Fetal fibronectin is a protein produced at various placental sites and, especially, by the fetal membranes.1–3 Among its potential functions, fetal fibronectin seems to be important for the adherence of the membranes to the decidua. Although not produced in the lower genital tract, before 20 weeks of gestational age fetal fibronectin is commonly found in the cervix and vagina, but after 22 to 24 weeks is found there in only 3–4% of pregnant women.4,5 The likely explanation for these findings is that until the membranes are tightly fixed to the decidua at about 20 weeks, fetal fibronectin routinely seeps downward from the space between the membranes and decidua into the cervix and vagina. After that time, downward seepage often seems to be explained by some pathologic intrauterine process such as chorioamnionitis or bleeding, and this association may explain the relationship observed between cervical or vaginal fetal fibronectin and subsequent preterm birth and neonatal infections.
Traditionally, a level of 50 ng/mL or more has been used to define a positive fetal fibronectin test.2,4 Previously, we have shown an almost linear relationship between fetal fibronectin levels and subsequent preterm birth.6,7 An increase in fetal fibronectin levels over time is associated with an increased risk of preterm birth.8 The relationship between cervical or vaginal fetal fibronectin and preterm birth has been studied extensively in developed countries, and in those studies, fetal fibronectin has emerged as one of the strongest predictors of subsequent preterm birth and especially early preterm birth.2–5,9–13 Fetal fibronectin has not previously been studied in resource-limited settings, among women who are human immunodeficiency virus (HIV)-infected, and in conjunction with a detailed examination of placental histology. Our objectives were 1) to determine whether vaginal fetal fibronectin levels were associated with mother-to-child transmission of HIV, placental histology, and receipt of antibiotics among women in sub-Saharan Africa enrolled in a randomized clinical trial of antibiotics for prevention of preterm birth and mother-to-child transmission of HIV and 2) to determine whether the relationships between vaginal fetal fibronectin concentrations and gestational age and birth weight were similar to those previously reported in developed countries.
The HIV Prevention Trials Network 024 study was a randomized, double blind, placebo controlled, phase III study of antibiotics to reduce both preterm birth and mother-to-child transmission of HIV. This trial has been described in detail previously.14,15 Enrollment began in July 2001, and the last delivery occurred in August 2003. The study was conducted in four sub-Saharan African sites: Lusaka, Zambia; Blantyre and Lilongwe in Malawi; and Dar es Salaam in Tanzania. Each of these sites had a sponsoring U.S. university. The institutional review boards of each U.S. university and the research ethics committees of each of the African sites approved the protocol. After appropriate counseling, each woman provided signed informed consent.
Women were enrolled from prenatal care clinics in each of the sites at 20 to 24 weeks of gestation. Enrollment required serologic confirmation of HIV infection status; in three of four sites, for every five HIV-infected women, one HIV-uninfected woman was enrolled. HIV-uninfected women were included in the trial for two reasons; the first was to reduce stigma for the HIV-infected women in the study, and the second was to increase the generalizability of the preterm birth outcome to all women, not just those who were HIV-infected. For the preterm birth, birth weight, and mother-to-child transmission of HIV outcome analyses, only those women whose newborn outcomes were known, who delivered a liveborn, singleton infant, and who had at least one fetal fibronectin value available for study were included.
When data presented at an interim Data Safety and Monitoring Board review justified the conclusion that the antibiotic regimen did not show a meaningful reduction of mother-to-child transmission of HIV, the protocol team was advised to cease enrollment; thus, no further participants were recruited as of February 21, 2003. Furthermore, because at that time there was a nonsignificant increase in mother-to-child transmission in the treatment group, the Data Safety and Monitoring Board asked that no further antibiotics be dispensed or administered, meaning that after March 5, 2003, some already enrolled women did not receive a labor and delivery antibiotic or placebo dose. Thus, only 1,820 of the 2,433 (75%) women in the analysis subset were eligible to receive their labor and delivery dose. Because we analyzed the data with and without those delivering after March 5, 2003, and found no differences in the rate of preterm birth and chorioamnionitis, we present the data on the entire population of women with available fetal fibronectin values who received the prenatal antibiotic or placebo treatment at 20 to 24 weeks, whether they received the labor and delivery dose or not.
