Professor L. Poston, Maternal and Fetal Research Unit, Department of Women's Health, GKT, St Thomas' Campus (King's College, London), St Thomas' Hospital, London, SE1 7EH, UK.
Objective To assess efficacy of cervico-vaginal fetal fibronectin as a predictor of spontaneous preterm birth in a high risk antenatal population, and to evaluate the psychological impact of fetal fibronectin testing.
Design An observational study.
Setting The antenatal clinic at a tertiary referral hospital.
Population One hundred and forty-six pregnant women with known risk factors for spontaneous preterm birth.
Methods Women designated as ‘at risk’ for preterm delivery by clinical history were screened for fetal fibronectin at 24 and again at 27 weeks of gestation. Anxiety levels were assessed by questionnaire and compared with anxiety levels of 206 low risk women also tested for fetal fibronectin. Fetal fibronectin results were disclosed to the woman and her clinician.
Main outcome measures Maternal anxiety and efficacy of the 24-week fetal fibronectin test to predict delivery before 30, 34 and 37 weeks of gestation.
Results Maternal anxiety was higher pretesting in those at high risk compared with low risk women undergoing the test. Among the high risk women, anxiety was raised to clinically significant levels in those receiving a positive fetal fibronectin screening test result. In all women, 5%, 9% and 21% delivered <30, <34 or <37 weeks of gestation, respectively. Nine percent (n= 13) tested positive for fetal fibronectin at 24 weeks. Predictive power for fetal fibronectin (24 weeks) was greatest for delivery <30 weeks of gestation, with a likelihood ratio of 15 for a positive test (6/13 positive women delivered before 30 weeks).
Conclusions Fetal fibronectin was most efficient as a predictor of preterm spontaneous delivery <30 weeks of gestation, but was associated with high levels of anxiety.
Preterm birth is a major problem worldwide with little variation in the incidence in recent decades among developed countries, including the UK.1 Although advances in obstetric and neonatal intensive care have dramatically improved survival rates (e.g. >90% before 32 weeks of gestation in a recent New Zealand national study),2 neonatal morbidity remains high3 with many having neurologic and development disabilities.4–6
The development of accurate methods of prediction of risk of early preterm delivery would facilitate appropriate use of resources, both through a selective increase in surveillance and in timely provision of effective interventions, such as antenatal steroids. Promising prophylactic interventions for labour prevention including progesterone,7 clindamycin8 or cerclage9 could also be targeted at this population. Prediction of women at risk based on previous obstetric history is unreliable10 and has led to investigation of other approaches including assessment of cervical length,11 bacterial vaginosis12 and biochemical markers.13 Of these, the presence of fetal fibronectin in cervico-vaginal secretions has been most intensively studied.14–16 A recent systematic review17 concluded that fetal fibronectin was most accurate in predicting spontaneous preterm birth within 7–10 days of testing among women with symptoms of threatened preterm birth (likelihood ratio 5.4, 95% CI 4.4–6.7), but was less effective in asymptomatic women. However, the analysis included both high or low risk populations. In this study, we present data to suggest that fetal fibronectin can be effective in prediction of preterm delivery among a UK inner city population of asymptomatic women with known risk for preterm delivery on the basis of clinical history.
Screening tests may also cause anxiety and stress has been associated with enhanced risk of preterm delivery.18,19 We were therefore concerned that while fetal fibronectin may be a useful test for prediction of preterm labour, it may in itself enhance risk through raising levels of antenatal stress.
Women recognised to be at increased risk of spontaneous preterm delivery were recruited consecutively over a two-year period from the general antenatal clinics of Guy's and St Thomas' Hospitals, London, UK. They were screened for fetal fibronectin at 24 and 27 weeks of gestation with the principal aim of determining whether the 24-week test efficiently predicted early delivery <30 weeks. This we believed to be a clinically relevant endpoint as the vast majority of mortality and morbidity occurs in deliveries before this gestation. In view of the potential effects of the fetal fibronectin test on anxiety, psychological assessments were made at both 24 and 27 weeks of gestation to assess the impact of the test on the anxiety level of the women. Comparisons were made with a group of low risk women similarly tested.
