OBJECTIVE: To assess the accuracy of vaginal fetal fibronectin sampling without use of a sterile speculum examination as a screening test for predicting spontaneous preterm birth.
METHODS: A historical cohort of patients who were followed up with serial fetal fibronectin testing between 1998 and 2001 was identified. All patients were considered to be at high risk for preterm delivery and were screened with fetal fibronectin testing without using a speculum at 2- to 3-week intervals from 22 weeks to 32 weeks of gestation. Charts were reviewed for fetal fibronectin results and pregnancy outcome data. Groups were compared using χ2 analysis or Fisher exact test with significance defined as P < .05.
RESULTS: A total of 1,396 fetal fibronectin tests from 416 pregnancies were performed via the “blind” sampling technique. Overall, 24.9% of pregnancies delivered spontaneously before 37 weeks; 9.1% delivered spontaneously before 34 weeks. For delivery before 34 weeks of gestation, the test had a sensitivity of 44.7%, a specificity of 88.4%, a positive predictive value of 27.9%, and a negative predictive value of 94.1%. For delivery within 14 and 21 days of a single fetal fibronectin assessment, the test had a sensitivity of 52% and 45.5%, a specificity of 94.5% and 94.9%, a positive predictive value of 14.6% and 22.5%, and a negative predictive value of 99.1% and 98.2%, respectively.
CONCLUSION: “Blind” vaginal fetal fibronectin sampling has high negative predictive values and specificities in predicting spontaneous preterm birth.
LEVEL OF EVIDENCE: II-2
Vaginal fetal fibronectin testing via the &#x201C;blind&#x201D; sampling technique has high negative predictive values and specificities in predicting spontaneous preterm birth.
From the *Department of Obstetrics and Gynecology, New York University School of Medicine, New York, New York; and †Department of Obstetrics and Gynecology, Yale University School of Medicine, New Haven, Connecticut.
The research presented in this paper was funded in part by an unrestricted educational grant from Adeza Biomedical, Sunnyvale, California
Presented at the 51st Annual Scientific Meeting of the Society for Gynecologic Investigation, Houston, Texas, March 25–27, 2004.
Address reprint requests to: Ashley S. Roman, MD, MPH, 550 First Avenue, NB 9E2, New York, NY 10016; e-mail: email@example.com.
Received July 29, 2004. Received in revised form October 14, 2004. Accepted October 27, 2004.
Fetal fibronectin is an extracellular matrix protein found at the interface between the chorion and decidua parietalis. It is normally undetectable from the cervix and vagina after 20 weeks of gestation. When detected in cervical or vaginal secretions after this gestational age, fetal fibronectin concentrations of more than 50 ng/mL has been associated with an increased risk of spontaneous preterm delivery. Fetal fibronectin assay of cervicovaginal secretions has proven to be a valuable asset in the prediction of spontaneous preterm delivery since the test was first described by Lockwood and colleagues in 1991.1
Performance of the fetal fibronectin test as originally described, however, can be unwieldy. Most studies describe a technique that requires a speculum examination with visually directed sampling of secretions in the posterior fornix and/or cervical os.1–12 Many patients are reluctant to undergo this invasive examination at regular 14- to 21-day intervals, thus limiting the use of this test in clinical practice. Furthermore, the speculum examination must be performed by a trained medical professional.
To improve patient acceptance of fetal fibronectin testing, we have adopted a protocol of “blind” sampling of vaginal secretions. The labia are digitally separated and a Dacron swab is blindly inserted into the vagina and directed toward the posterior fornix where it is left in place for 30 seconds without the aid of a speculum. This technique serves two benefits: 1) it improves patient tolerance of the examination, and 2) it can be performed by the office nursing staff and does not require physician supervision. The objective of this study was to assess the accuracy of “blind” vaginal fetal fibronectin sampling in predicting spontaneous preterm birth in a large population at risk for preterm delivery.
