OBJECTIVE: To investigate accuracy of fetal fibronectin testing to predict preterm birth in twin gestations with symptoms of preterm labor.
METHODS: We reviewed charts of all patients with twin gestations who underwent fetal fibronectin testing and presented with complaints of preterm labor between January 1, 2000, and June 30, 2004. We also reviewed the charts of all singleton gestations with similar complaints that had fetal fibronectin testing between January 1, 2000, and December 31, 2001. All samples were processed using a rapid fetal fibronectin detection system. We calculated the sensitivity, specificity, positive predictive value, and negative predictive value of fetal fibronectin testing in singleton and twin gestations in predicting delivery within 14 days of testing.
RESULTS: Four hundred twenty-nine singletons and 87 twins met the inclusion criteria. The birth rate before 34 weeks of gestation for singletons was 3.5% compared with the rate of twin pregnancies of 28.7%. Fetal fibronectin predicted delivery within 14 days of testing with a sensitivity, specificity, and positive and negative predictive values in singleton gestations of 82%, 90%, 17%, and 99%, respectively. In twin gestations, fetal fibronectin predicted delivery within 14 days of testing with a sensitivity, specificity, and positive and negative predictive values of 71%, 74%, 19%, and 97%, respectively.
CONCLUSION: As noted in singleton pregnancies, fetal fibronectin testing in twins has a high negative predictive value. Fetal fibronectin evaluation may be a useful tool in screening twins with symptoms of preterm labor, because a negative result places these women at a low risk for delivering within 2 weeks of testing.
LEVEL OF EVIDENCE: II
Twins presenting with complaints of labor with a negative fetal fibronectin have a low likelihood of delivering within 2 weeks of testing.
From the 1Department of Obstetrics and Gynecology and3 Division of Maternal Fetal Medicine, Baystate Medical Center, Springfield, Massachusetts; and 2Department of Obstetrics and Gynecology, Albany Medical Center, Albany New York.
The authors thank Allison Lee for her assistance with data abstraction and Kristine L. Cooper, MS, for statistical support.
Presented at the American College of Obstetricians and Gynecologists District I meeting in Quebec, Canada, September 15–18, 2005.
Corresponding author: Glenn Markenson, Division of Maternal Fetal Medicine, Baystate Medical Center, 759 Chestnut Street, Springfield MA; e-mail: firstname.lastname@example.org.
Financial Disclosure Dr. Markenson is on the Speakers Bureau for Adeza Biomedical Corporation, Sunnyvale, CA, the manufacturer of the fetal fibronectin detection system used in this study.
Women with multifetal gestations that present with symptoms of preterm labor can be a diagnostic challenge. These symptoms are common in multiple pregnancies, and it is difficult to determine which patients will truly deliver early. Due to the potential morbidities of prematurity, over treatment is frequent. This results in exposure to potentially unnecessary interventions such as hospitalization, tocolytic use, and prolonged bedrest. To prevent unneeded interventions in singleton gestations with symptoms of preterm labor, fetal fibronectin testing may be helpful.1 In women with threatened preterm labor, the negative predictive value of a fetal fibronectin test is 99% in ruling out delivery within 14 days.2 Unfortunately, the positive predictive value of the test is only 16%, and thus the value of this test appears to be in avoiding unnecessary therapeutic interventions.3 The data are limited regarding the use of fetal fibronectin screening for preterm birth in multiple gestations with symptoms of preterm labor. Studies focusing on fetal fibronectin testing in twins have enrolled only asymptomatic pregnancies as part of a preterm labor screening program.4–6 Information regarding the use of fetal fibronectin in twins with symptoms of preterm labor comes from studies enrolling both singletons and twins, but these investigations include only limited numbers of twins. The goal of this study is to investigate the accuracy of fetal fibronectin testing in twin gestations with symptoms of preterm labor, compared with its use in singleton pregnancies with symptoms of preterm labor.
