OBSTETRICS

Contingent Use of Fetal Fibronectin Testing andCervical Length Measurement in Women WithPreterm LabourFranois Audibert, MD, MSc, Suzanne Fortin, MD, Edgard Delvin, PhD, Anissa Djemli, PhD,Suzanne Brunet, RT, Johanne Dub, MD, FRCOG, William D. Fraser, MD, MScDpartement dObsttrique Gyncologie, Centre Hospitalier Universitaire Sainte-Justine, Universit de Montral, Montral (Quebec)

Abstract

Objective: To evaluate the contingent use of fetal fibronectin (fFN)testing and cervical length (CL) measurement to predict pretermdelivery, and to validate the use of phosphorylated IGFBP-1 as apredictor of preterm delivery.

Methods: We recruited 71 women with a clinical diagnosis ofpreterm labour between 24 and 34 weeks, and tested for thepresence of fFN and IGFBP-1 in the cervicovaginal secretions ofall women immediately before CL measurement.

Results: Among the 66 women with complete outcome, four wereexcluded from the final analysis as two had assessment for fFNbut no CL measurement, and another two had CL measured butno screening for fFN. Among 62 women with complete results, themean gestational age at recruitment was 29.4 2.5 weeks. Sixwomen (9.6%) delivered within two weeks of assessment, and 14(22.5%) delivered before 34 weeks. A positive fFN test resulted ina sensitivity of 83%, a specificity of 84%, a positive predictivevalue of 36%, and a negative predictive value of 98% for deliverywithin two weeks; for CL < 25 mm, these figures were 50%, 52%,10%, and 91%, respectively, and for a positive IGFBP-1, they were17%, 93%, 20%, and 91%, respectively. A policy of contingent useof fFN (in which the test was assumed to be positive if CL 15 mm,and fFN was only measured if the CL was between 16 and 30 mm)gave sensitivity, specificity, positive and negative predictive valuesof 80%, 61%, 17%, and 97%, respectively for delivery within twoweeks. Using this contingent use protocol, only one third ofwomen needed fFN screening after CL measurement.

Conclusion: In this study, IGFBP-1 screening did not predict pretermdelivery and fFN screening provided the best predictive capacity.A policy of contingent use of testing for fFN after CL measurement,or contingent use of CL measurement after fFN screening(depending on available resources) is a promising approach tolimit use of resources.

Rsum

Objectif : valuer lutilisation contingente du dpistage de lafibronectine ftale (FNf) et de la mesure de la longueur cervicale(LC) pour prdire laccouchement prterme, ainsi que validerlutilisation de lIGFBP-1 phosphoryle titre de facteur prdictifde laccouchement prterme.

Mthodes : Nous avons recrut 71 femmes ayant obtenu undiagnostic clinique de travail prterme entre 24 et 34 semaines,et nous avons cherch dterminer la prsence de FNf etdIGFBP-1 dans les scrtions cervicovaginales de toutes lesfemmes, immdiatement avant la mesure de la LC.

Rsultats : Parmi les 66 femmes ayant connu une issue complte,quatre ont t exclues de lanalyse finale puisque deux dentreelles avaient subi un dpistage de la FNf sans quune mesure dela LC ne soit effectue, tandis que la LC des deux autres avait tmesure sans quun dpistage de la FNf ne soit men. Chez les62 femmes prsentant des rsultats complets, lge gestationnelmoyen au moment de ladmission ltude tait de 29,4 2,5 semaines. Six femmes (9,6 %) ont accouch dans les deuxsemaines suivant lvaluation et 14 (22,5 %) ont accouch avantla 34e semaine de gestation. Un rsultat positif au test FNf donnaitlieu une sensibilit de 83 %, une spcificit de 84 %, uncoefficient de prvision dun test positif de 36 % et un coefficientde prvision dun test ngatif de 98 % pour ce qui est dunaccouchement dans les deux semaines; en ce qui concerne uneLC

