Fetal Fibronectin, Inflammatory Biomarkers, and Salivary Estriol Testing for Preterm Labor

Number: 0166

Policy

Aetna considers the fetal fibronectin (fFN) immunoassay test medically necessary for evaluating symptomatic pregnant women at high-risk for preterm delivery (see background section for selection criteria). Repeat fFN immunoassay test is considered medically necessary if members remain symptomatic two or more weeks after a previous negative test.

Aetna considers the fetal fibronectin test experimental and investigational for routine screening of the general obstetric population and for all other indications including high-risk women who are asymptomatic for preterm labor, for identifying optimal candidates for cerclage, and following insertion of a cervical cerclage because its effectiveness for these indications has not been established.

Aetna considers biomarkers of intra-uterine inflammation (in amniotic fluid) including angiogenin, C-reactive protein, cytokines (e.g., interleukin-1, interleukin-6, interleukin-8), maternal matrix metalloproteinase-9, microRNA expression in the cervix, procalcitonin, and tumor necrosis factor-alpha experimental and investigational for evaluating pregnant women at high-risk for preterm delivery because their effectiveness for this indication has not been established.

Aetna considers human chorionic gonadotrophin and phosphorylated insulin-like growth factor binding protein-1 (in cervico-vaginal fluid) experimental and investigational for evaluating pregnant women at high-risk for preterm delivery because their effectiveness for this indication has not been established.

Aetna considers salivary estriol (SalEst) test experimental and investigational because the test results are not available rapidly enough to assist in decisions concerning the immediate care of the member.

Aetna considers DNA methylation, ferritin, fructose bisphosphonate aldolase A, heat shock protein beta-1, peroxiredoxin-1, pyruvate kinase M1/M2, transferrin, uric acid and vimentin experimental and investigational as biomarkers for preterm labor because their effectiveness for this indication has not been established.

Aetna considers proteomic biomarkers (e.g., 14-3-3 protein sigma, annexin A5, inter-α-trypsin inhibitor heavy chain H4, protein S100-A8, and protein S100-A12) experimental and investigational for evaluating pregnant women at high-risk for preterm delivery because their effectiveness for this indication has not been established.

Aetna considers serum levels of angiopoietin-1 (Ang-1), angiopoietin-2 (Ang-2), and the Ang-1/Ang-2 ratio levels experimental and investigational for evaluating pregnant women at high-risk for preterm delivery because their effectiveness for this indication has not been established.

Aetna considers albumin/vitamin D-binding protein, beta-2 adrenoceptor genotyping, cervical phosphorylated insulin-like growth factor binding protein-1, as well as maternal tumor necrosis factor-α G308A polymorphism and interferon-γ A874T polymorphism experimental and investigational for evaluating pregnant women at high-risk for preterm delivery because their effectiveness for this indication has not been established.

Aetna considers placental alpha-microglobulin-1 (PAMG-1) and phosphorylated insulin-like growth factor binding protein 1 (pIGFBP-1) experimental and investigational for evaluating pregnant women at high-risk for preterm delivery because their effectiveness for this indication has not been established.

Aetna considers testing of maternal methylenetetrahydrofolate reductase polymorphisms experimental and investigational for prediction of preterm delivery because of insufficient evidence.

Background

This policy is in accordance with the American College of Obstetricians and Gynecologists (ACOG) Committee Opinion on Fetal Fibronectin.

Fetal fibronectin (fFN) assay has recently been approved by the Food and Drug Administration (FDA) for clinical use in identifying patients at risk for preterm delivery.  The fFN immunoassay is a qualitative test for the detection of fFN protein in cervico-vaginal secretions.

A number of studies have examined the value of the fFN test as a predictor of likelihood of preterm delivery in women with symptoms of preterm labor.  The data indicate that a negative test has a maximal negative predictive value of approximately 96 % for not delivering within the next 2 weeks, while a positive test has a 15 to 20 % positive predictive value for preterm delivery.  Despite these data, there have been no prospective interventional studies demonstrating a decrease in preterm deliveries or improved perinatal outcomes based on the knowledge of the results of this test.  No study has examined the efficacy of fFN on the incidence, morbidity, and mortality of preterm delivery.  However, there may be selected cases in which quickly available results may be helpful in assessing the patient's risk of preterm delivery allowing for an impact on clinical decisions.

In a Cochrane review on fFN testing for reducing the risk of preterm birth, Berghella and colleagues (2008) concluded that although fFN is commonly used in labor and delivery units to help in the management of women with symptoms of preterm labor, currently there is not sufficient evidence to recommend its use.  Since this review found an association between knowledge of fFN results and a lower incidence of preterm birth before 37 weeks, further research should be encouraged.

Klebanoff and associates (2008) examined if salivary progesterone (P) or estriol (E3) concentration at 16 to 20 weeks' gestation predicts preterm birth or the response to 17alpha-hydroxyprogesterone caproate (17OHPC) and if 17OHPC treatment affected the trajectory of salivary P and E3 as pregnancy progressed.  This was a secondary analysis of a clinical trial of 17OHPC to prevent preterm birth.  Baseline saliva was assayed for P and E3.  Weekly salivary samples were obtained from 40 women who received 17OHPC and 40 who received placebo in a multi-center randomized study of 17OHPC to prevent recurrent preterm delivery.  Both low and high baseline saliva P and E3 were associated with a slightly increased risk of preterm birth.  However, 17OHPC prevented preterm birth comparably, regardless of baseline salivary concentrations of P and E3.  Moreover, 17OHPC did not alter the trajectory of salivary P over pregnancy, but it significantly blunted the rise in salivary E3 as well as the rise in the E3/P ratio.  The authors concluded that 17OHPC flattened the trajectory of E3 in the second half of pregnancy, suggesting that the drug influences the fetoplacental unit.

Selection Criteria for Fetal Fibronectin (fFN) Immunoassay

According to the ACOG Committee on Obstetric Practice (1997), the fFN test is only appropriate for use in symptomatic pregnant women with all of the following characteristics:

• Amniotic membranes intact; and
• Cervical dilatation is minimal (less than 3 cm); and
• Sampling is performed no earlier than 24 weeks, 0 days and no later than 34 weeks, 6 days of gestation.

In addition, the fFN test is only useful if results are available quickly enough (generally considered to be under 4 hours) so that the results can assist in decisions concerning the immediate care of pregnant women.