A study-specific leaflet was included in the ‘pregnancy pack’ sent to all women at booking informing them of the study. Eligibility criteria included previous preterm birth or preterm prelabour rupture of the membranes before 37 weeks of gestation, previous late miscarriage (16+0 to 23+6 weeks of gestation), a current multiple pregnancy or prior cervical surgery.
The study was approved by the Local Research Ethics Committee and all women gave informed written consent. The likely implications of a positive and negative result were explained at the time of consent. Each woman had an obstetric ultrasound assessment at 20 weeks of gestation to rule out major anomalies and confirm gestational age. Where there was a discrepancy of 10 or more days between the menstrual and ultrasound assessment of gestational age, the ultrasound generated dates were used. Fetal fibronectin tests were planned for 23+1 to 24+6 weeks of gestation (‘24 week test’) and 26+0 to 27+6 weeks gestation (‘27 week test’). The purpose of repeat testing was to identify the temporal relationship of the fetal fibronectin tests and to compare the efficacy of the 24 and 27 week test as a predictive assay. A sterile speculum examination for sample collection was performed. Dacron swabs were rotated in the posterior vaginal fornix for 15 seconds to ensure saturation. Samples were frozen in buffer at −70°C until assay. Fetal fibronectin concentrations were determined using a commercially available quantitative ELISA assay (PTDcheck, Adeza Biomedical, Sunnyvale, California, USA) with specific monoclonal antibody (FDC-6) against the onco-fetal domain of fetal fibronectin. Results were made available to the woman by telephone within three working days of the test. Concentrations greater than 50 ng/mL were designated positive.15 Women with positive results were seen by the lead investigator within 24 hours of being informed of their results. At this consultation, the implication of a positive test was further emphasised and the symptoms and signs of preterm labour were discussed. The importance of early attendance upon appearance of any indications of preterm labour was stressed. Written leaflets were given to reinforce the verbal advice. All management and treatment decisions were at the discretion of the woman's consultant obstetrician, but was not based on fetal fibronectin results. Women with negative results were informed of the results in writing, with copies sent to their obstetrician, general practitioner and midwife. Delivery details were obtained by the investigators from obstetric notes and the hospital computer database. Births that followed spontaneous labour or preterm prelabour rupture of the membranes were designated as ‘spontaneous’.
Anxiety was assessed using the six-item short form of the state scale of the Speilberger State–Trait Anxiety Inventory.20 Women rated the extent to which they were currently experiencing feelings of anxiety on a four-point scale. In addition, questions were asked in relation to anxiety about the outcome of the pregnancy, and anxiety in anticipation of, and after (if appropriate), the test. Inclusion for this part of the study required proficiency in English. The questionnaires were completed immediately before undergoing the 24 week and 27 week fetal fibronectin test and at six weeks postpartum. The standardised mean score in this test is 35 for non-pregnant women; above 42 is considered diagnostic of a generalised anxiety disorder. As a control, a second group of 206 low risk pregnant women who had taken part in a low risk study and who had screened negative for fetal fibronectin at the same time points were also asked to complete anxiety questionnaires.
All demographic data and test results were entered onto a customised database (Microsoft Access) and registered under the UK Data Protection Act. Outcomes investigated were spontaneous preterm birth, <30 weeks, <34 weeks and <37 weeks of gestation. Post-test probability of preterm birth was calculated, and results were also expressed as the likelihood ratio of a given test result.21 Overall pregnancy survival was calculated using Kaplan–Meier survival curves and the log-rank test was used to compare overall pregnancy survival between fetal fibronectin positive and negative women. Risk ratios were calculated using binomial regression (Stata 7.0, Statacorp. 2001 Stata Statistical Software. Release 7.0, Stata, College Station, Texas, USA), unadjusted and adjusted for known predictors of preterm delivery. Analysis of anxiety data was conducted using SPSS for windows. T tests were used to compare anxiety in women at high risk of preterm delivery and women at low risk. T tests were used to compare anxiety in women at high risk of preterm delivery who received positive fetal fibronectin test results and women at high risk of preterm delivery who received negative fetal fibronectin test results.