In addition to the study visits that occurred at randomization (20–24 weeks), 28 weeks, and 36 weeks, women also received routine prenatal care at their sites. This care included HIV-related counseling and for HIV-infected women, the provision of nevirapine given according to the HIV Network for Prevention Trials 012 regimen. At 20–24 weeks of pregnancy, all enrolled women received either antibiotics (metronidazole 250 mg and erythromycin 250 mg) or identically-appearing placebos, all orally, three times daily for 7 days. With onset of labor contractions or premature rupture of membranes, study participants initiated a second oral course of antibiotics consisting of metronidazole 250 mg and ampicillin 500 mg or identically-appearing placebos every 4 hours until delivery. The rationale for choosing the specific antibiotics has been described previously.14,15
Structured questionnaires were used to collect information on maternal demographics, obstetric history, medical history, sexual history, as well as use of concomitant antibiotics and other medications. Gestational age was determined using three methods. The first was based on patient recall of the last menstrual period. The second was based on the uterine size in centimeters measured from the pubic symphysis to the uterine fundus with a tape. In addition, at delivery or within 48 hours of birth, the neonate was assessed for gestational age using a modified Ballard examination. Because this study was done in low-resource African sites, ultrasound was not available. Birth weights were collected to provide a relative confirmation of the gestational age data. We therefore had both obstetric and pediatric estimations of gestational age as well as birth weight. In this article, we present results based on each of the methods of gestational age determination, the mean number of weeks from randomization to delivery, as well as birth weight. Preterm birth outcomes are presented as less than 32 weeks, less than 35 weeks, and less than 37 weeks of gestational age.
At enrollment at 20 weeks to 24 weeks (mean gestational age [±standard deviation] 22.0±0.03 weeks) and at the 28 week visit (mean gestational age 28.2±0.04 weeks), vaginal and cervical fluid samples were collected to evaluate fetal fibronectin levels and the presence of gonorrhea and trichomoniasis. Vaginal swabs were also collected to assess bacterial vaginosis based on the Nugent criteria.16 Serum tests for syphilis were performed. All women who tested positive for syphilis were treated with penicillin, as were women with gonorrhea. Women with bacterial vaginosis and trichomoniasis were not routinely treated.
At delivery, the placental samples were collected, placed in formalin, and sent to the site pathology laboratories where they were processed into paraffin tissue blocks. These were transported to the pathology laboratory at the University Teaching Hospital in Zambia, where final tissue slide processing and histologic examinations were performed. The diagnosis of histologic chorioamnionitis was based on qualitative assessment of the membranes, cord, and placenta basalis. A grading system that classified neutrophilic infiltration into none, slight, or marked, and chronic and acute based on the predominant infiltrate of mononuclear or polymorphonuclear (PMN) cells was used.17 For this article, the diagnosis of histologic chorioamnionitis required a marked level of PMN infiltration. The pathologist reading the slides had no knowledge of the clinical data.
The samples for fetal fibronectin were obtained from the posterior vaginal fornix with a polyester swab before the performance of any other portion of the pelvic examination. The samples were placed in buffer, frozen at –70°C within 2 hours of collection, and subsequently shipped to Adeza Biomedical Company (Sunnyvale, CA) for analysis using enzyme-linked immunosorbent assay methodology. The within-run coefficient of variation for this assay ranged from 2.9–4.2%, and the between-run coefficient of variation ranged from 7.6–11.8%. The samples were analyzed by Adeza without any clinical information available.