Although a negative test appears to be useful in ruling out imminent preterm delivery (i.e., within 2 weeks), the clinical implications of a positive result have not been fully evaluated.

If the test is to be clinically useful, the results must be available from the laboratory in a timely manner (generally considered to be under 4 hours) so that the test can effect decisions concerning the immediate care of the patient.  This is in accordance with the ACOG Committee Opinion, which states that “[i]f the test is to be clinically useful, the results must be available from the laboratory in a timely manner.”

ACOG does not recommend the fFN test for screening asymptomatic women to determine risk of preterm delivery.  ACOG's Division of Practice Activities concluded that "this test is not recommended as a routine screening procedure for the general prenatal population".  A recent clinical trial of the fFN test in 108 women at low risk of preterm delivery concluded that bi-weekly fFN determinations in asymptomatic women between 24 and 34 weeks' gestation “are of limited clinical value for the prediction of preterm birth”.

Keeler and colleagues (2009) determined the relationship between fFN testing prior to ultrasound-indicated cerclage and obstetric outcome.  Singleton pregnancies between 18 and 24 weeks' gestation with an ultrasound-diagnosed short cervix (less than 25 mm) and funneling (greater than 25 %) of the chorio-amniotic membranes into the endocervical canal were analyzed.  The fFN testing was performed and patients were randomized to cerclage or no-cerclage.  Groups were stratified by fFN result.  Cerclage patients were compared with no-cerclage patients.  The primary outcome was delivery prior to 35 weeks' gestation.  Spontaneous preterm birth prior to 35 weeks' gestation occurred in 15 (44.1 %) fFN-positive-cerclage patients and 16 (55.2 %) fFN-positive no-cerclage patients (p = 0.45).  Similarly, it occurred in 16 (17.8 %) fFN-negative cerclage patients and 11 (17 %) fFN-no-cerclage patients (p = 0.99).  The authors concluded that fFN testing did not identify optimal candidates for cerclage.

Inflammatory biomarkers are being investigated as predictors of preterm birth.  Gedc and Ford (2010) stated that there is overwhelming evidence that intra-uterine infection and inflammation play an important role in the pathogenesis of spontaneous preterm labor, preterm prelabor rupture of the membranes and fetal injury resulting in long-term sequelae.  Early diagnosis of subclinical infection and inflammation may therefore aid clinicians institute interventions focusing on such adverse outcomes.  Biomarkers of intra-uterine inflammation (e.g., interleukin-6) although sensitive, are not specific.  Thus, decision to deliver remote from term because of intra-uterine infection and/or inflammation should be based on clinical signs and/or bacterial culture or Gram stain of amniotic fluid.  In patients with preterm contractions and intact membranes, the risk of delivery is 1 % within the week following a negative fFN in cervico-vaginal secretions.  This aids to decide if antenatal steroids should be administered to patients presenting with preterm contractions between 24 and 34 weeks' gestation.  Biomarkers in cervical secretions and amniotic fluid identify those who may benefit from cerclage when the cervix is shortened (less than 25 mm) and dilated in the second trimester.  The authors concluded that so far, few interventions utilizing inflammatory biomarkers have shown clinical benefit.  They noted that future efforts should focus on the quest for accurate biomarkers that can be obtained non-invasively and allow early prediction of subclinical disease to initiate appropriate risk-specific intervention.

In a case-control study nested in a large, prospective, multi-center cohort trial (n = 5,337), Kramer and colleagues (2010) examined the role of mid-trimester maternal plasma cytokines and C-reactive protein (CRP) as predictors of spontaneous preterm birth.  Cohort women had an interview, examination, and venipuncture at 24 to 26 weeks.  Frozen plasma samples in women with spontaneous preterm birth (n = 207) and approximately 2 term controls per case (n = 444) were analyzed using Luminex multi-analyte profiling technology.  Fresh placentas were fixed, stained, and blindly assessed for histological evidence of infection/inflammation, decidual vasculopathy, and infarction, and vaginal swabs were analyzed for bacterial vaginosis and fFN concentration.  High maternal matrix metalloproteinase-9 (MMP-9) concentration, but none of the other cytokines or CRP, was significantly associated with spontaneous preterm birth [adjusted odds ratio = 1.7 (1.1 to 2.4)] and showed a dose-response relation across quartiles.  No association was observed, however, between maternal MMP-9 and placental infection/inflammation, bacterial vaginosis, or vaginal fFN concentration.  The authors concluded that these findings require confirmation in future studies, but suggest that a systemic immune response implicating MMP-9 may have an etiologic role in spontaneous preterm birth.

Conde-Agudelo et al (2011) examined the accuracy of novel biomarkers to predict spontaneous preterm birth in women with singleton pregnancies and no symptoms of preterm labor.  Electronic searches in PubMed, Embase, Cinahl, Lilacs, and Medion, references of retrieved articles, and conference proceedings were carried out.  No language restrictions were applied.  Observational studies that evaluated the accuracy of biomarkers proposed in the last decade to predict spontaneous preterm birth in asymptomatic women were selected.  These researchers excluded studies in which biomarkers were evaluated in women with preterm labor.  Two reviewers independently extracted data on study characteristics, quality, and accuracy.  Data were arranged in 2 × 2 contingency tables and synthesised separately for spontaneous preterm birth before 32, 34, and 37 weeks of gestation.  They used bi-variate meta-analysis to estimate pooled sensitivities and specificities, and calculated likelihood ratios (LRs).  A total of 72 studies, including 89,786 women and evaluating 30 novel biomarkers, met the inclusion criteria.  Only 3 biomarkers (proteome profile and prolactin in cervicovaginal fluid, and matrix metalloproteinase-8 in amniotic fluid) had positive LRs greater than 10.  However, each of these biomarkers was evaluated in only 1 small study.  Four biomarkers had a moderate predictive accuracy (interleukin-6 and angiogenin [a potent inducer of neovascularization], in amniotic fluid; human chorionic gonadotrophin and phosphorylated insulin-like growth factor binding protein-1, in cervico-vaginal fluid).  The remaining biomarkers had low predictive accuracies.  The authors concluded that none of the biomarkers evaluated in this review meet the criteria to be considered a clinically useful test to predict spontaneous preterm birth.  They stated that further large, prospective cohort studies are needed to evaluate promising biomarkers such as a proteome profile in cervico-vaginal fluid.

