysis of ROC curves. By ROC analysis, a
the fetal fibronectin test could improve predictive effec- cutoff preterm labor index value, a score of 4 or greater,
tiveness in patients with negative preterm labor index was established as a positive predictor for preterm deliv-
scores. Preterm labor index–negative patients were di- ery in cases with intact membranes. Baumgarten’s re-
vided into a fibronectin-positive group and a fibronectin- port, which included cases with and without PROM,
negative group. No differences were found between determined the same tocolysis index cutoff, 4, as our
these two subgroups in maternal background factors study excluding PROM cases. The preterm labor index,
such as age, parity, gestational age, history of preterm therefore, was a good predictor of preterm delivery in
birth, or history of smoking (Table 2). Rates of preterm cases with intact membranes, as well as delivery within 1
delivery, delivery within 1 week, low birth weight, and week of admission. Early identification of preterm labor
need of treatment in the neonatal intensive care unit cases can help in prearranging neonatal intensive care
were significantly higher in the fetal fibronectin–positive and deciding whether to administer corticosteroids and
subgroup than in the fetal fibronectin–negative subgroup tocolytic agents.
(Table 4). Oncofetal fibronectin in cervicovaginal secretions has
been reported to be a good predictor of preterm delivery
in preterm labor cases with intact membranes. Lock-
DISCUSSION wood et al8 investigated the efficacy of fetal fibronectin in
The tocolysis index, incorporating four clinical factors cervical and vaginal secretions to predict preterm deliv-
for predicting preterm delivery (uterine contraction, rup- ery in 117 patients with preterm labor, and reported

Table 3. Sensitivity, Specificity, and Predictive Values of the Preterm Labor Index and Fetal Fibronectin for Delivery Within
1 Week
Sensitivity (%) Specificity (%) PPV (%) NPV (%)
PLI (%) 77.7 79.5 66.7 84.5
fFN 73.8 74.2 60.8 83.9
PLI or fFN 57.7 89.2 73.2 80.5
PLI and fFN 90.4 57.8 52.2 92.2
PPV ⫽ positive predictive value; NPV ⫽ negative predictive value; PLI ⫽ preterm labor index; fFN ⫽ fetal fibronectin.
PLI ⱖ 4 and fFN ⱖ 50 ng/mL were defined as positive.

126 Sakai et al Predictors of Preterm Delivery OBSTETRICS & GYNECOLOGY


Table 4. Outcomes of Pregnancy in Patients With a Preterm Labor Index Score of 3 or Less According to Fetal Fibronectin
fFN ⱖ 50 ng/mL fFN ⬍ 50 ng/mL
Variable (n ⫽ 38) (n ⫽ 78) P
Gestational age (wk) 28.2 ⫾ 6.0 29.1 ⫾ 5.3 .237
PLI score 1.78 ⫾ 1.0 1.60 ⫾ 1.2 .201
Preterm delivery (%) 26 (68.4) 36 (46.2) .024
Delivery within 7 d (%) 11 (28.9) 7 (9.0) .010
Prolongation period (d) 39.9 ⫾ 38.2 41.4 ⫾ 28.7 .677
Birth weight (g) 2350 ⫾ 846 2692 ⫾ 561 .024
Neonates requiring intensive care (%) 14 (36.8) 8 (10.3) .001
fFN ⫽ fetal fibronectin; PLI ⫽ preterm labor index.
Preterm delivery was defined as delivery at less than 37 weeks’ gestation. Prolongation periods were defined as intervals between admission and
delivery. Data are shown as the mean ⫾ standard deviation.