At study visits conducted at birth and at 4–6 weeks and 3 months, HIV-1-exposed infant venous blood was collected to prepare dried blood spots. Nucleic acids from the dried blood spots were extracted along with internal controls using silica bead technology (BioMerieux, Durham, NC). Human immunodeficiency virus RNA was detected with the BioMerieux NucliSens HIV-1 QL assay for the Malawi and Zambia sites and with the Roche HIV-1Amplicor Monitor assayv.1 for the Tanzania site in a reference laboratory (University of North Carolina, Chapel-Hill, NC). For confirmation of the infant’s HIV-1 infection, a second sample was tested whenever possible; otherwise, the same sample was repeatedly tested. The HIV Prevention Trials Network Clinical Laboratory tested 10% of randomly selected dried blood spots with positive results and an equal number of dried blood spots with negative results. The methods for the measurement of CD4 counts and viral loads have been described previously.15
All statistical analyses were conducted using SAS 9.1 (SAS Institute Inc., Cary, NC). Uncorrected χ2 tests and two-tailed t tests were used to compare proportions and means, respectively. Logistic regression modeling was used when appropriate. To estimate the HIV-1 infection rates at birth and at 4–6 weeks, we used a censored multinomial model with the following possible outcomes: 1) first positive test result occurring at birth, 2) first positive test result at the 4–6 week visit, or 3) negative tests at both birth and 4–6 weeks.18 For analysis purposes, the birth visit was defined as occurring between birth and one week of age. The 4–6 week visit was extended to 4–8 weeks to include neonates who returned for a follow-up visit to that age. Subjects whose HIV-1 infection status was unknown at a visit (either due to lost to follow-up, or missing scheduled visits) were treated as having a censored outcome.
Figure 1 depicts the derivation of the study population. Of the 2,661 women enrolled in the HIV Prevention Trials Network 024 study, 2,353 had liveborn singletons, had fetal fibronectin results, and are those cases presented in the preterm birth and birth weight analyses. Of these, 1,960 had HIV-infected mothers and 314 had HIV-uninfected mothers. Of the 1,960 women with HIV, 1,909 had infant HIV results available. Table 1 shows the characteristics of the study population as well as those with a fetal fibronectin value 50 ng/mL or more at both 20–24 weeks and 28 weeks. At both points, a lower proportion of the women in Dar es Salaam had high fetal fibronectin values, and at 28 weeks there was a significant trend relating high fetal fibronectin values to low maternal education. At both time points, women with syphilis were more likely to have elevated fetal fibronectin values, with the results at 28 weeks statistically significant. At 20–24 weeks, a lower proportion of women with bacterial vaginosis and women with trichomoniasis had high fetal fibronectin values. However, there were no statistically significant differences between HIV-positive and HIV-negative women in the percent with high fetal fibronectin levels. Also, among women with HIV, at both 20–24 weeks and at 28 weeks, fetal fibronectin levels were not associated with either maternal CD4 counts or viral load (data not shown.).
Table 2 presents the distribution of fetal fibronectin values at 20–24 weeks and at 28 weeks. At 20–24 weeks, 4.2% of the values were 50 ng/mL or more, 2.1% of the values were between 50 and 99 ng/mL, 0.8% between 100 and 199 ng/mL, and 1.3% more than 200 ng/mL. Values were generally similar at 28 weeks.
Table 3 shows the distribution of fetal fibronectin values in the antibiotic and placebo treated groups at 20–24 weeks and at 28 weeks. In neither period was there any difference between the antibiotic treatment and placebo groups. Therefore, from this point forward in the analyses, the antibiotic and placebo groups were combined.
We first compared fetal fibronectin values to several mortality outcomes. Maternal mortality occurred in 2.4% of the women, but was not associated with fetal fibronectin levels. Stillbirths were associated with elevated fetal fibronectin levels at 20–24 weeks, (10.3% compared with 4.2 %, P=.020) but not at 28 weeks (5.9% compared with 4.9%, P=.737). Neonatal death was associated with elevated fetal fibronectin levels at 28 weeks (13.3% compared with 4.5%, P=.001), but not at 20 –24 weeks (5.4% compared with 4.2%, P=.571). Table 4 shows mean fetal fibronectin values related to a variety of preterm birth and birth weight outcomes at 20–24 weeks and at 28 weeks. There were no significant differences in mean fetal fibronectin levels for those with and without each outcome at 20–24 weeks. However, at 28 weeks, each outcome was associated with higher mean fetal fibronectin levels, and most differences were significant.