The Royal College of Obstetricians and Gynaecologists’ clinical guideline on “Cervical cerclage” (RCOG, 2011) stated that fetal fibronectin testing following insertion of a cervical cerclage is not recommended.

Bamberg et al (2012) evaluated mid-trimester amniotic fluid concentrations of 3 major pro-inflammatory cytokines (interleukin 6 [IL-6], interleukin 8 [IL-8], and tumor necrosis factor-alpha [TNF-α]) in asymptomatic pregnancies with adverse outcomes.  A prospective follow-up study at the Charite University Hospital, Berlin, Germany of women with uncomplicated singleton pregnancies at 2nd trimester and amniocentesis was carried out.  Concentrations of IL-6, IL-8, and TNF-α were measured by enzyme-linked immunosorbent assay following amniotic fluid assessment by mid-trimester amniocentesis performed from gestation days 15 weeks 0 days up to 20 weeks 6 days.  Values from normal pregnancies were compared to those from pregnancies having adverse outcomes of spontaneous abortion, preterm delivery, pre-eclampsia, or eclampsia.  Main outcome measure IL-6, IL-8 and TNF-α in relation to adverse pregnancy outcome.  A total of 298 consecutive patients were evaluated.  Median patient age was 35 years (range of 19 to 43).  Controls consisted of 273 women who delivered without further complications after 37 weeks gestation.  The range values of IL-6, IL-8, and TNF-α in the control group were 4.9 to 2,620 pg/ml, 36.2 to 5,843 pg/ml, and 8.0 to 28.2 pg/ml, respectively.  Patients with adverse pregnancy outcome (n = 25) were classified into 3 groups:

1. spontaneous abortion group (n = 4),
2. preterm delivery group (n = 17), and
3. pre-eclampsia/eclampsia group (n = 4). 

There were no significant differences in IL-6, IL-8, and TNF-α between controls and study groups, regardless of the type of complication (p > 0.05).  The authors concluded that mid-trimester amniotic fluid concentrations of the pro-inflammatory cytokines IL-6, IL-8, and TNF-α are not predictive of adverse pregnancy outcome in terms of spontaneous abortion, preterm delivery or pre-eclampsia/eclampsia in this study population.

Galazis and colleagues (2013) noted that preterm birth (PTB) is a major cause of neonatal mortality and morbidity.  Women with polycystic ovary syndrome (PCOS) are at high-risk of PTB.  There is a need for research studies to investigate the mechanisms linking PCOS and PTB, to facilitate screening, and develop novel preventative strategies.  These researchers listed all the proteomic biomarkers of PTB and integrated this list with the PCOS biomarker database to identify commonly expressed biomarkers of the 2 conditions.  They carried out a systematic review of PTB biomarkers and update of PCOS biomarker database.  All eligible published studies on proteomic biomarkers for PTB and PCOS identified through various databases were evaluated.  For the identification of the relevant studies, the following search terms were used: "proteomics", "proteomic", "preterm birth", "preterm labour", "proteomic biomarker" and "polycystic ovary syndrome".  This search was restricted to humans only.  A database on proteomic biomarkers for PTB was created while an already existing PCOS biomarker database was updated.  The 2 databases were integrated and biomarkers that were co-expressed in both women with PCOS and PTB were identified and investigated.  A panel of 6 proteomic biomarkers was similarly differentially expressed in women with PTB and women with PCOS compared to their respective controls (normal age-matched women in the case of PCOS studies and women with term pregnancy in the case of PTB studies).  These biomarkers include pyruvate kinase M1/M2, vimentin, fructose bisphosphonate aldolase A, heat shock protein beta-1, peroxiredoxin-1 and transferrin.  The authors concluded that these proteomic biomarkers (pyruvate kinase M1/M2, vimentin, fructose bisphosphonate aldolase A, heat shock protein beta-1, peroxiredoxin-1 and transferrin) can be potentially used to better understand the pathophysiological mechanisms linking PCOS and PTB.  This would help to identify subgroups of women with PCOS at risk of PTB and hence the potential of developing preventative strategies.

Clowse et al (2013) stated that while increased disease activity is the best predictor of PTB in women with systemic lupus erythematosus (SLE), even women with low disease activity are at increased risk of this complication.  Biomarkers that would identify at-risk pregnancies could allow interventions to prevent PTB.  In this study, measures of SLE activity, inflammation, placental health and renal function between 20 and 28 weeks gestation (mid-gestation) were correlated to PTB and gestational age at delivery in a prospective cohort of pregnant women with SLE.  Of the 40 pregnancies in 39 women, all with mild-moderate SLE disease, 9 (23.7 %) of the 38 live births were delivered preterm.  Low C4 was the only marker of SLE activity associated with younger gestational age at delivery.  Elevated ferritin and lower estradiol correlated with younger gestational age at delivery.  Renal function remained normal during all pregnancies at mid-gestation and did not correlate with PTB.  Higher serum uric acid, however, correlated with younger gestational age at delivery.  The authors concluded that in women with SLE with mild-moderate disease activity, ferritin, estradiol and uric acid levels at mid-gestation may predict PTB.  They stated that these markers may prove to be clinically useful in identifying pregnancies at particularly high-risk for adverse outcomes.

Bastek and Elovitz (2013) stated that biomarkers associated with spontaneous PTB and pre-eclampsia have been discovered in patients who experience these adverse obstetrical outcomes.  The identification of such biomarkers holds promise in both facilitating the early identification of those patients at greatest risk and enhancing the understanding of these disease processes to determine therapeutic interventions.  These investigators reviewed the existing literature to determine the utility of biomarkers in the risk stratification of spontaneous PTB and pre-eclampsia.  They found that despite the promise of some biomarkers in identifying patients at increased risk for spontaneous PTB and/or pre-eclampsia, the use of biomarkers in clinical practice to predict adverse obstetrical outcome remains challenging.  Although data from small discovery studies may be encouraging, progress with biomarker research remains limited by the lack of validation of these discovered biomarkers.  Furthermore, owing to the heterogeneity of existing studies, generalizable conclusions are difficult to understand, meta-analyses are challenging to perform, and agreement on cut-point standardization is difficult.  The authors concluded that the identification of an abnormal biomarker level does not guarantee whether or when an adverse clinical event might occur.  The lack of understanding of the true etiologies of these disease processes resulting in the absence of definitive interventions to prevent spontaneous PTB and pre-eclampsia from occurring.