MATERIALS AND METHODS
After institutional review board approval was obtained, a search of all fetal fibronectin tests in a faculty maternal-fetal medicine practice between 1998 and 2001 was performed. Once patients were identified, a search of patient outpatient records and hospital charts was performed for the following information: demographic information, risk factors for preterm delivery (see below), gestational age at delivery, indication for delivery, gestational age at which fetal fibronectin was sampled, and fetal fibronectin results. Patients were included in this study if they were asymptomatic at time of fetal fibronectin sampling and had one or more of the following risk factors for preterm delivery: history of delivery before 37 weeks, history of second trimester pregnancy loss, incompetent cervix, history of multifetal pregnancy reduction during the index pregnancy, history of second or third trimester bleeding during the index pregnancy, multiple gestation, uterine anomaly, and history of preterm contractions or labor during the index pregnancy. At time of data analysis, all patients had completed their pregnancies. Patients were excluded from analysis if pregnancy outcome data were unavailable (23 patients).
Fetal fibronectin testing was performed according to an established protocol. Serial fetal fibronectin samples were collected every 2–3 weeks starting at 22 weeks of gestation and continuing until 32 weeks of gestation or delivery (if before 32 weeks). At these visits, a nurse or physician would insert a Dacron swab into the vagina without the aid of a speculum and leave the swab in place for at least 30 seconds. Fetal fibronectin swabs were collected more than 12 hours from the last vaginal examination and more than 48 hours after the last reported intercourse. Vaginal swabs were sent for quantitative determination of fetal fibronectin concentration using an enzyme-linked immunosorbent assay (Fetal Fibronectin Immunoassay; Adeza Biomedical Company, Sunnyvale, CA). A fetal fibronectin concentration of 50 ng/mL or greater was considered to be positive.
Patients and physicians were not blinded to fetal fibronectin results. If an assay result was positive, screening for bacterial vaginosis was performed, and the patient was referred for contraction monitoring. If bacterial vaginosis was diagnosed, the patient was treated with antibiotics according to the guidelines of the Centers for Disease Control and Prevention.13 If regular contractions were noted and cervical change was documented either by dilation or by sonographic evidence of effacement, tocolysis was initiated, and glucocorticoids for fetal lung maturity were administered.
For the purpose of this study, preterm delivery was defined as spontaneous delivery (due to preterm, premature rupture of membranes or preterm labor) before 34 weeks of gestation. A gestational age cutoff for prematurity of 34 weeks (as opposed to 37 weeks) was chosen because evidence indicates that infants delivered between 34 and 37 weeks of gestation experience morbidity and mortality at rates similar to term infants delivered between 37 and 40 weeks.14 The vast majority of preterm infants with clinically relevant morbidity and mortality due to prematurity are born at less than 34 weeks of gestation.
Each fetal fibronectin test was correlated with gestational age at delivery. A fetal fibronectin test result was considered to be a true positive if delivery occurred within the specified time frame (eg, within 14 or 21 days) after the positive result. Similarly, a test result was considered to be a false negative if spontaneous delivery occurred within 14 or 21 days of a negative result. If a second test was performed within this 14-day or 21-day window, its results were not taken into consideration when evaluating the predictive value of the first test. For calculating statistics for spontaneous delivery before 34 weeks of gestation, each patient with one or more positive fetal fibronectin test results at any time was designated as positive for fetal fibronectin; patients who tested negative for every specimen were designated as negative.
All data are expressed as means and standard deviations or number of observations with appropriate proportions. Sensitivity, specificity, positive and negative predictive values were calculated for the above outcomes. Statistical analysis was performed using χ2 analysis or Fisher exact test, where appropriate. The threshold of significance was defined as P < .05.
A total of 1,396 fetal fibronectin tests from 416 pregnancies in 416 patients met criteria for analysis. Spontaneous preterm delivery before 37 0/7 weeks of gestation occurred in 24.9% of patients; 9.1% of patients delivered spontaneously before 34 0/7 weeks of gestation. The median number of tests performed per patient was 4 (range 1–7); 89 of 1,396 test results were positive (6.4%). In Table 1, demographic characteristics of the population are shown.
The accuracy of “blind” fetal fibronectin sampling in predicting spontaneous delivery is shown in Table 2. For delivery within 14 days and 21 days, any single fetal fibronectin test had a sensitivity of 52% and 45.5%, specificity of 94.5% and 94.9%, positive predictive value of 14.6% and 22.5%, and negative predictive value of 99.1% and 98.2%, respectively. For spontaneous delivery before 34 weeks, fetal fibronectin had a sensitivity of 44.7%, specificity of 88.4%, positive predictive value of 27.9%, and negative predictive value of 94.1%.