INTRODUCTION

Preterm birth remains a major cause of perinatal morbid-ity and mortality,1 and its rate has not declined over thelast two decades despite the improvement in perinatal man-agement.2 Assessing the probability of preterm delivery isstill a clinical challenge and is important because standardclinical interventions (tocolysis, corticosteroid administra-tion, and transfer to a tertiary care facility) are potentiallyrisky and expensive. Previous studies have shown that adiagnosis of preterm labour based on digital examinationwas less reliable than a diagnosis based on objective tests,such as detection of fetal fibronectin in cervicovaginalsecretions and ultrasound measurement of cervical length.3

Fetal fibronectin, an extracellular matrix glycoprotein local-ized at the maternalfetal interface of the amniotic mem-branes between the chorion and the decidua, is found atvery low levels in cervicovaginal secretions under normalconditions. Levels 50 ng/mL at or after 22 weeks gesta-tion have been associated with an increased risk of sponta-neous preterm birth.36 A recent meta-analysis has shownthat birth before 37 weeks was significantly decreased inpatients whose management was based on knowledge offFN results compared with controls whose fFN resultswere not known.7

Transvaginal CL measurement is the other validated test topredict preterm birth in women with threatened pretermlabour as well as in asymptomatic high-risk and low-riskwomen.3,814 A CL measurement of 25 mm or less is gener-ally considered an excellent indicator of an increased risk ofpreterm delivery, particularly among women with pretermlabour.

Several studies have reported that fFN screening and CLmeasurement provided similar results in predicting the riskof preterm delivery.1520 However, whether combined fFNand CL measurement improves the prediction of pretermdelivery and how the tests should be combined remainunclear.3 Availability of one of the two tests may be an issuein some facilities because the ultrasound expertise for CLmeasurement may not be always available in small centres.On the other hand, the additional cost of fFN testing maybe difficult to justify in centres where CL measurement isreadily available. Therefore, the use of CL or fFN as thefirst-line test might be a more rational option, limiting the

use of a second test to selected cases in a contingentapproach.

More recently, a rapid test for the determination of the insu-lin-like growth factor binding protein phosphorylatedisoform in endocervical secretions has been proposed.2124

Although IGFBP-1 is synthesized in the decidua and theliver, decidual cells predominantly secrete thephosphorylated form, which in normal conditions is notpresent in amniotic fluid and cervical secretions.

The objective of the present study was to determine the per-formance of a contingent use of fFN testing and ultrasoundCL measurement for the prediction of preterm delivery inpatients with preterm labour. A secondary objective was tovalidate the use of cervical IGFBP-1 measurement for theprediction of preterm delivery.

MATERIALS AND METHODS

Women admitted to our tertiary care unit with a clinicaldiagnosis of preterm labour and intact membranes between24 and 34 weeks were approached to participate in the studyand were included after providing informed consent.Preterm labour was defined by the presence of regular uter-ine contractions, lasting at least 30 seconds and occurring atleast four times per 30 minutes, and significant cervicalchanges on digital examination. Women were excluded ifthey had confirmed or suspected rupture of membranes,cervical dilatation > 3 cm, cervical cerclage, vaginal bleed-ing, placenta previa, placental abruption, severe intrauterinegrowth restriction, preeclampsia, or medically indicatedpreterm delivery before 34 weeks.

The study investigations were carried out either on admis-sion or within 24 hours of admission if a digital examinationhad been performed in the 24 hours before the patientsinclusion in the study. Each subject was first examined witha vaginal speculum. A Dacron swab was rotated in the pos-terior fornix of the vagina and sent to the laboratory. Thepresence or absence of fFN was measured by a qualitativetest (Full Term, Hologic, Marlborough, MA), and resultswere expressed as positive or negative. A second swab(Actim Partus, Somagen, Edmonton AB) was taken in thecervix and held for 15 seconds, then dipped into a sterilemedium and held for another 10 seconds. Following this, adipstick was used to determine if the test was positive (twoblue lines), suggesting a concentration of IGFBP-1 in thecervical secretions higher than 10 mg/L, or negative (singleblue line after 5 minutes). Immediately after the sampling, atransvaginal sonographic measurement of the cervicallength was performed by a trained sonographer, or by a resi-dent under direct supervision of a faculty member, using astandard protocol (empty bladder, minimal pressure,measurement of the maximum length between the internal