Schneuer et al (2014) evaluated angiopoietin-1 (Ang-1), angiopoietin-2 (Ang-2), and the Ang-1/Ang-2 ratio levels in the first trimester of pregnancy, their association with adverse pregnancy outcomes, and their predictive accuracy.  This cohort study measured serum Ang-1 and Ang-2 levels in 4,785 women with singleton pregnancies attending first trimester screening in New South Wales, Australia.  Multi-variate logistic regression models were used to assess the association and predictive accuracy of serum biomarkers with subsequent adverse pregnancy outcomes (small for gestational age, PTB, pre-eclampsia, miscarriage greater than 10 weeks, and stillbirth).  Median (interquartile range) levels for Ang-1, Ang-2, and the Ang-1/Ang-2 ratio for the total population were 19.6 ng/ml (13.6 to 26.4), 15.5 ng/ml (10.3 to 22.7), and 1.21 (0.83 to 1.73), respectively.  Maternal age, weight, country of birth, and socioeconomic status significantly affected Ang-1, Ang-2, and the Ang-1/Ang-2 ratio levels.  After adjusting for maternal and clinical risk factors, women with low Ang-2 levels (less than 10th percentile) and high Ang-1/Ang-2 ratio (greater than 90th percentile) had increased risk of developing most adverse pregnancy outcomes.  Compared with the Ang-1/Ang-2 ratio alone, maternal and clinical risk factors had better predictive accuracy for most adverse pregnancy outcomes.  The exception was miscarriage (Ang-1/Ang-2 ratio area under receiver operating characteristic curve = 0.70; maternal risk factors = 0.58).  Overall, adding the Ang-1/Ang-2 ratio to maternal risk factors did not improve the ability of the models to predict adverse pregnancy outcomes.  The authors concluded that these findings suggested that the Ang-1/Ang-2 ratio in first trimester is associated with most adverse pregnancy outcomes, but do not predict outcomes any better than clinical and maternal risk factor information.

Kacerovsky et al (2014) analyzed the findings of studies on proteomic biomarkers for spontaneous PTB.  Three electronic databases (Medline, Embase, and Scopus) were searched for studies in any language reporting the use of proteomic biomarkers for PTB published between January 1994 and December 2012.  Retrieved citations were screened, and relevant studies were selected for full-text reading, in triplicate.  The search yielded 529 citations, 51 were selected for full-text reading and 8 studies were included in the review.  A total of 64 dysregulated proteins were reported.  Only 14-3-3 protein sigma, annexin A5, protein S100-A8, protein S100-A12, and inter-α-trypsin inhibitor heavy chain H4 were reported in more than 1 study, but results could not be combined due to heterogeneity in type of sample and analytical platform.  The authors concluded that according to the existing literature, there are no specific proteomic biomarkers capable of accurately predicting PTB.

C-Reactive Protein and Procalcitonin

Dulay and colleagues (2015) stated that the arsenal of maternal and amniotic fluid (AF) immune response to local or systemic infection includes among others the acute-phase reactants IL-6, CRP and procalcitonin (PCT).  If these molecules can be used as non-invasive biomarkers of intra-amniotic infection (IAI) in the subclinical phase of the disease remains incompletely known.  These researchers used time-matched maternal serum, urine and AF from 100 pregnant women who had an amniocentesis to rule out IAI in the setting of preterm labor, preterm premature rupture of membranes (PPROM) or systemic inflammatory response (SIR: pyelonephritis, appendicitis, pneumonia) to infection.  Cord blood was analyzed in a subgroup of cases.  These investigators used sensitive immunoassays to quantify the levels of inflammatory markers in the maternal blood, urine and AF compartment.  Microbiological testing and placental pathology was used to establish infection and histological chorio-amnionitis.  Procalcitonin was not a useful biomarker of IAI in any of the studied compartments.  Maternal blood IL-6 and CRP levels were elevated in women with subclinical IAI.  Compared to clinically manifest chorio-amnionitis group, women with SIR have higher maternal blood IL-6 levels rendering some marginal diagnostic benefit for this condition.  Urine was not an useful biological sample for assessment of IAI using any of these 3 inflammatory biomarkers.  The authors concluded that in women with subclinical IAI, the large overlapping confidence intervals (CIs) and different cut-offs for the maternal blood levels of IL-6, CRP and PCT likely made interpretation of their absolute values difficult for clinical decision-making.

Furthermore, an UpToDate review on “Diagnosis of preterm labor and overview of preterm birth” (Lockwood, 2015) does not mention CRP and procalcitonin as diagnostic tools.

Interleukin-1

Nadeau-Vallee et al (2016) noted that PTB is a leading cause of neonatal mortality and morbidity worldwide, and represents a heavy economic and social burden.  Despite its broad etiology, PTB has been firmly linked to inflammatory processes.  Pro-inflammatory cytokines are produced in gestational tissues in response to stressors and can prematurely induce uterine activation, which precedes the onset of preterm labor.  Of all cytokines implicated, interleukin (IL)-1 has been largely studied, revealing a central role in preterm labor.  However, currently approved IL-1-targeting therapies have failed to show expected efficacy in pre-clinical studies of preterm labor.  The authors summarized animal and human studies in which IL-1 or IL-1-targeting therapeutics were implicated with preterm labor; focused on novel IL-1-targeting therapies and diagnostic tests; and developed the case for commercialization and translation means to hasten their development.

Also, an UpToDate review on “Diagnosis of preterm labor and overview of preterm birth” (Lockwood, 2015) does not mention IL-1 as a diagnostic tool.