Table 5 presents the comparisons of pregnancy outcomes for those women with fetal fibronectin values less than 50 ng/mL compared with those with fetal fibronectin levels 50 ng/mL or more at both 20–24 weeks and at 28 weeks. At 20–24 weeks, there were no significant differences in any outcome. However, at 28 weeks, there were significant differences in every outcome, with the most impressive differences in outcome (usually greater than four-fold) in less than 32 weeks preterm birth, regardless of whether the gestational age was determined by last menstrual period, fundal height measurements, or neonatal Ballard scores. There was also a 199 g decrease in mean birth weight for those women with fetal fibronectin levels 50 ng/mL or more (P<.001) and a nearly five-fold increase in the percent of infants born weighing 1,500 g or more (P=.015). The time from randomization to delivery was also significantly decreased in women with high fetal fibronectin levels. To confirm that the relationships in Table 5 were not due to confounding, we ran a series of multivariate regression analyses, adjusting for a number of characteristics including site, maternal age, history of preterm birth, years of education, body mass index, and treatment group. At 28 weeks, the odds ratio (OR) for a fetal fibronectin value 50 ng/mL or more to be associated with preterm birth less than 32 weeks as determined by fundal height measurements at enrollment was 6.3 (95% confidence interval [CI]3.2–12.3), for preterm birth less than 35 weeks as determined by fundal height measurements, 3.4 (95% CI 2.2–5.3), and for preterm birth less than 37 weeks as determined by fundal height measurements, 2.3 (95% CI 1.5–3.3). The OR for a fetal fibronectin 50 ng/mL or more at 28 weeks to be associated with a birth weight less than 1,500 g was 5.0 (95% CI 1.6–15.1), and for birth weight less than 2,500 g, the OR was 2.7 (95% CI 1.7–4.2).
Table 6 presents selected outcomes at 20–24 weeks and 28 weeks based on specific fetal fibronectin ranges. In this analysis, at 20–24 weeks there were no significant differences in outcome as the fetal fibronectin levels increased. However, at 28 weeks, as the fetal fibronectin values increased, the outcomes generally worsened, and every test of trend was significant. However, there were few apparent differences in outcomes between women with values less than 20 ng/mL compared with values of 20–49 ng/mL, either at 20–24 weeks or at 28 weeks. Clearly, however, at 28 weeks, as the values rose above 50 ng/mL, the outcomes progressively worsened.
We also evaluated the relationship between various placental histologic findings and fetal fibronectin levels at 20–24 weeks and 28 weeks. In both periods neither marked PMN nor mononuclear cell infiltration into any site was associated with increasing fetal fibronectin levels. (Data not shown.)
Table 7 presents the relationship between fetal fibronectin levels and mother-to-child transmission of HIV. Cases of mother-to-child transmission of HIV are stratified according to presumed timing of transmission: infants with positive diagnostic assays at birth (presumed in utero transmission), infants with negative diagnostic assays at birth but positive assay results at 6 weeks (presumed intrapartum/early postnatal transmission), and overall (both in utero and intrapartum/early postnatal transmission). The positive association between fetal fibronectin and mother-to-child transmission of HIV at 20 to 24 weeks was relatively weak and the direction of the association was reversed in the later time period.
This study confirms many previous observations related to fetal fibronectin and adverse pregnancy outcomes1–13 and suggests that vaginal fetal fibronectin values at 28 weeks predict preterm birth approximately as well in low-resource countries and among HIV-infected women as they do in developed countries and among HIV-uninfected women. The observed association between elevated fetal fibronectin levels and neonatal death is likely explained by the increased preterm birth rate in those women.