Subgroup analysis of pregnancies with singleton and multiple gestations was performed. There were 247 patients with a singleton gestation; 9 (4%) delivered spontaneously before 34 weeks. Predictive values for preterm delivery using the “blind” technique in singletons are shown in Table 3. For delivery within 14 and 21 days of testing in singletons, “blind” fetal fibronectin sampling had 50% and 53.3% sensitivity, 93.5% and 94.0% specificity, 7.7% and 15.4% positive predictive value, and 99.4% and 99% negative predictive value. In singletons, “blind” fetal fibronectin testing had a 44.4% sensitivity, 88.2% specificity, 12.5% positive predictive value, and 97.7% negative predictive value for delivery before 34 weeks.
There were 169 patients carrying multiple gestations; 29 (17.2%) patients delivered spontaneously before 34 weeks of gestation. Predictive values for preterm delivery using the “blind” technique in multiple gestations are shown in Table 4. For delivery within 14 and 21 days of testing in multiple gestations, “blind” fetal fibronectin sampling had 52.9% and 41.4% sensitivity, 95.5% and 95.9% specificity, 24.3% and 32.4% positive predictive value, and 98.7% and 97.2% negative predictive, respectively. In multiple gestations, “blind” fetal fibronectin sampling had a 44.8% sensitivity, 88.6% specificity, 44.8% positive predictive value, and 88.6% negative predictive value in predicting spontaneous delivery before 34 weeks of gestation.
Since its use was first described in 1991, cervicovaginal fetal fibronectin has proven to be a valuable clinical test.1–12,15 The primary benefit of the test is in its negative predictive value, that is, identifying patients who will not deliver preterm or within a specified time interval. As a result, patients with negative test results can be managed on outpatient basis and avoid unnecessary—and often poorly proven—interventions such as tocolysis and bedrest.
A limitation to using this test, particularly when used at regular intervals as a screening tool in high-risk patients, is the requirement of a speculum examination when sampling secretions. As a result, a trained medical professional must perform the examination with the patient in lithotomy position with moderate discomfort to the patient. One prior study by Coleman and colleagues16 evaluated the correlation between examinations with a speculum and without a speculum in a small number of patients. All patients who tested negative with a speculum examination also tested negative without a speculum, but 27% (6/22) who tested positive with a speculum tested negative when the test was performed without a speculum at the same examination. Two of these 6 patients delivered spontaneously within the next 10 days, a finding that raises concern about false-negative results with the “blind” sampling technique. This study is also limited by the relatively small number of samples evaluated and the lack of statistical calculations for sampling without a speculum and prediction of preterm delivery.
The findings detailed in our study indicate that sampling for fetal fibronectin via the “blind” sampling technique carries similar negative predictive values and specificities for spontaneous preterm birth as speculum-directed sampling in asymptomatic, high risk women. We found a specificity of 94.5% for delivery within 14 days compared with 89% reported in a recent meta-analysis,17 and a negative predictive value of 99.1% compared with 98.7% in a similar high-risk population.2 For predicting spontaneous delivery before 34 weeks, “blind” sampling has a specificity of 88.3% compared with 89% reported in a recent meta-analysis17 and a negative predictive value of 94.1% compared with a range of 88.7–97.8% reported in similar populations.3,12 Although the sensitivity of “blind” sampling is slightly lower than with speculum-directed testing (for delivery within 14 days, 52% compared with 67%;17 for delivery before 34 weeks, 44.7% compared with 53%17), the primary utility of testing is its negative predictive value, regardless of sampling technique.
The strength of this study is that it evaluates the use of this sampling technique in a large, heterogeneous, high-risk population. Furthermore, this study evaluates the use of this technique in everyday clinical practice. The retrospective nature of this investigation, however, poses several limitations. Physicians were not blinded to fetal fibronectin results. Consequently, patients with positive test results received closer surveillance and may have been more likely to receive tocolysis. The benefit of tocolytics in preventing spontaneous preterm delivery is controversial. In fact, most evidence suggests that the use of tocolysis in the setting of spontaneous preterm labor only prolongs pregnancy up to 48 hours.18 Thus, it is unlikely that the use of tocolytics had any effect on the predictive value of the test when examining an interval to delivery of 14 days or delivery rate before 34 weeks. Furthermore, the “blind” vaginal samples in our study were not compared with speculum-directed samples obtained at the same time.
