Prediction of Preterm Delivery with Serial Cervical Length Measurements After Threatened Preterm Labor

Overview

Prospective cohort study on pregnant women discharged from the hospital after the first episode of threatened preterm labor. Cervical length (CL) will be measured with transvaginal US upon initial presentation (i.e at the time of hospital admission), at the time of hospital discharge, and respectively 2, 4, 8 and 12 weeks later. Pregnant women undelivered after the 1st episode of threatened preterm labor will be invited to participate in the study if CL upon discharge is \< 25 mm. The study will investigate the potential association between cervical shortening over time and time of delivery, to assess if spontaneous preterm delivery can be predicted by CL.

Preterm labor is the result of labor occurring prior to 36 6/7 weeks' gestation (i.e spontaneous preterm labor) leading to delivery. Preterm labor is a clinical diagnosis, defined as the appearance of regular contractions accompanied by a change in cervical dilation, effacement, or both, or initial presentation with regular contractions and cervical dilation of at least 2 cm prior to 36 6/7 weeks' gestation. The pathogenesis of preterm labor has not yet been fully understood, but it has been hypothesized that it could be caused by early idiopathic activation of the normal process of labor or instead, that it could be the result of pathological insults (Goldenberg 2008, Hamilton 2013, Slattery 2002). In Western countries, spontaneous PTB is responsible for 75% of perinatal mortality and more than half of long-term morbidity (McCormick 1985). Although most preterm newborns survive, they have an increased risk of neuro-behavioral disability, cerebral palsy, mental and developmental retardation. Moreover, spontaneous PTB also represents an important cause of chronic broncho pneumopathies, visual and auditory abnormalities and other chronic diseases of the pediatric age including school failure. As contractions are the most commonly recognized antecedent of preterm birth, cessation of uterine contractions has been the primary focus of therapeutic intervention to prevent the burden of prematurity. Preterm labor is currently considered a syndrome with a multifactorial etiology, among which inflammation, uteroplacental ischemia or hemorrhage, uterine overdistension, stress and other immune-mediated processes certainly play a role (Romero 2006): these factors do not necessarily represent a direct cause, but contribute to a transition from uterine quiescence to preterm labor or preterm premature rupture of membranes (pPROM) (Goldenberg 2005). There are many maternal and fetal conditions that have been associated with PTB, including maternal demographic characteristics (such as low socio-economic status, race), nutritional status, obstetric history (history of prior spontaneous PTB), psychological aspects, risk behaviors (such as drug and tobacco use), infections, short CL, as well as biological and genetic markers (Goldenberg 2008, Andrews 2000). Cervical length is measured with a 7-9 MHz transvaginal probe by trained obstetricians-gynecologists following the same standardized approach (Berghella 2003). Short CL (\< 20 mm) prior to 24 weeks' gestation has been recognized as a risk factor for preterm delivery among asymptomatic women; furthermore, short CL (\< 25 mm) has also been identified as a risk factor for spontaneous preterm delivery among patients presenting with painful uterine contractions (ACOG 2012). Threatened preterm labor (TPL) is the most common diagnosis that leads to hospitalization during pregnancy (Bennet 1998) and its annual costs exceed $800 million in the USA (Nicholson 2000). The association between preterm labor and preterm delivery is weak, as less than 10% of women diagnosed preterm labor actually give birth within 7 days of presentation (Fucks 2004). Therefore, clinicians define as threatened preterm labor a condition characterized by uterine contractions associated with cervical changes, that does not lead to preterm delivery, due a spontaneous resolution of the labor process, interventions aimed at possible causes of the labor process (such as treatment of a concomitant infection), or symptomatic therapeutic interventions (such as administration of tocolytics). Although approximately 75% of women presenting with TPL remain undelivered after a 48-hours course of tocolytics, their risk of preterm delivery remains high, as approximately 30% of them deliver before 37 weeks (Simhan 2007). However, it is a challenge for clinicians to identify who will deliver prematurely after an episode of threatened preterm labor. Cervical length (CL) measured by transvaginal sonography has been shown to be an accurate predictor of spontaneous preterm delivery among women presenting with painful uterine contractions (Berghella 2017, Sotiriadis 2016). However, studies have mainly focused on the role of a single CL measure collected at the time of the initial presentation instead of evaluating the predictive value of serial CL evaluations as threatened preterm labor resolves. We have previously showed in a secondary analysis of a RCT that serial CL evaluations may help identify those who will deliver prematurely after their first episode of threatened preterm labor (Chiossi 2020). We now intend to validate these findings prospectively, with a cohort study on the role of serial CL measures among women who remained undelivered after their first episode threatened preterm labor. Transvaginal CL measure has the potential to promptly identify shortening of the uterine cervix, a condition associated with preterm delivery. Such evaluation could be included in the antenatal care of women discharged from the hospital after an episode of threatened preterm labor to stratify their risk of preterm birth, to rationalize care and resource-utilization, as well as to improve pregnancy outcome. Pregnant women admitted to the hospital with the diagnosis of threatened preterm labor (regular contractions accompanied a change in cervical dilation, effacement, or both, or initial presentation with regular contractions and cervical dilation of at least 2 cm), who remained undelivered, will be invited to participate if CL upon discharge is \< 25 mm. Cervical length will be evaluated at follow up appointments 2, 4, 8 and 12 weeks after discharge and linked to the actual gestational age at delivery. Using a CL threshold of 10 mm we will estimate the risk of preterm delivery at different time points: each risk assessment will be based on the CL measured from hospital admission up to that time. At hospital discharge, patients will be classified at "low risk" of preterm delivery if their CL will be \> 10 mm, while "high risk" subjects will have a shorter CL. The risk of preterm birth will then be assessed on all the women who will remain undelivered until the 1st follow up appointment, 2 weeks later: pregnancies with CL measurements \> 10 mm from hospital discharge up to 2 weeks later will be considered "low risk", while those with CL \< 10 mm on at least one assessment will be classified as "high risk". If delivery will not occur after 4 weeks (2nd follow up visit), pregnancies who will maintain CL measurement \> 10 mm from hospital admission through the following 2 visits will be classified as "low risk", while "high risk" women will have a shorter CL on at least one of the 4 assessments. Risk estimates will be calculated in a similar fashion on women still pregnant 8 and 12 weeks after hospital discharge (3rd and 4th follow up visits). Cervical length will also be evaluated upon hospital admission, and the risk of preterm delivery will also be assessed taking into account such value. Descriptive statistics will be used to characterize the socio-demographic features and the obstetric characteristics of women categorized as "low risk" (i.e pregnancies whose CL persistently remains above 10 mm) or high risk (i.e those with at least one CL measurement \< 10 mm): categorical variables will be presented as absolute and percentage frequencies, they will be tested with Chi square test or Fisher's exact test as appropriate. Continuous variables will be summarized as mean +/- SD and compared with Student's t test. A level of statistical significance of p \< 0.05 will be considered. To control for confounding, multivariate logistic regression models will be used to describe the association between delivery prior to 37 weeks' gestation and CL (measured from hospital admission to each subsequent evaluation). Socio-demographic variables such as maternal age, BMI at entry to care, ethnicity, smoking, and education will be considered as potential confounders. Similarly, obstetric features such as parity, tocolysis at the initial hospitalization or at any subsequent hospital admission, urine culture collected on the initial hospitalization, and gestational age at enrolment will also be tested as potential.