One hundred and forty-eight women were screened at 24 weeks of gestation and 89% returned for testing at 27 weeks. An additional nine women had been recruited but were screened outside the preset gestational age limits and therefore excluded from analysis. Eleven of the women had an elective preterm delivery and were included in the survival analyses, graphs and log-rank tests up until the point of delivery only. Of these two were delivered prior to 30 weeks and are excluded from all analyses, four had an elective preterm delivery by 34 weeks, and five between 34 and 37 weeks. Over half of the women had more than one risk factor for spontaneous preterm birth (Table 1). The demographic characteristics of the population screened are also shown in Table 1.
Table 1. Demographic characteristics and risk factors for preterm delivery in study population (n= 146). Values are presented as n* (%) or median [range].
Some have more than one risk factor.
Previous late miscarriage
Previous preterm delivery, 24–29 weeks
Previous preterm delivery, 30–33 weeks
Previous preterm delivery, 34–36 weeks
Previous preterm prelabour rupture of membranes
More than one of the above
Smoked in pregnancy
Smoked before pregnancy
Left school >16 years
74 (51) (one unknown)
Age at expected date of delivery
Primary outcome data were available for the remaining 146 women screened at 24 weeks of gestation. Nine percent of subjects (n= 13) were fetal fibronectin positive at 24 weeks and 8% (n= 10) at 27 weeks. The incidence of spontaneous preterm birth was 5% before 30 weeks of gestation, 9% before 34 weeks and 21% before 37 weeks.
Kaplan–Meier ‘survival’ curves, comparing the number of fetal fibronectin positive and negative women tested at 24 weeks of gestation who remain undelivered, are shown in Fig. 1. There was an excess of spontaneous early births in the fetal fibronectin positive group with a two- to three-week delay between positive testing and the gestational age at delivery. The log-rank test showed a highly significant difference between the survival curves (P= 0.0015).
Table 2 summarises the predictive power of the fetal fibronectin tests for delivery before 30, 34 and 37 weeks. Three-quarters of spontaneous preterm deliveries below 30 weeks of gestation were correctly identified by a positive ‘24 week’ fetal fibronectin test. Most women who tested positive at 24 weeks were also positive at 27 weeks of gestation but five became positive who were previously negative. The 27 week test had a similar predictive value to the 24 week test for delivery before 30 weeks (Table 2). The unadjusted risk ratio for spontaneous preterm delivery before 30 weeks of gestation of the 24 week test was 30.7 with a 95% CI of 6.9–136.9. Adjustment for the different a priori risk factors and different interventions administered had little effect on the predictive ability; multiple pregnancy (risk ratio 30.7), previous late miscarriage (31), preterm birth (30.7) and preterm prelabour rupture of the membranes (30.9), smoking pre-pregnancy (29.5), age at delivery (30.4), body mass index (28.1), cervical suture (26.7), other antibiotic (30.5), antenatal steroid use (31.6) and tocolytics (30.3).
Table 2. Vaginal fetal fibronectin at 24 and 27 weeks of gestation as a predictor of spontaneous preterm labour before 30, 34 and 37 weeks of gestation.
24 week test
Fetal fibronectin −ve
Fetal fibronectin +ve
Fetal fibronectin −ve
Fetal fibronectin +ve
27 week test
Fetal fibronectin −ve
Fetal fibronectin +ve
Fetal fibronectin −ve
Fetal fibronectin +ve
The results of the anxiety analysis are shown in Table 3. Women at high risk for preterm delivery were more anxious than women at low risk, both before the 24 and the 27 week fetal fibronectin test. At 27 weeks of gestation, women who had received a positive test result at 24 weeks were significantly more anxious than those whose previous test result was negative [t(100)= 1.70, P < 0.05], with the anxiety scores above the cutoff indicating clinically significant levels of anxiety. Among women who received a negative test result at 24 weeks, anxiety levels immediately prior to repeat testing were unchanged.
Table 3. Anxiety index [mean (SD)] in women at low and high risk of preterm delivery, receiving positive and negative test results following fetal fibronectin screening.
Low risk of preterm delivery
High risk of preterm delivery
Significance (low vs high)
Not included in the analysis as confounded by anxiety affects of preterm birth and NICU.