Based on a MEDLINE search of the medical literature from 1951 through 2003 using keywords “fetal fibronectin” and “preterm birth,” this investigation is the second study and the largest to date to assess the accuracy of sampling for fetal fibronectin without the use of a speculum in predicting spontaneous preterm birth. The “blind” sampling technique for fetal fibronectin testing has the potential to enhance patient acceptance of testing.
1. 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 preterm delivery. N Engl J Med 1991;325:669–74.
2. Leeson SC, Maresh MJA, Martindale EA, Mahmood T, Muotone A, Hawkes N, et al. Detection of fetal fibronectin as a predictor of preterm delivery in high risk asymptomatic pregnancies. Br J Obstet Gynaecol 1996;103:48–53.
3. Nageotte MP, Casal D, Senyei AE. Fetal fibronectin in patients at increased risk for premature birth. Am J Obstet Gynecol 1994;170:20–5.
4. Hellemans P, Gerris J, Verdonk P. Fetal fibronectin detection for prediction of preterm birth in low risk women. Br J Obstet Gynaecol 1995;102:207–12.
5. Lockwood CJ, Wein R, Lapinski R, Casal K, Berkowitz G, Alvarez M, et al. The presence of cervical and vaginal fetal fibronectin predicts preterm delivery in an inner-city obstetric population. Am J Obstet Gynecol 1993;169:798–804.
6. Greenhagen JB, Van Wagoner J, Dudley D, Hunter C, Mitchell M, Logston V, et al. Value of fetal fibronectin as a predictor of preterm delivery for a low-risk population. Am J Obstet Gynecol 1996;175:1054–6.
7. Bittar RE, Yamasaki AA, Sasaki S, Zugaib M. Cervical fetal fibronectin in patients at increased risk for preterm delivery. Am J Obstet Gynecol 1996;175:178–81.
8. 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.
9. 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.
10. Goepfert AR, Goldenberg RL, Mercer B, Iams J, Meis P, Moawad A, et al. The preterm prediction study: quantitative fetal fibronectin values and the prediction of spontaneous preterm birth. Am J Obstet Gynecol 2000;183:1480–3.
11. Iams JD, Casal D, MacGregor JA, Goodwin TM, Kreaden US, Lowensohn R, et al. Fetal fibronectin improves the accuracy of diagnosis of preterm labor. Am J Obstet Gynecol 1995;173:141–5.
12. Morrison JC, Neaf RW III, Botti JJ, Katz M, Belluomini JM, McLaughlin BN. Prediction of spontaneous preterm birth by fetal fibronectin and uterine activity. Obstet Gynecol 1996;87:649–55.
13. Sexually transmitted diseases treatment guidelines 2002. Centers for Disease Control and Prevention. MMWR Recomm Rep 2002;51(RR-6):1–78.
14. Fox MD, Allbert JR, McCaul JF, Martin RW, McLaughlin BN, Morrison JC. Neonatal morbidity between 34 and 37 weeks' gestation. J Perinatol 1993;13:349–53.
15. Goldenberg RL, Mercer BM, Meis PJ, Copper RL, Das A, McNellis D. The preterm prediction study: fetal fibronectin testing and spontaneous preterm birth. Obstet Gynecol 1996;87:643–8.
16. Coleman MAG, McCowan LME, Pattison NS, Mitchell M. Fetal fibronectin detection in preterm labor: evaluation of a prototype bedside dipstick technique and cervical assessment. Am J Obstet Gynecol 1998;179:1553–8.
17. Leitich H, Kaider A. Fetal fibronectin: how useful is it in the prediction of preterm birth? Br J Obstet Gynaecol 2003;110(suppl):66–70.
18. Higby K, Xenakis EM, Pauerstein CJ. Do tocolytic agents stop preterm labor? A critical and comprehensive review of efficacy and safety. Am J Obstet Gynecol 1993;168:1247–56.