Data from the National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network Preterm Prediction Study and others6,7 suggest that using a cutoff of 50 ng/mL at gestational age 24 weeks or more identifies women at high risk for preterm birth.1–7 These studies also suggest that there is little difference in preterm birth among women using any fetal fibronectin level cutoff less than 50 ng/mL, but several studies have shown that as fetal fibronectin values increase above 50 ng/mL, the preterm birth rate rises. This study confirms that observation.
We have previously observed that fetal fibronectin is a better predictor of preterm birth at or after 24 weeks gestational age than it is at earlier gestational ages.4,5 This study confirms that fetal fibronectin is a poor predictor of preterm birth at 20–24 weeks (mean gestational age 22 weeks) yet becomes a strong predictor at 28 weeks. The reason for this difference is unknown. However, we suspect that many elevated values at 20–24 weeks are due to a nonpathologic seepage of fetal fibronectin from the uterus to the vagina as a result of as yet incomplete attachment of the membranes to the decidua. Furthermore, we hypothesize that later seepage of fetal fibronectin to the cervix and vagina is often due to a pathologic process such as infection or bleeding at the choriodecidual interface (Vullaganti M, Goldenberg RL, Copper R. Fetal fibronectin at less than 20 weeks is associated with chorioamnionitis in preterm deliveries [meeting abstract]. Am J Obstet Gynecol 2000;182:S53).2,19,20
We were, therefore, surprised that there was no association between high fetal fibronectin levels and various placental histologic findings, especially in view of our earlier observations that high fetal fibronectin levels at 24 weeks were associated with subsequent clinical chorioamnionitis (Vullaganti et al),19,20 histologic chorioamnionitis, and neonatal sepsis. One explanation may be that the previous study was restricted to infants delivering preterm, whereas in the present study, most placentas were from term infants. The difference in gestational age at the time of fetal fibronectin testing may also play a role. For example, our previous studies would suggest that fetal fibronectin values obtained at less then 24 weeks are poorly associated with preterm birth,5 whereas after 24–26 weeks, the association between placental inflammation and preterm birth diminishes as gestational age progresses.4,20 It may be that had we collected the first fetal fibronectin sample at approximately 24–26 weeks, rather then at a mean of 22 weeks, the associations would have been stronger. On the other hand, it also may be that contrary to our hypotheses, high fetal fibronectin levels in the vagina, despite being a predictor of spontaneous preterm birth in this and many other studies, is not associated with, or caused by, chorioamnionitis.21
One previous study has suggested that the use of antibiotics does not affect fetal fibronectin levels.22 Our large randomized study comparing metronidazole and erythromycin treatment to a placebo at 20–24 weeks gestational age, strongly suggests that the use of these antibiotics have little effect on fetal fibronectin levels drawn approximately 4 weeks later.
The relationship between fetal fibronectin values and mother-to-child transmission of HIV was inconsistent, with a positive association at 20–24 weeks but with a negative association at 28 weeks. Because an elevated fetal fibronectin value may be explained by choriodecidual disruption caused either by bleeding or infection,2,19 it may be that these conditions explain both the increased fetal fibronectin values as well as an increased risk of mother-to-child transmission of HIV. We do not have an explanation for the difference in results with changes in gestational age, other than one or the other occurred by chance.
In summary, high vaginal fetal fibronectin levels at 28 weeks, but not at 20–24 weeks, were strongly associated with preterm birth–related outcomes and predicted these outcomes in an African setting much as they do in more developed areas. Surprisingly, fetal fibronectin values were not related to histologic chorioamnionitis and were not changed by antibiotic treatment, but were associated with maternal syphilis. The relationships between elevated fetal fibronectin levels at 20–24 weeks with mother-to-child transmission of HIV require further evaluation.
1. Sadovsky Y, Friedman SA. Fetal fibronectin and preterm labor [letter]. N Engl J Med 1992;326:709.
2. 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.