MicroRNA Expression in the Cervix

Sanders et al (2015) stated that PTB is a leading cause of infant mortality and can lead to poor life-long health and adverse neurodevelopmental outcomes.  The pathophysiologic mechanisms that precede preterm labor remain elusive, and the role that epigenetic phenomena play is largely unstudied.  These researchers examined the association between microRNA (miRNA) expression levels in cervical cells obtained from swabs collected during pregnancy and the length of gestation.  They analyzed cervical samples obtained between 16 and 19 weeks of gestation from 53 women in a prospective cohort from Mexico City, and followed them until delivery.  Cervical miRNA was extracted and expression was quantified using the NanoString nCounter Analysis System.  Linear regression models were used to examine the association between miRNA expression levels and gestational age at delivery, adjusted for maternal age, education, parity, body mass index, smoke exposure, and inflammation assessed on a Papanicolaou smearr.  These investigators identified 6 miRNAs that were significantly associated with gestational age at the time of delivery, including miR-21, 30e, 142, 148b, 29b, and 223.  Notably, per each doubling in miR-21 expression, gestations were 0.9 (95 % confidence interval [CI]: 0.2 to 1.5) days shorter on average (p = 0.009).  Per each doubling in miR-30e, 142, 148b, 29b, and 223 expression, gestations were shorter by 1.0 to 1.6 days.  The predicted targets of the miRNAs were enriched for molecules involved in DNA replication and inflammatory processes.  The authors concluded that the levels of specific miRNAs in the human cervix during pregnancy are predictive of gestational age at delivery, and should be validated in future studies as potential biomarkers of preterm birth risk.

Elovitz et al (2015) examined if miRNA profiles in maternal blood are different in women who are destined to have a preterm, compared with a term, birth.  A nested case-control study was performed with maternal serum that was collected as part of a larger prospective cohort.  MiRNA expression profiles in maternal serum were compared between women who ultimately had a preterm birth (n = 40) compared with term birth (n = 40).  MiRNA expression profiles were created with the use of the Affymetrix GeneChip miRNA Array.  The data were analyzed with the significance of analysis of microarrays and principle components analyses.  A false discovery rate of 20 % was used to determine the most differentially expressed miRNAs.  Of the 5,640 miRNAs that were analyzed on the array, 4 miRNAs were significantly different between cases and control subjects; 2 of the 4 miRNAs were mature miRNAs.  The fold difference in expression was less than 2 for all 4 miRNAs.  The authors concluded that miRNA profiles in maternal blood were not significantly different in women who were destined to have a preterm, compared with a term, birth.  They stated that miRNAs in maternal blood are unlikely to become clinically useful biomarkers for the prediction of PTB.

Also, an UpToDate review on “Diagnosis of preterm labor and overview of preterm birth” (Lockwood, 2015) does not mention miRNAs diagnostic tools.

Albumin/Vitamin D-Binding Protein

In a retrospective, cohort study, Liong and colleagues (2015) identified cervico-vaginal fluid (CVF) biomarkers predictive of spontaneous PTB in women with symptoms of preterm labor.  Subjects were women with a singleton pregnancy admitted to the Emergency Department between 22 and 36 weeks of gestation presenting with symptoms of preterm labor.  Two-dimensional electrophoresis was used to analyze the CVF proteome.  Validation of putative biomarkers was performed using enzyme-linked immunosorbent assay (ELISA) in an independent cohort.  Optimal concentration thresholds of putative biomarkers were determined and the predictive efficacy for PTB was compared with that of fetal fibronectin.  Main outcome measure was prediction of spontaneous preterm labor within 7 days.  Differentially expressed proteins were identified by proteomic analysis in women presenting with “threatened” preterm labor without cervical change who subsequently delivered preterm (n = 12 women); ELISA validation using an independent cohort (n = 129 women) found albumin and vitamin D-binding protein (VDBP) to be significantly altered between women who subsequently experienced PTB and those who delivered at term.  Prediction of preterm delivery within 7 days using a dual biomarker model (albumin/VDBP) provided 66.7 % sensitivity, 100 % specificity, 100 % positive predictive value (PPV) and 96.7 % negative predictive value (NPV), compared with fetal fibronectin yielding 66.7, 87.9, 36.4 and 96.2 %, respectively (n = 64).  Using the maximum number of screened samples, the predictive utility of albumin/VDBP yielded a sensitivity of 77.8 %, specificity and PPV of 100 % and NPV of 98.0 % (n = 109).  The authors concluded that the dual biomarker model of albumin/VDBP is more effective than fetal fibronectin in predicting spontaneous preterm delivery in symptomatic women within 7 days.  Moreover, they stated that a clinical diagnostic trial is needed to test this model on a larger population to confirm these findings and to further refine the predictive values.

Beta-2 Adrenoceptor Genotyping

In a case-control study, Miller and associates (2015) examined if beta-2 adrenoceptor (β2 AR) genotype is associated with shortening of the cervix or with PTB risk among women with a short cervix in the 2nd trimester.  A total of 439 women, including 315 with short cervix and 124 with normal cervical length were included in this study.  Nulliparous women with cervical length less than 30 mm upon a 16- to 22-week trans-vaginal sonogram and controls frequency-matched for race/ethnicity with cervical lengths greater than or equal to 40 mm were studied; β2 AR genotype was determined at positions encoding for amino acid residues 16 and 27.  Genotype distributions were compared between case and control groups.  Within the short cervix group, pregnancy outcomes were compared by genotype, with a primary outcome of PTB less than 37 weeks.  Genotype data were available at position 16 for 433 women and at position 27 for 437.  Using a recessive model testing for association between short cervix and genotype, and adjusted for ethnicity, there was no statistical difference between cases and controls for Arg16 homozygosity (OR 0.7, 95 % CI: 0.4 to 1.3) or Gln27 homozygosity (OR 0.9, 95 % CI: 0.3 to 2.7).  Among cases, Arg16 homozygosity was not associated with protection from PTB or spontaneous PTB.  Gln27 homozygosity was not associated with PTB risk, although sample size was limited.  The authors concluded that β2 AR genotype did not appear to be associated with short cervical length or with PTB following the 2nd-trimester identification of a short cervix; influences on PTB associated with β2 AR genotype did not appear to involve a short cervix pathway.