sensitivity, specificity, positive predictive value, and neg- change and with negative fibronectin results were less
ative predictive value of 81.7%, 82.5%, 83.1%, and likely to deliver preterm. The preterm labor index in
81.0%, respectively. Many other studies of the predictive these patients without cervical change was graded as 2– 4
value of fetal fibronectin reported good sensitivity (44 – points (uterine contraction, 2 points; spotting, 1 point;
90%) and specificity (72–96%) for predicting preterm marked genital bleeding, 2 points). Our data support
delivery in symptomatic patients.9 –15 In our study, sen- Rinehart’s findings. When the preterm labor index was 4
sitivity and specificity for preterm delivery within 1 week or greater, preterm delivery was likely irrespective of
of fetal fibronectin were 73.8% and 74.2%, respectively, fetal fibronectin. Measurement of fetal fibronectin offers
somewhat lower than in a recent meta-analysis (sensitiv- little additional benefit in such high-risk cases.
ity 89%, specificity 86%).9 Although fibronectin in cervi- Although many studies have evaluated many poten-
covaginal secretions is undoubtedly a good predictor of tial biologic markers for preterm birth,17,18 the preterm
preterm delivery, the assay is relatively expensive (about labor index should be used for patients with preterm
$219). labor without PROM because it is a good predictor of
We evaluated the predictive value of the preterm preterm delivery overall as well as delivery within 1
labor index and compared the predictive ability of the week. In patients with the preterm labor index of 3 or
preterm labor index with that of fibronectin. Sensitivity lower, fetal fibronectin should be measured when possi-
and specificity for preterm delivery by preterm labor ble.
index were 71.7% and 81.2%, respectively, without sig-
nificant differences from those shown by fetal fibronectin
in preterm labor cases with intact membranes. There- REFERENCES
fore, our study demonstrated that the preterm labor 1. McCormick MC. The contribution of low birth weight to
index and fetal fibronectin have a similar ability to infant mortality and childhood morbidity. N Engl J Med
distinguish between patients with high and low risk of 1985;312:82–90.
preterm delivery. 2. Gibbs RS, Romero R, Hillier SL, Eschenbach DA, Sweet
Our study suggests that the preterm labor index is a RA. A review of premature birth and clinical infection.
good predictor of both preterm delivery and delivery Am J Obstet Gynecol 1992;166:1515–28.
within 1 week for preterm labor patients without 3. Iams JD. Current status of prematurity prevention. JAMA
PROM. The preterm labor index is a very simple 1989;262:265–6.
method that can predict whether apparent labor will 4. King JF, Grant A, Keirse MJ, Chalmers I. Beta-mimetics in
result in preterm delivery. We therefore recommend the preterm labour: An overview of the randomized con-
preterm labor index as a rapid and cost-effective predic- trolled trials. Br J Obstet Gynaecol 1988;95:211–22.
tor. Whenever resources are available, fetal fibronectin 5. Baumgarten K, Gruber W. Tocolyseindex. In: Duden-
should be measured in patients whose preterm labor hausen JW, Saling E, eds. Perinatale Medizin. Stuttgart,
index is 3 or less because fibronectin-positive patients Germany: Georg Thieme Verlag, 1974:197–9.
whose preterm labor index is 3 or less still are considered 6. Richter R. Evaluation of success in treatment of threaten-
at high risk for preterm delivery. When both the preterm ing premature labor by beta mimetic drugs. Am J Obstet
labor index and fibronectin are negative, one might Gynecol 1977;127:482–6.
withhold corticosteroids and tocolysis and discharge the 7. Cunningham FG, MacDonald PC, Gant NF, Leveno KJ,
patient home. Recently, Rinehart et al16 reported that Gilstrap LC III. Williams obstetrics. 19th ed. Norwalk,
patients with findings of preterm labor including cervical Connecticut: Appleton & Lange, 1993:862.

VOL. 101, NO. 1, JANUARY 2003 Sakai et al Predictors of Preterm Delivery 127
8. Lockwood CJ, Senyei AE, Dische MR. Fetal fibronectin in by fetal fibronectin and uterine activity. Am J Obstet
cervical and vaginal secretions defines a patient population Gynecol 1996;87:649–55.
at high risk for preterm delivery. N Engl J Med 1991;325: 15. Malak TM, Sizmur F, Bell SC, Tayler DJ. Fetal fibronectin
669–74. in cervicovaginal secretions as a predicter of preterm birth.
9. Leitich H, Egarter C, Kaider A, Hohlagschwandtner M, Br J Obstet Gynaecol 1996;103:648–53.
Berghammer P, Husslein P. Cervicovaginal fetal fibronec- 16. Rinehart BK, Terrone DA, Isler CM, Barrilleaux PS,
tin as a marker for preterm delivery: A meta-analysis. Am J
Bufkin L, Morrison JC. Pregnancy outcome in women
Obstet Gynecol 1999;180:1169–76.
with preterm labor symptoms without cervical change.
10. Goepfert AR, Goldenberg RL, Mercer B, Iams J, Meis P, Am J Obstet Gynecol 2001;184:1004–7.
Moawad A, et al. The preterm prediction study: Quanti-
17. Mauldin JG, Newman RB. Preterm birth risk assessment.
tative fetal fibronectin values and the prediction of sponta-
neous preterm birth. Am J Obstet Gynecol 2000;183: Semin Perinatol 2001;25:215–22.
1480–3. 18. Goldenberg RL, Iams JD, Mercer BM, Meis PJ, Moawad
11. Nelson LH, Anderson RL, O’Shea TM, Swain M. Expect- A, Das A, et al. The Preterm Prediction Study: Toward a
ant management of preterm premature rupture of the multiple-marker test for spontaneous preterm birth. Am J
membranes. Am J Obstet Gynecol 1994;17:350–8. Obstet Gynecol 2001;185:643–51.
12. Irion O, Matute J, Bischof P. Prediction de la prematurite
par la fibronectine oncofoetale. J Gynecol Obstet Biol Address reprint requests to: Shigeru Saito, MD, Toyama Med-
Reprod (Paris) 1995;24:624–9. ical and Pharmaceutical University, Department of Obstetrics
13. Burrus DR, Ernest JM, Veille JC. Fetal fibronectin, inter- and Gynecology, Toyama-Sugitani 2630, Toyama 930-0194,
leukin-6, and C-reactive protein are useful in establishing Japan; E-mail: s30saito@ms.toyama-mpu.ac.jp.
prognostic subcategories of idopathic preterm labor. Am J
Obstet Gynecol 1995;173:1258–62.
14. Morrison JC, Naef RW, Botti JJ, Katz M, Belluomini JM, Received March 4, 2002. Received in revised form July 30, 2002.
McLaughlin BN. Prediction of spontaneous preterm birth Accepted August 8, 2002.

128 Sakai et al Predictors of Preterm Delivery OBSTETRICS & GYNECOLOGY

Vous aimerez peut-être aussi