OBSTETRICS

308 APRIL JOGC AVRIL 2010

ABBREVIATIONS

CL cervical lengthfFN fetal fibronectinIGFBP-1 insulin-like growth factor binding protein-1

and external os, before and after Valsalva manoeuvre).25,26

The clinical team were not blinded to the results of fFNtesting and CL measurement, and these results were avail-able in the medical record. The results of IGFBP-1 testingwere not disclosed to the clinician in charge and were notreported in the medical record. The standard local manage-ment protocol for preterm labour was applied to the studypatients: administration of corticosteroids, administrationof tocolytics for a maximum of 48 hours, and bed rest wereprescribed by the attending physician, depending on clinicalevaluation and on the results of investigations, includingfFN and CL measurement. The reproducibility of CL mea-surement or fFN assessment was not tested in this study.

The outcome of the pregnancy was recorded in a databasewith other information pertinent to the project. The out-comes of interest were delivery within two weeks of admis-sion to the study and delivery before 34 weeks.

Predictive values and likelihood ratios with their 95% confi-dence interval were first calculated for each test consideredseparately, then for different combinations of both markersused in a contingent manner. The contingent use of fFNtest was considered positive if cervical length was 15 mmor between 16 and 30 mm with positive fFN, as describedby Schmitz et al.20 The test was considered negative if cervicallength was > 30 mm or between 16 and 30 mm with negativefFN (Figure 1). The contingent use of CL test was con-sidered positive if fFN was positive, or if fFN was negativewith a CL 15 mm. The test was considered negative if fFNwas negative and CL > 15 mm (Figure 2). Predictive valueswere compared using the McNemar chi-square test. Allanalyses were performed with Stata 10.0 software(StataCorp LP, College Station, TX).

The Ethics Committee of Centre Hospitalier UniversitaireSainte-Justine approved this prospective study.

RESULTS

Seventy-one women were recruited for the study betweenJanuary 2006 and January 2007. The outcome of pregnancycould not be determined for five women who had been dis-charged and delivered in another centre. Among the 66remaining women, two had a fFN assessment but no CLmeasurement, and another two had CL measured but noevaluation of fFN. These cases were excluded from furtheranalysis. Sixty-two women were included in the final analy-sis, including seven twin pregnancies, and 37 (60%) womenwere initially transferred from another centre because ofpreterm labour. Six women (9.6%) delivered within twoweeks, and 14 women (22.5%) delivered before 34 weeks.General characteristics and pregnancy outcome in studysubjects are summarized in Table 1.

The values of the different tests in predicting delivery withintwo weeks and before 34 weeks are shown in Table 2. Thebest single test for the prediction of delivery within twoweeks was fFN (sensitivity 83% and specificity 84%),whereas CL measurement had a slightly better sensitivitythan fFN (71 vs. 50%, respectively) for predicting deliverybefore 34 weeks. The specificity of fFN was significantlybetter than CL measurement (P < 0.001) and similar to thespecificity of contingent testing. With a policy of contingentuse of fFN (fFN tested only for CL more than 15 mm butless than 30 mm), the fFN testing could have been avoidedin 40 of 62 women (65%). With a policy of contingent mea-surement of CL (CL measured only if negative fFN), the CLmeasurement could have been avoided in 14 of 62 women(23%). Overall, we found no significant difference betweenthe sensitivity of fFN testing, CL measurement, and theircombinations.

The predictive values of the IGFBP-1 test were very poor,with a sensitivity of 17% and 14% for the prediction ofpreterm delivery within two weeks and before 34 weeks,respectively.

Contingent Use of Fetal Fibronectin Testing and Cervical Length Measurement in Women With Preterm Labour

APRIL JOGC AVRIL 2010 309

Figure 2. Contingent use of cervical length measurementFigure 1. Contingent use of fetal fibronectin testing

DISCUSSION

The results of this study confirm that both cervicovaginalfFN testing and endovaginal CL measurement providegood prediction of delivery within two weeks or before34 weeks in women with threatened preterm labour. Inaddition, we have tested two different strategies combiningboth tests, with the aim of improving the predictive value whiledecreasing the need for additional resources. For the pre-diction of delivery within two weeks, the best performancewas provided by fFN testing alone, whereas contingent useof CL measurement was the best predictor of deliverybefore 34 weeks. However, the sensitivities did not differsignificantly between the various combinations; this meansthat, depending on local resources, a choice can be madeamong these options. In facilities where vaginal ultrasoundequipment and expertise are readily available (especially intertiary care centres), CL measurement is a good test for theinitial triage. When CL results are in an intermediate range,fFN testing clearly provides useful additional informationto decide whether the woman with resolved preterm labourcan be discharged. In facilities where vaginal ultrasound isnot routinely offered, fFN testing offers an excellent optionto decide if a woman with preterm labour requires treat-ment and referral to a tertiary facility. Whatever the strategyused, it is important to perform the fFN swab samplingbefore any other vaginal examination (CL measurement ordigital examination) as routinely recommended for fFN

testing. The swab is simply discarded if the test is deemedunnecessary on the basis of CL measurement

There is no consensus about the benefits of combining fFNtesting and CL measurement in women with pretermlabour. Rozenberg et al.,16 using a one-step combination(both tests performed for every patient), found the combi-nation of tests to have limited value, whereas several otherstudies reported increased predictive values when both testswere combined in various ways.15,1820 The selective use offFN after CL measurement, similar to the contingent use wetested, was proposed by Hincz et al.18 and by Schmitz et al.20

Both studies found an improved specificity when fFN test-ing was limited to cases of intermediate measurement of CL(2131 mm for Hincz, and 1630 mm for Schmitz). Gomezet al.19 found a significant improvement in the prediction ofpreterm delivery when fFN was tested after a CL < 30 mm.We were unable to find any report of the contingent use ofCL measurement after fFN testing.

Our study has some limitations. First, the clinicians provid-ing care for the women, except for IGFBP-1, were notblinded to the results of the tests, and the results of the testsmight have changed the management of the pregnancy.However, the primary goal of this study was not to validatefFN and CL as markers of the risk of preterm delivery,which has been done previously, but rather to test their con-tingent use. Another limitation of the study is its relativelysmall sample size. We cannot exclude a lack of statisticalpower to detect subtle differences in the predictive valuesbetween the various combinations of tests. However, webelieve that this study provides important information toobstetric care providers who have to choose a rationalalgorithm for management in cases of threatened pretermlabour, adapted to the local resources. The reproducibilityof measurements was not tested in this study. However,numerous reports have confirmed that both cervical lengthmeasurement and fFN testing had a good to excellentreproducibility.3 We chose not to perform a cost-effectivenessanalysis,27 because the cost of cervical length measurementis highly dependent on the availability of vaginal ultrasoundand expertise. Depending on the type of facility andavailable staff, the choice of the first-line test, fFN testing, orCL measurement, provides similar predictive values amongwomen with threatened preterm labour.

A secondary objective of this study was to validate the useof phosphorylated IGFBP-1 as a marker of an increasedrisk of preterm delivery. In contrast to previousreports,2124,2830 our study found that IGFBP-1 was a verypoor predictor of preterm delivery. Only five women out of62 tested had a positive result, and the sensitivity for thedetection of preterm delivery was extremely low, below20%. The reasons for this poor performance are unclear.

OBSTETRICS

310 APRIL JOGC AVRIL 2010

Table 1. General characteristics, pregnancy outcome,and test results

Characteristics N = 62

Maternal age in years, mean SD 27.6 6.2Nulliparous, n (%) 29 (46.8)Maternal transfer, n (%) 37 (59.7)Gestational age at inclusion in weeks, mean SD 29.4 2.5Gestational age at delivery in weeks, mean SD 36.5 3.2Received tocolytics, n (%) 44 (71.0)Delivery within 2 weeks, n (%) 6 (9.7)Delivery before 34 weeks, n (%) 14 (22.6)Delivery before 37 weeks, n (%) 23 (37.1)Admission to delivery interval in days, mean SD 49.9 22.8Median cervical length in mm (range) 26.5 (051)

Cervical length < 15 mm 16 (28.8)Cervical length < 25 mm 30 (48.4)Cervical length 30 mm 24 (38.7)

Positive fFN, n (%)Positive IGFBP-1, n (%)

14 (22.6)5 (8.1 )

We followed the manufacturers instructions carefully. Thesample size of the current study is similar to those of previ-ous reports. As with all new techniques, there is a potentialpublication bias, since positive results are more likely to bepublished than studies with negative results.31 We thereforebelieve that the use of IGFBP-1 as a marker of pretermdelivery with intact membranes requires further prospectiveand adequately powered studies, and cannot yet match thereliability of fFN.

CONCLUSION

We have confirmed that fFN testing in patients selected byuse of cervical sonography is more specific for predictingpreterm birth than cervical length measurement alone, andis as effective as fFN testing in all women. Using this con-tingency testing can reduce the number of fFN tests per-formed by 65%. On the other hand, cervical ultrasoundafter fFN triage is an acceptable option, depending on theresources available. The effect of the contingent use of

these tests on preterm birth rates, the duration of hospital-ization, and overall costs now must be evaluated inprospective studies.

REFERENCES

1. Goldenberg RL, Culhane JF, Iams JD, Romero R. Epidemiology and causesof preterm birth. Lancet 2008;371(9606):7584.

2. Iams JD, Romero R, Culhane JF, Goldenberg RL. Primary, secondary, andtertiary interventions to reduce the morbidity and mortality of pretermbirth. Lancet 2008;371(9607):16475.

3. Iams JD. Prediction and early detection of preterm labor. Obstet Gynecol2003;101:40212.

4. Leitich H, Egarter C, Kaider A, Hohlagschwandtner M, Berghammer P,Husslein P. Cervicovaginal fetal fibronectin as a marker for pretermdelivery: a meta-analysis. Am J Obstet Gynecol 1999;180:116976.

5. Goldenberg RL, Mercer BM, Iams JD, Moawad AH, Meis PJ, Das A, et al.The preterm prediction study: patterns of cervicovaginal fetal fibronectin aspredictors of spontaneous preterm delivery. National Institute of ChildHealth and Human Development Maternal-Fetal Medicine Units Network.Am J Obstet Gynecol 1997;177:812.

6. Lockwood CJ, Senyei AE, Dische MR, Casal D, Shah KD, Thung SN, et al.Fetal fibronectin in cervical and vaginal secretions as a predictor of pretermdelivery. N Engl J Med 1991;325:66974.

Contingent Use of Fetal Fibronectin Testing and Cervical Length Measurement in Women With Preterm Labour

APRIL JOGC AVRIL 2010 311

Table 2. Predictive values for preterm birth of cervical length, fFN, IGFBP-1, and contingent use of fFN and cervicallength

DeliverySensitivity% (95% CI)

Specificity% (95% CI)

LR+(95% CI)

LR(95% CI)

PPV% (95% CI)

NPV% (95% CI)

Within 2 weeks(prev 9.7%)

fFN + 83 (36100) 84 (7292) 5.2 (2.610.4) 0.2 (0.01.2) 36 (1365) 98 (89100)IGFBP-1 17 (064) 93 (8398) 2.3 (0.317.6) 0.9 (0.61.3) 20 (0.572) 91 (8197)CL < 25 mm 50 (1288) 52 (3865) 1.0 (0.42.4) 1.0 (0.42.2) 10 (226) 91 (7598)Contingent fFN 50 (1288) 64 (5077) 1.4 (0.63.3) 0.8 (0.31.8) 13 (334) 92 (7998)Contingent CL 83 (36100) 62 (4875) 2.2 (1.43.6) 0.3 (0.01.6) 19 (639) 97 (85100)

< 34 weeks(prev 22.6%)

fFN + 50 (2377) 85 (7294) 3.4 (1.48.1) 0.6 (0.31.0) 50 (2377) 85 (7294)IGFBP-1 14 (243) 94 (8399) 2.3 (0.412.4) 0.9 (0.71.1) 40 (585) 79 (6689)CL < 25 mm 71 (4292) 58 (4372) 1.7 (1.12.7) 0.5 (0.21.2) 33 (1753) 87 (7196)Contingent fFN 64 (3587) 71 (5683) 2.2 (1.23.4) 0.5 (0.21.0) 39 (2061) 87 (7396)Contingent CL 71 (4292) 67 (5280) 2.1 (1.33.6) 0.4 (0.21.0) 38 (2059) 89 (7497)

< 37 weeks(prev 37.1%)

fFN + 48 (3560) 92 (8699) 6.2 (1.920.0) 0.5 (0.40.8) 79 (6889) 75 (6486)IGFBP-1 13 (521) 95 (89100) 2.5 (0.514) 0.9 (0.81.1) 60 (4872) 65 (5377)CL < 25 mm 74 (5290) 67 (5081) 2.2 (1.33.7) 0.4 (0.20.8) 57 (3774) 81 (6493)Contingent fFN 56 (3477) 74 (5887) 2.2 (1.24.2) 0.6 (0.41.0) 56 (3477) 74 (5887)Contingent CL 65 (4384) 72 (5585) 2.3 (1.34.1) 0.5 (0.30.9) 58 (3777) 78 (6190)

fFN+: positive fetal fibronectin test; CL: cervical length; contingent fFN: CL measured in all cases, and fFN only for CL between 16 and 30 mm; contingent CL: fFNtested in all cases, and CL only for negative fFN; LR+: likelihood ratio for a positive result; LR:likelihood ratio for a negative result; PPV: positive predictive value;NPV: negative predictive value; prev: prevalence;P > 0.05 for all comparisons between sensitivities; P< 0.001 vs. CL.

7. Berghella V, Hayes E, Visintine J, Baxter JK. Fetal fibronectin testing forreducing the risk of preterm birth. Cochrane Database Syst Rev2008:CD006843.

8. Iams JD, Goldenberg RL, Meis PJ, Mercer BM, Moawad A, Das A, et al.The length of the cervix and the risk of spontaneous premature delivery.National Institute of Child Health and Human Development MaternalFetalMedicine Unit Network. N Engl J Med 1996;334:56772.

9. Coleman MA, Keelan JA, McCowan LM, Townend KM, Mitchell MD.Predicting preterm delivery: comparison of cervicovaginal interleukin(IL)-1beta, IL-6 and IL-8 with fetal fibronectin and cervical dilatation. Eur JObstet Gynecol Reprod Biol 2001;95:1548.

10. Owen J, Yost N, Berghella V, Thom E, Swain M, Dildy GA, 3rd, et al.Mid-trimester endovaginal sonography in women at high risk forspontaneous preterm birth. JAMA 2001;286:13408.

11. Berghella V, Roman A, Daskalakis C, Ness A, Baxter JK. Gestational age atcervical length measurement and incidence of preterm birth. ObstetGynecol 2007;110(2 Pt 1):3117.

12. Grimes-Dennis J, Berghella V. Cervical length and prediction of pretermdelivery. Curr Opin Obstet Gynecol 2007;19:1915.

13. Crane JM, Hutchens D. Transvaginal sonographic measurement of cervicallength to predict preterm birth in asymptomatic women at increased risk:a systematic review. Ultrasound Obstet Gynecol 2008;31:57987.

14. Schmitz T, Kayem G, Maillard F, Lebret MT, Cabrol D, Goffinet F.Selective use of sonographic cervical length measurement for predictingimminent preterm delivery in women with preterm labor and intactmembranes. Ultrasound Obstet Gynecol 2008;31:4216.

15. Rizzo G, Capponi A, Arduini D, Lorido C, Romanini C. The value of fetalfibronectin in cervical and vaginal secretions and of ultrasonographicexamination of the uterine cervix in predicting premature delivery forpatients with preterm labor and intact membranes. Am J Obstet Gynecol1996;175(5):114651.

16. Rozenberg P, Goffinet F, Malagrida L, Giudicelli Y, Perdu M, Houssin I,et al. Evaluating the risk of preterm delivery: a comparison of fetalfibronectin and transvaginal ultrasonographic measurement of cervicallength. Am J Obstet Gynecol 1997;176(1 Pt 1):1969.