Cervical Phosphorylated Insulin-Like Growth Factor Binding Protein-1

Conde-Agudelo and Romero (2016) evaluated the accuracy of the cervical phosphorylated insulin-like growth factor binding protein-1 (phIGFBP-1) test to predict PTB in women with and without symptoms of preterm labor through the use of formal methods for systematic reviews and meta-analytic techniques.  Data sources included PubMed, Embase, Cinahl, Lilacs, and Medion (all from inception to June 30, 2015), reference lists, conference proceedings, and Google scholar.  Study eligibility criteria were cohort or cross-sectional studies that reported on the predictive accuracy of the cervical phIGFBP-1 test for PTB.  Two reviewers selected studies, assessed the risk of bias, and extracted the data.  Summary receiver-operating characteristic curves, pooled sensitivities and specificities, and summary likelihood ratios were generated.  A total of 43 studies met the inclusion criteria, of which 15 provided data on asymptomatic women (n = 6,583) and 34 on women with an episode of preterm labor (n = 3,620).  Among asymptomatic women, the predictive accuracy of the cervical phIGFBP-1 test for PTB at less than 37, less than 34, and less than 32 weeks of gestation was minimal, with pooled sensitivities and specificities and summary positive and negative likelihood ratios ranging from 14 % to 47 %, 76 % to 93 %, 1.5 to 4.4, and 0.6 to 1.0, respectively.  Among women with an episode of preterm labor, the test had a low predictive performance for delivery within 7 and 14 days of testing, and PTB at less than 34 and less than 37 weeks of gestation with pooled sensitivities and specificities and summary positive and negative likelihood ratios that varied between 60 % and 68 %, 77 % and 81 %, 2.7 and 3.5, and 0.4 and 0.5, respectively.  A negative test result in women with an episode of preterm labor had a low-to-moderate accuracy to identify women who are not at risk for delivering within the next 48 hours (summary negative likelihood ratio of 0.28 in all women and 0.23 in women with singleton gestations).  The authors concluded that cervical phIGFBP-1 has the potential utility to identify patients with an episode of preterm labor who will not deliver within 48 hours.  However, its overall predictive ability for the identification of symptomatic and asymptomatic women at risk for PTB is limited.

Maternal Tumor Necrosis Factor-α G308A Polymorphism and Interferon-γ A874T Polymorphism

Liu and colleagues (2015) examined the association between tumor necrosis factor-α (TNF-α) G308A polymorphism and interferon-γ (INF-γ) A874T polymorphism and risk of PTB by performing a meta-analysis of available studies.  Articles were chosen based on PubMed, Embase, Web of science, and China Biology Medicine (CBM) databases with no language restriction from their inceptions to March 1, 2014.  Specific inclusion criteria were used to evaluate articles.  Meta-analysis was performed by using a random or fixed effect model with STATA 11.0 software.  These researchers estimated the summary odds ratios (ORs) with its corresponding 95 % CI to assess the association.  A total of 21 eligible case-control studies with 2,103 cases and 5,070 controls were finally included into this meta-analysis.  Pooled analysis showed that A allele of TNF-α G308A was not associated with increased PTB risk (OR = 0.84, 95 % CI: 0.65 to 1.07, p = 0.167 for G versus A).  Stratifying analysis for ethnicity and different definition of PTB also indicated that A allele was not associated with increased PTB risk.  However, the meta-analysis showed that INF-γ A874T polymorphism was associated with the increased risk of PTB (OR = 1.14, 95 % CI: 1.11 to 1.73, p = 0.004 for A versus T).  Stratifying analysis was not performed due to the small sample size.  The authors concluded that TNF-α G308A polymorphism was not associated with an increased risk of PTB, but INF-γ A874T polymorphism may contribute to increasing susceptibility to PTB.  Moreover, they stated that detection of polymorphism of INF-γ A874T might be a promising biomarker for the diagnosis and prognosis of preterm delivery.

DNA Methylation and Preterm Birth

Barcelona de Mendoza et al (2017) stated that the causes of many cases of PTB remain enigmatic.  Increased understanding of how epigenetic factors are associated with health outcomes has resulted in studies examining DNA methylation (DNAm) as a contributing factor to PTB.  However, few studies on PTB and DNAm have included African American women, the group with the highest rate of PTB.  These researchers analyzed the existing studies on DNAm and PTB among African American women.  Studies (n = 10) were limited by small sample size, cross-sectional study designs, inconsistent methodologies for epigenomic analysis, and evaluation of different tissue types across studies.  African Americans comprised less than 50 % of the sample in 50 % of the studies reviewed.  Despite these limitations, there was evidence for an association between DNAm patterns and PTB.  The authors concluded that future research on DNAm patterns and PTB should use longitudinal study designs, repeated DNAm testing, and a clinically relevant definition of PTB and should include large samples of high-risk African American women to better understand the mechanisms for PTB in this population.

Placental Alpha-Microglobulin-1 (PAMG-1) and Phosphorylated Insulin-Like Growth Factor Binding Protein 1 (pIGFBP-1)

Fuchs and colleagues (2017) examined the accuracy of cervical phosphorylated insulin-like growth factor binding protein-1 (phIGFBP-1) test alone or in combination with cervical length (CL), to predict PTB in symptomatic women.  These researchers performed a prospective cohort study from 2012 to 2015 including singleton pregnancies with symptoms of preterm labor, intact membranes and CL of less than 25 mm at 24 to 34 weeks of gestation.  Studied outcome were spontaneous delivery within 7 and 14 days of testing and spontaneous PTB at less than 34 and less than 37 weeks of gestation.  Among 180 women, 21 (11.7 %) had a positive phIGFBP-1 test.  Spontaneous PTB occurred within 7 days, 14 days of testing and before 34 weeks and 37 weeks in 7.8 %, 10.6 %, 12.9 % and 28.8 %, respectively.  The phIGFBP-1 test had a low predictive performance for all studied outcomes varying for positive likelihood ratios (2.8 to 3.4) and negative likelihood ratios (0.8).  Combining phIGFBP-1 and CL did not increase its predictive ability.  After adjustment, positive phIGFBP-1 test was no more independently associated with a delivery within 7 days (p = 0.55), unlike CL of less than 15 mm (p = 0.04).  The authors concluded that phIGFBP-1 test alone or in combination with CL had a low predictive accuracy to predict PTB in symptomatic women.

An UpToDate review on “Diagnosis of preterm labor” (Lockwood, 2017) states that “Like fFN, placental α-microglobulin-1 (PAMG-1) or phosphorylated insulin-like growth factor binding protein 1 (pIGFBP-1) in vaginal or cervical secretions suggests disruption of the fetal membranes (ROM or labor) and are potential markers of an increased risk of preterm birth.  However, the utility of these tests has not been validated in either large or randomized clinical trials.  In the largest study, which included 796 women with signs and symptoms of preterm labor, the sensitivities of PAMG-1 and fFN for spontaneous preterm birth within 7 days were 50 % (3/6) and 67 %(4/6), respectively, and specificities were 98.4 % (619/629) and 85.7 %539/629), respectively”.