PARTICIPANTS AND METHODS
We reviewed the charts of all twin pregnancies that underwent fetal fibronectin testing at Baystate Medical Center in Springfield, Massachusetts from January 1, 2000, through June 30, 2004. For comparison, we also reviewed the charts of all singleton gestations that underwent fetal fibronectin testing from January 1, 2000, through December 31, 2001. This was a convenience sample of patients based on the availability of an obstetric database that was implemented during that time frame. Approval for this study was granted by the institutional review board of Baystate Medical Center. Only those patients who presented to the Baystate Medical Center obstetric triage area with complaints concerning preterm labor, such as contractions, pelvic pressure, and low back pain, and who met the following criteria were included in the study: fetal fibronectin testing between 24.0 and 34.9 weeks of gestation, intact membranes, and cervical dilation less than 3 cm. Patients were excluded if they had intercourse, a vaginal examination or a vaginal ultrasound within 24 hours of fetal fibronectin testing. In addition, any patients with a cervical cerclage or those that required a preterm delivery within 14 days of testing due to maternal or fetal complications were excluded. If a patient presented to the triage area more than one time during their pregnancy for evaluation of preterm labor, only the first fetal fibronectin test was included in our analysis. Fetal fibronectin sampling was performed before a digital cervical examination following the manufacturer’s guidelines. A speculum examination was performed and a Dacron (E.I. du Pont de Nemours & Co., Inc., Wilmington, DE) swab was placed in the posterior vaginal fornix, allowing it to absorb the vaginal secretions for 10 seconds. All samples were processed at Baystate Reference Laboratory within 24 hours using the rapid fetal fibronectin TLi system (Adeza Biomedical, Sunnyvale, CA) qualitative method. Results as recommended by the manufacturer were reported by the laboratory as either positive (50 ng/mL or more) or negative (less than 50 ng/mL). Laboratory personnel were not blinded to the clinical situation. Clinical information was obtained from the Baystate Medical Center laboratory and perinatal databases. Data collection included patient demographic information, cervical dilation at the time of fetal fibronectin sampling, and gestational age at delivery.
We calculated the sensitivity, specificity, positive predictive value, and negative predictive value of fetal fibronectin testing in singleton and twin gestations in predicting delivery within 14 days of testing. Statistical tests performed include Student t test, Wilcoxon rank sum and Pearson χ2 test as appropriate and exact methods were used to calculate 95% confidence intervals. Generalized linear models were used to estimate the relative negative predictive values and corresponding confidence intervals. All statistics were calculated using R 2.4.1 GUI software (The R Foundation for Statistical Computing, Vienna, Austria). Statistical significance was defined as P<.05.
A total of 429 singletons and 87 sets of twins met the inclusion criteria. The demographic and clinical characteristics for twin pregnancies based on fetal fibronectin testing are summarized in Table 1. When comparing maternal age, race, gestational age at time of testing, gravidity, and parity, there were no significant differences between the fetal fibronectin–positive and-negative groups. Those that were fetal fibronectin negative, however, delivered at a later mean gestational age (35.5 weeks compared with 34.0 weeks) and the mean interval from testing to delivery was longer (43.8 days compared with 34 days). In addition, the median cervical dilation in the fetal fibronectin–negative group was significantly less (0.0 cm) compared with the positive group (0.75 cm). Not surprisingly, women with a history of preterm delivery, defined as delivery at less than 37 weeks gestation, were significantly more likely to have a positive fetal fibronectin result.
As expected, the twin patient group had a significantly higher rate of preterm birth, with a mean gestational age at delivery 3.6 weeks earlier for twins compared with singletons (Table 2). In addition, the rate of delivery within 14 days of testing was more than doubled in the twin group compared with singletons. Of interest, the higher preterm birth risk in twins was also reflected in a corresponding and significantly higher fetal fibronectin screen positive rate in twin gestations (29.9%) compared with singletons (12.1%).
Predictive values for preterm delivery for both singleton and twins are presented in Table 3. The negative predictive value for delivery 7 days or less and 14 days or less of testing was similar between the two groups; singletons 99% at both intervals, and twins 100% at 7 days or less and 97% at 14 days or less. However, the negative predictive value dropped in twins at 21 days or less (87%) and for predicting delivery at less than 34 weeks (77%), compared with singletons, which had a 98% negative predictive value at both of these periods.
A comparison of the negative predictive values in singletons and twins for the different intervals from testing to delivery (7 days or less, 14 days or less, 21 days or less, and less than 34 weeks) are shown in Table 4 using relative negative predictive values. Relative negative predictive values can be interpreted in the following manner: relative negative predictive value more than 1 indicates that the test result is more indicative in twins than singletons that delivery will not occur. A relative negative predictive value less than 1 indicates that a negative test is less indicative in twins than singletons of not delivering, and a relative negative predictive value equal to 1 indicates there is no difference between the two groups. There seems to be no statistically significant difference in the negative predictive value of fetal fibronectin testing between singletons and twins for delivery 7 days or less, 14 days or less, and 21 days or less from the time of testing. However, in the case of predicting delivery at less than 34 weeks, the relative negative predictive value (0.76, confidence interval 0.58–0.99) indicates that the negative predictive value of fetal fibronectin testing in twins is significantly less indicative that delivery will not occur before 34 weeks as compared with singletons (Table 4).
Preterm birth represents a leading cause of infant morbidity and mortality in the United States. Twelve and one half percent of all births in 2004 occurred before 37 weeks of gestation, and 2% occurred before 32 weeks of gestation. These numbers continue to increase largely in part to the increase in multiple gestations. National statistics for the United States in 2004 revealed the overall rate of twins was 32.2 per 1,000 total live births. This reflects a 2% increase over the previous year and a 70% increase in twins since 1980. Twins are at a considerable risk for preterm birth, with a 60% birth rate before 37 weeks gestation and a 12% birth rate before 32 weeks. Therefore, although twins represent only 2–3% of all neonates born, they contribute disproportionately to the overall burden of preterm birth and neonatal mortality.7
Available interventions to improve neonatal outcomes in pregnancies at risk for preterm delivery are limited. National Institute of Health guidelines8 recommend a single course of antenatal corticosteroids for women at risk for preterm delivery. This intervention has been shown to have maximal benefit when administered between 48 hours and seven days of delivery. The American College of Obstetricians and Gynecologists (ACOG) recommendations9 advise against the use of multiple courses. Nevertheless, identifying women at the greatest risk for delivery within 7 days of presentation remains a difficult task. There have been a large number of studies that investigated the usefulness of fetal fibronectin to predict preterm birth in symptomatic patients. These studies found that only rarely would a preterm birth occur after a negative test.10 However, the efficacy of this test in a large symptomatic twin population has yet to be well defined. In the largest study to date examining the role of fetal fibronectin in symptomatic patients, only 37 of the total of 763 pregnancies studied were twin gestations.3
Previous efforts to establish the role of fetal fibronectin testing in twin gestations are limited to the screening of asymptomatic patients during routine obstetric visits.4–6 In this context, the positive predictive value for preterm birth was 68–73% and the negative predictive value for preterm birth was 70–89%.4–6 Review of our data suggests that although patients presenting with twin pregnancies and symptoms of preterm labor have a higher risk of delivering preterm compared with singletons, the negative predictive value of fetal fibronectin, within 2 weeks of testing is similar to a singleton population (Table 5). Using the likelihood ratios noted in Table 3, a substantial difference in pretest and posttest odds was found in the twin population. For example, the pretest odds of delivering within 14 days are 0.09, that is, before administering the test the odds of delivering in the next 14 days are 0.09 or about 1 in 10. After a positive test result, the odds increase to 0.24 or about one in four. Similarly, after a negative test result, the odds decrease to 0.04 or about one in 20. For these patients, there may be a role for less intensive monitoring and fewer interventions compared with those with a positive result. At our institution, patients presenting to the triage unit with symptoms concerning for preterm labor, that are found to have a negative fetal fibronectin result, may be discharged home with precautions and a strong reassurance that delivery in the next 14 days is highly unlikely.
The major strength of this study is the large number of twin gestations available for analysis. However, due to the retrospective design of the study, we were unable to evaluate the clinical effect of the fetal fibronectin results. Questions regarding whether fetal fibronectin results in twin gestations would change practitioners’ practice patterns and effect the use of resources remain unanswered. Prediction of preterm birth in multiple gestations is a critical issue; therefore, prospective studies to determine the true clinical impact and cost-effectiveness of fetal fibronectin testing for this population are warranted.
1. American College of Obstetrics and Gynecologists. Assessment of risk factors for preterm birth. ACOG Practice Bulletin No. 31. Obstet Gynecol 2001;98:709–16.
2. Revah A, Hannah ME, Sue-A-Quan AK. Fetal fibronectin as a predictor of preterm birth: an overview. Am J Perinatol 1998;15:613–21.
3. Peaceman AM, Andrews WW, Thorp JM, Cliver SP, Lukes A, Iams JD, et al. Fetal fibronectin as a predictor of preterm birth in patients with symptoms: a multicenter trial. Am J Obstet Gynecol 1997;177:13–8.
4. Ruiz RJ, Fullerton J, Brown CE. The utility of fFN for the prediction of preterm birth in twin gestations. J Obstet Gynecol Neonatal Nurs 2004;33:446–54.
5. Oliveira T, de Souza E, Mariani-Neto C, Camano L. Fetal fibronectin as a predictor of preterm delivery in twin gestations. Int J Gynaecol Obstet 1998;62:135–9.
6. Tolino A, Ronsini S, Zullo F, Pellicano M, Regine V, Nappi C. Fetal fibronectin as a screening test for premature delivery in multiple pregnancies. Int J Gynaecol Obstet 1996;52:3–7.
7. Martin JA, Hamilton BE, Sutton PD, Ventura SJ, Menacker F, Kirmeyer S. Births: final data for 2004. Natl Vital Stat Rep 2006;55:1–101.
8. Antenatal corticosteroids revisited: repeat courses. NIH Consens Statement 2000;17:1–18.
9. American College of Obstetrics and Gynecologists. Antenatal corticosteroid therapy for fetal maturation. ACOG Committee Opinion No. 273. Obstet Gynecol 2002;99:871–3.
10. Honest H, Bachmann LM, Gupta JK, Kleijnen J, Khan KS. Accuracy of cervicovaginal fetal fibronectin test in predicting risk of spontaneous preterm birth: systematic review. BMJ 2002;325:301–10.
Figure. No caption available.