3. Lockwood CJ, Wein R, Lapinski R, Casal D, 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.
4. Goldenberg RL, Mercer BM, Meis PJ, Copper RL, Das A, McNellis D. The preterm prediction study: fetal fibronectin testing and spontaneous preterm birth. NICHD Maternal Fetal Medicine Units Network. Obstet Gynecol 1996;87:643–8.
5. Goldenberg RL, Klebanoff M, Carey JC, MacPherson C, Leveno KJ, Moawad AH, et al. Vaginal fetal fibronectin measurements from 8 to 22 weeks’ gestation and subsequent spontaneous preterm birth. Am J Obstet Gynecol 2000;183:469–75.
6. Lu GC, Goldenberg RL, Cliver SP, Kreaden US, Andrews WW. Vaginal fetal fibronectin levels and spontaneous preterm birth in symptomatic women. Obstet Gynecol 2001;97:225–8.
7. 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. The National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network. Am J Obstet Gynecol 2000;183:1480–3.
8. Goldenberg RL, Mercer BM, Iams JD, Moawad AH, Meis PJ, Das A, et al. The preterm prediction study: patterns of cervicovaginal fetal fibronectin as predictors of spontaneous preterm delivery. National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network. Am J Obstet Gynecol 1997;177:8–12.
9. 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.
10. Nageotte MP, Casal D, Senyei AE. Fetal fibronectin in patients at increased risk for premature birth. Am J Obstet Gynecol 1994;170:20–5.
11. Goldenberg RL, Iams JD, Mercer BM, Meis PJ, Moawad AH, Copper RL, et al. The Preterm Prediction Study: the value of new vs standard risk factors in predicting early and all spontaneous preterm births. NICHD MFMU Network. Am J Public Health 1998;88:233–8.
12. 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.
13. Iams JD, Casal D, McGregor 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.
14. Goldenberg RL, Mwatha A, Read J, Adeniyi-Jones S, Sinkala M, Msmanga G, et al. The HPTN 024 Study: the efficacy of antibiotics to prevent chorioamnionitis and preterm birth. Am J Obstet Gynecol 2006;194:650–61.
15. Taha TE, Brown ER, Hoffman IF, Fawzi W, Read JS, Sinkala M, et al. A phase III clinical trial of antibiotics to reduce chorioamnionitis-related perinatal HIV-1 transmission. AIDS 2006;20:1313–21.
16. Nugent RP, Krohn MA, Hillier SL. Reliability of diagnosing bacterial vaginosis is improved by a standardized method of gram stain interpretation. J Clin Microbiol 1991;29:297–301.
17. Goldenberg RL, Mudenda V, Read JS, Brown ER, Sinkala M, Kamiza S, et al. HPTN 024 study: histologic chorioamnionitis, antibiotics and adverse infant outcomes in a predominantly HIV-infected African population. Am J Obstet Gynecol 2006;195:1065–74.
18. Little RJ, Rubin DB. Statistical analysis with missing data. 2nd ed. Hoboken (NJ): John Wiley and Sons; 2000. p. 278–81.
19. Goldenberg RL, Thom E, Moawad AH, Johnson F, Roberts J, Caritis SN. The preterm prediction study: fetal fibronectin, bacterial vaginosis, and peripartum infection. NICHD Maternal Fetal Medicine Units Network. Obstet Gynecol 1996;87:656–60.
20. Goldenberg RL, Hauth JC, Andrews WW. Intrauterine infection and preterm delivery. N Engl J Med 2000;342:1500–7.
21. Yoon BH, Romero R, Moon JB, Oh SY, Han SY, Kim JC, et al. The frequency and clinical significance of intra-amniotic inflammation in patients with a positive cervical fetal fibronectin. Am J Obstet Gynecol 2001;185:1137–42.
© 2007 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.
22. Pastore LM, Royce RA, Jackson TP, Thorp JM Jr, Savitz DA, Kreaden US. Association between bacterial vaginosis and fetal fibronectin at 24–29 weeks’ gestation. Obstet Gynecol 1999;93:117–23.