This study, the largest to date in high risk asymptomatic women, provides further evidence that the presence of fetal fibronectin in the cervico-vaginal secretions is predictive of preterm delivery. A negative test reduced the possibility of delivery to levels similar to the background risk. Importantly, early testing may have use as a clinical tool in the prediction of delivery before 30 weeks of gestation (i.e. the deliveries at greatest risk of mortality and morbidity). However, this observation has to be considered in light of the novel evidence that fetal fibronectin testing increased levels of anxiety in high risk women.
The elevation of anxiety to clinically significant levels, nearly three weeks after a positive fetal fibronectin result, is a cause for concern. Anxiety may contribute to preterm delivery18,19 or lead to alterations in maternal–fetal haemodynamics22,23 and the fetal fibronectin test could potentially increase the risk of delivery. Evidence from other screening tests shows that anxiety can be reduced or avoided by addressing common concerns and misinterpretations about the meaning of a positive test result.24 As there is evidence that anxiety reduction may prolong gestation,25 benefit could be gained in women testing negative for fetal fibronectin. Unfortunately, there was no demonstrable reduction in anxiety after a negative test result, although this reassessment took place immediately prior to the speculum examination for the 27 week test and maternal anxiety may have been elevated in anticipation of the either the procedure itself or the subsequent result.
The risk ratio of 30.7 for delivery before 30 weeks in the high risk women with positive tests at 24 weeks compared very favourably with the likelihood ratio of 3.64 for delivery before 34 weeks, the earliest gestation reported in a recent systematic review of the predictive value of fetal fibronectin testing.18 Screening at 24 weeks of gestation may be preferable to target surveillance or steroids, particularly in view of the current concern regarding repeated dosage.26 One-fifth of very preterm babies (27/28 weeks of gestation) cited in the recent CESDI report27 received more than one course of steroids. The report also highlights that steroids should have been administered earlier in a number of cases. The fetal fibronectin test results in this study were made known to the woman's clinician, but it is unlikely that would have influenced gestational age at delivery in the study cohort, as the interventions used in this study did not influence the risk ratios for delivery. Any possible additional effects regarding anxiety need further investigation.
The performance of the fetal fibronectin test in this study was better than that of the largest longitudinal study17 in low risk asymptomatic women. Improved predictive capacity among high risk women could relate to the high values for fetal fibronectin observed, the risk of preterm delivery being proportional to fetal fibronectin concentrations to values up to 300 ng/mL.28
Few previous studies have addressed the issue of prediction of spontaneous preterm birth by fetal fibronectin testing in high risk women, and are limited by either very few women delivering before 30 weeks29 or not providing this data.30,31 Morrison et al.32 probed the value of fetal fibronectin in prediction of early preterm delivery in an investigation of fetal fibronectin in conjunction with uterine contractions, but despite the women being symptomatic, the test did not perform as well as that reported here.
The mechanism of appearance of fetal fibronectin in cervico-vaginal secretions of women destined to deliver prematurely remains unknown, although it has been suggested that intrauterine infection may play a role.33 Given the association between intrauterine infection and early preterm birth,34,35 this study would also support this hypothesis. The association of positive fetal fibronectin testing and infection has provided the rationale for using antibiotics as an intervention for preterm labour prevention but the studies to date in asymptomatic women ‘at risk’ have produced discordant results. Evidence suggests antibiotic use may either have little impact,36 detrimental effects37–39 or result in a possible reduction8,40 in preterm delivery rates. It has been suggested that future clinical trials of antibiotic prophylaxis should be concentrated in women most likely to benefit (i.e. those at risk of deliveries <32 weeks of gestation).41 The fetal fibronectin test, as described in this study, may thus provide an efficient screening method for detection of asymptomatic women at risk of early preterm delivery and for targeted antibiotic therapy, as antibiotics such as clindamycin given early may be beneficial.8
Despite the recent conclusion that fetal fibronectin is best for prediction of delivery in symptomatic women,16 this study demonstrates the predictive value of fetal fibronectin in screening for spontaneous preterm births <30 weeks of gestation in an asymptomatic high risk population. For the first time, we have identified a negative psychological impact of the test. Although fetal fibronectin is an effective method of identification of women at risk, until effective interventions are established, use in clinical practice may be limited to conservative management and increased surveillance.
The authors would like to thank Mr Steve Sarre for help with the database, Ms Brigid Offley Shore for contributing to patient recruitment and Dr Myfanwy Morgan for advice on socio-economic variables.