In a systematic review and meta-analysis, Melchor and colleagues (2018) evaluated the accuracy of PAMG-1, fFN and phIGFBP-1 tests in predicting spontaneous PTB (sPTB) within 7 days of testing in women with symptoms of preterm labor.  The test performance of each biomarker was also assessed according to pre-test probability of sPTB less than or equal to 7 days.  The Cochrane, Medline, PubMed and ResearchGate bibliographic databases were searched from inception until October 2017.  Cohort studies that reported on the predictive accuracy of PAMG-1, fFN and phIGFBP-1 for the prediction of sPTB within 7 days of testing in women with symptoms of preterm labor were included.  Summary receiver-operating characteristics (ROC) curves and sensitivity, specificity, PPV, NPV and LR+ and LR- were generated using indirect methods for the calculation of pooled effect sizes with a bi-variate linear mixed model for the logit of sensitivity and specificity, with each diagnostic test as a co-variate, as described by the Cochrane Handbook for Systematic Reviews of Diagnostic Test Accuracy.  Bi-variate mixed model pooled sensitivity of PAMG-1, fFN and phIGFBP-1 for the prediction of sPTB less than or equal to 7 days was 76 % (95 % CI: 57 to 89 %), 58 % (95 % CI: 47 to 68 %) and 93 % (95 % CI: 88 to 96 %), respectively; pooled specificity was 97 % (95 % CI: 95 to 98 %), 84 % (95 % CI: 81 to 87 %) and 76 % (95 % CI: 70 to 80 %) respectively; pooled PPV was 76.3 % (95 % CI: 69 to 84 %) (p < 0.05), 34.1 % (95 % CI: 29 to 39 %) and 35.2 % (95 % CI: 31 to 40 %), respectively; pooled NPV was 96.6 % (95 % CI: 94 to 99 %), 93.3 % (95 % CI: 92 to 95 %) and 98.7 % (95 % CI: 98 to 99 %), respectively; pooled LR+ was 22.51 (95 % CI: 15.09 to 33.60) (p < 0.05), 3.63 (95 % CI: 2.93 to 4.50) and 3.80 (95 % CI: 3.11 to 4.66), respectively; and pooled LR- was 0.24 (95 % CI: 0.12 to 0.48) (p < 0.05), 0.50 (95 % CI: 0.39 to 0.64) and 0.09 (95 % CI: 0.05 to 0.16), respectively.  The areas under the ROC curves for PAMG-1, fFN and phIGFBP-1 for sPTB less than or equal to 7 days were 0.961, 0.874 and 0.801, respectively.  The authors concluded that in the prediction of sPTB within 7 days of testing in women with signs and symptoms of preterm labor, the PPV of PAMG-1 was significantly higher than that of phIGFBP-1 or fFN.  Other diagnostic accuracy measures did not differ between the 3 biomarker tests.  As prevalence affects the predictive performance of a diagnostic test, use of a highly specific assay for a lower-prevalence syndrome such as sPTB may optimize management.

The authors stated that this study had several drawbacks.  First, it may be under-powered, as these researchers were unable to attain convergence in the low‐risk and intermediate‐risk groups.  Thus, the model had to be simplified to include random study effects only, without random effects for sensitivity and specificity by study.  Additionally, the 3 risk groups showed variation in sensitivity and specificity.  These metrics should not be dependent on the prevalence of the disease, as compared to predictive values.  This heterogeneity could be due to within‐study error and/or other population characteristics.  The reason for variation across populations should be further investigated.  Second, the 3 biomarkers were not studied in the same number of subjects.  Furthermore, only 45 % of studies were considered to have a low risk of bias.  Studies that did not specify the definition of PTB as spontaneous were ranked as having higher risk of bias as they may have included medically indicated deliveries.  This effect was expected to be minimal, given that few studies fell into this group, and more than 2/3 of PTBs were spontaneous.  Given the complexity and heterogeneity of available studies, the performance end-points were limited to sPTB less than or equal to 7 days of testing, as not all studies included information for other end-points, such as sPTB within 48 hours or 14 days of testing.  Alternative analysis methods may produce different pooled estimates from those found in the present study.  A simple pooled estimate may be obtained directly by summing numerators and denominators from the raw data across studies, but had severe limitations as it assumed lack of heterogeneity across the studies.  Therefore, random‐effects models are needed to describe variability in test accuracy across studies, which in addition assumes independence of sensitivity and specificity.  The bi-variate mixed‐effects model of Reitsma et al (2005) is a widely accepted method for meta‐analysis of diagnostic tests, as it overcomes the limitations of simple pooling by jointly modeling sensitivity and specificity.  Apart from accounting for between‐study variability due to both random error and inherent differences between the studies owing to different design, population or study procedures, the bi-variate mixed‐effects model method also allowed for inclusion of co-variates.  Finally, this review did not examine the impact of test performance on patient management and resource economics.  This is an area for future research.  Additionally, the heterogeneity of each test's performance using individual‐patient data and meta‐regression techniques looking at co-variates such as gestational age, contraction frequency and CL at presentation should also be examined in future studies.

Furthermore, an UpToDate review on “Preterm labor: Clinical findings, diagnostic evaluation, and initial treatment” (Lockwood, 2018) states that “Like fFN, placental α-microglobulin-1 (PAMG-1) or phosphorylated insulin-like growth factor binding protein 1 (pIGFBP-1) in vaginal or cervical secretions suggests disruption of the fetal membranes (ROM or labor) and are potential markers of an increased risk of preterm birth.  However, the utility of these tests has not been validated in either large or randomized clinical trials.  In the largest study, which included 796 women with signs and symptoms of preterm labor, the sensitivities of PAMG-1 and fFN for spontaneous preterm birth within 7 days were 50 %(3/6) and 67 % (4/6), respectively, and specificities were 98.4 % (619/629) and 85.7 % (539/629), respectively”.