17. Goldenberg RL, Iams JD, Das A, Mercer BM, Meis PJ, Moawad AH, et al.The Preterm Prediction Study: sequential cervical length and fetalfibronectin testing for the prediction of spontaneous preterm birth.National Institute of Child Health and Human Development Maternal-FetalMedicine Units Network. Am J Obstet Gynecol 2000;182:63643.

18. Hincz P, Wilczynski J, Kozarzewski M, Szaflik K. Two-step test: thecombined use of fetal fibronectin and sonographic examination of theuterine cervix for prediction of preterm delivery in symptomatic patients.Acta Obstet Gynecol Scand 2002;81:5863.

19. Gomez R, Romero R, Medina L, Nien JK, Chaiworapongsa T, Carstens M,et al. Cervicovaginal fibronectin improves the prediction of preterm deliverybased on sonographic cervical length in patients with preterm uterinecontractions and intact membranes. Am J Obstet Gynecol 2005;192:3509.

20. Schmitz T, Maillard F, Bessard-Bacquaert S, Kayem G, Fulla Y, Cabrol D,et al. Selective use of fetal fibronectin detection after cervical lengthmeasurement to predict spontaneous preterm delivery in women withpreterm labor. Am J Obstet Gynecol 2006;194:13843.

21. Paternoster DM, Muresan D, Vitulo A, Serena A, Battagliarin G,Dellavanzo M, et al. Cervical phIGFBP-1 in the evaluation of the risk ofpreterm delivery. Acta Obstet Gynecol Scand 2007;86:1515.

22. Elizur SE, Yinon Y, Epstein GS, Seidman DS, Schiff E, Sivan E.Insulin-like growth factor binding protein-1 detection in preterm labor:evaluation of a bedside test. Am J Perinatol 2005;22:3059.

23. Kurkinen-Raty M, Ruokonen A, Vuopala S, Koskela M, Rutanen EM,Karkkainen T, et al. Combination of cervical interleukin-6 and -8,phosphorylated insulin-like growth factor-binding protein-1 andtransvaginal cervical ultrasonography in assessment of the risk of pretermbirth. BJOG 2001;108:87581.

24. Kekki M, Kurki T, Karkkainen T, Hiilesmaa V, Paavonen J, Rutanen EM.Insulin-like growth factor-binding protein-1 in cervical secretion as apredictor of preterm delivery. Acta Obstet Gynecol Scand 2001;80:54651.

25. Berghella V, Bega G, Tolosa JE, Berghella M. Ultrasound assessment of thecervix. Clin Obstet Gynecol 2003;46:94762.

26. Berghella V, Talucci M, Desai A. Does transvaginal sonographicmeasurement of cervical length before 14 weeks predict preterm delivery inhigh-risk pregnancies? Ultrasound Obstet Gynecol 2003;21:1404.

27. Abenhaim HA, Morin L, Benjamin A. Does availability of fetal fibronectintesting in the management of threatened preterm labour affect theutilization of hospital resources? J Obstet Gynaecol Can 2005;27:68994.

28. Tanir HM, Sener T, Yildiz Z. Cervical phosphorylated insulin-like growthfactor binding protein-1 for the prediction of preterm delivery insymptomatic cases with intact membranes. J Obstet Gynaecol Res2009;35:6672.

29. Altinkaya O, Gungor T, Ozat M, Danisman N, Mollamahmutoglu L.Cervical phosphorylated insulin-like growth factor binding protein-1 inprediction of preterm delivery. Arch Gynecol Obstet 2009;279:27983.

30. Balic D, Latifagic A, Hudic I. Insulin-like growth factor-binding protein-1(IGFBP-1) in cervical secretions as a predictor of preterm delivery. J MaternFetal Neonatal Med 2008;21:297300.

31. Sridharan L, Greenland P. Editorial policies and publication bias: theimportance of negative studies. Arch Intern Med 2009;169:10223.

OBSTETRICS

312 APRIL JOGC AVRIL 2010