Maternal Methylenetetrahydrofolate Reductase Polymorphisms and Preterm Delivery

Wu and associates (2017) examined the association of maternal and neonatal methylenetetrahydrofolate reductase (MTHFR) C677T polymorphisms with PTB and low birth weight (LBW) susceptibility, respectively.  These investigators carried out a systematic search of Embase, Medline, China Biological Medicine (CBN) Database, Chinese National Knowledge Infrastructure (CNKI), and Wanfang Database before June, 2016.  The frequencies of maternal and neonatal MTHFR C677T genotypes in the cases and controls and other information were extracted by 2 independent investigators; ORs with 95 % CIs were adopted to estimate the relationships between MTHFR C677T polymorphisms and PTB as well as LBW by random or fixed effect models.  A total of 25 studies from 20 articles concerning maternal and neonatal MTHFR C677T gene polymorphism with PTB and LBW were included in this study.  Maternal MTHFR C677T polymorphism was associated with PTB risk under allele contrast (T versus C, OR = 1.36, 95 % CI: 1.02 to 1.81), homozygote (TT versus CC, OR = 1.70, 95 % CI: 1.07 to 2.68), and recessive (TT versus CT + CC, OR = 1.49, 95 % CI: 1.00 to 2.22) model, but not dominant or heterozygote model.  Maternal MTHFR C677T polymorphism was also associated with LBW risk under allele contrast (OR = 1.69, 95 % CI: 1.25 to 2.28), homozygote (OR = 2.26, 95 % CI: 1.44 to 3.54), dominant (OR = 1.71, 95 % CI: 1.19 to 2.47), recessive (OR = 1.79, 95 % CI: 1.42 to 2.26) model, but not heterozygote model.  No associations between neonatal MTHFR C677T polymorphism and PTB or LBW were found under all genetic models.  The authors concluded that identification of maternal MTHFR C677T mutation may play a key role for primary prevention of PTB as well as LBW and screening pregnant women of high risk in developing countries.

Fang and colleagues (2018) noted that reported associations between MTHFR gene polymorphisms and preterm delivery are conflicting.  In a  meta-analysis, these investigators summarized the existing evidence and evaluated these associations.  Eligible studies were retrieved from Medline, Embase, the CBM Database and the Cochrane Library.  They calculated pooled ORs and 95 % CIs within 5 genetic models using either random-effects or fixed-effects models dependent on study heterogeneity.  Potential publication bias was assessed using a Begg's test; sensitivity analysis was performed to evaluate the stability of the results.  A total of 13 studies involving 4,816 mothers who experienced preterm delivery and 34,506 normal controls were included.  Significant associations between MTHFR C677T polymorphism and the risk of preterm delivery were detected overall (ORT/C = 1.34, 95 % CI: 1.12 to 1.61; ORTT/CC = 1.60, 95 % CI: 1.21 to 2.11; ORCT/CC = 1.33, 95 % CI: 1.07 to 1.65; ORTT/(CC + CT) = 1.41, 95% CI 1.11-1.78; OR(TT + CT)/CC  = 1.36, 95% CI 1.11-1.66) and in an Asian population (ORT/C  = 1.80, 95% CI 1.24-2.62; ORTT/CC  = 2.13, 95% CI 1.27-3.57; ORCT/CC  = 1.93, 95% CI 1.37-2.71; OR(TT + CT)/CC = 2.03, 95 % CI: 1.49 to 2.77).  Negative associations of the A1298C polymorphism were only observed among Asian pregnant women (ORC/A = 0.66, 95 % CI: 0.50 to 0.88; ORCC/AA = 0.10, 95 % CI: 0.02 to 0.53; ORCC/(AA + AC) = 0.11, 95 % CI: 0.02 to 0.57; OR(CC + AC)/AA = 0.68, 95 % CI: 0.49 to 0.94).  The authors concluded that MTHFR 677 T may play a significant role in regard to the risk of preterm delivery, especially in the Asian population.

Furthermore, an UpToDate review on “Preterm labor: Clinical findings, diagnostic evaluation, and initial treatment” (Lockwood, 2018) does not mention MTHFR gene polymorphisms.

Table: CPT Codes / HCPCS Codes / ICD-10 Codes

Code	Code Description
Information in the [brackets] below has been added for clarification purposes.   Codes requiring a 7th character are represented by "+":
CPT codes covered if selection criteria are met:
82731	Fetal fibronectin, cervicovaginal secretions, semi-quantitative [not covered following insertion of a cervical cerclage]
CPT codes not covered for indications listed in the CPB:
0066U	Placental alpha-micro globulin-1 (PAMG-1), immunoassay with direct optical observation, cervico-vaginal fluid, each specimen
81291	MTHFR (5,10-methylenetetrahydrofolate reductase) (eg, hereditary hypercoagulability) gene analysis, common variants (eg, 677T, 1298C)
81401	Molecular pathology procedure, Level 2 (eg, 2-10 SNPs, 1 methylated variant, or 1 somatic variant [typically using nonsequencing target variant analysis], or detection of a dynamic mutation disorder/triple repeat)
82085	Aldolase [fructose bisphosphonate aldolase A]
82677	Estriol
82728	Ferritin
83006	Growth stimulation expressed gene 2 (ST2, Interleukin 1 receptor like-1)
83516	Immunoassay for analyte other than infectious agent antibody or infectious agent antigen; qualitative or semiquantitative, multiple step method
83518	qualitative or semiquantitative, single step method
83519	qualitative by radioimmunoassay (e.g., RIA)
83520	qualitative, not otherwise specified
84112	Placental alpha microglobulin-1 [PAMG-1], cervicovaginal secretion, qualitative)
84145	Procalcitonin (PCT)
84210	Pyruvate [pyruvate kinase M1/M2]
84466	Transferrin
84550	Uric acid; blood
86140 - 86141	C-reactive protein
88324	Immunohistochemistry or immunocytochemistry, each separately identifiable antibody per block, cytologic preparation, or hematologic smear
Other CPT codes related to this CPB:
59320	Cerclage of cervix, during pregnancy; vaginal
HCPCS codes not covered if selection criteria are met:
S3652	Saliva test, hormone level; to assess preterm labor risk
ICD-10 codes covered if selection criteria are met [not covered for biomarkers of intra-uterine inflammation such as cytokines or maternal matrix metalloproteinase-9]:
N88.3	Incompetence of cervix uteri
O09.211 - O09.219	Supervision of pregnancy with history of pre-term labor
O34.30 - O34.33	Maternal care for cervical incompetence
O47.00 - O47.9	False [threatened] labor
O60.00 - O60.03	Preterm labor without delivery
Z87.51	Personal history of pre-term labor

The above policy is based on the following references: