Preterm birth is classified by the presence of uterine contractions and/or amniorrhexis at clinical presentation. This classification does not include prior cervical change. We hypothesized that the rate of cervical shortening before preterm birth would not differ according to clinical presentation.
We analyzed data from a completed study of paired cervical ultrasound measurements to test our hypothesis. Cervical ultrasound measurements obtained 4 weeks apart in the second trimester were related to gestational age and clinical presentation at birth.
Of 2521 eligible women, 128 were delivered after preterm labor and 106 after preterm membrane rupture; 89 delivered preterm for a medical or obstetrical indication; 2198 delivered at term. The rate of change was similar in women who presented with preterm labor (−0.96 mm/week) and preterm ruptured membranes (−0.82 mm/week).
Cervical shortening occurs at the same rate before spontaneous preterm birth, regardless of presentation.
Various systems have been proposed to classify preterm birth to understand its pathogenesis and provide a framework for clinical care. Classifications have been based on gestational age at birth (early vs late), clinical presentation (preterm labor vs preterm membrane rupture without labor vs cervical insufficiency vs medical-obstetrical complications), presumed cause (infection vs hemorrhage vs uterine stretch), and/or whether parturition was initiated spontaneously or iatrogenically.
Gestational age at birth and clinical presentation are widely used because they are most easily identified and most often recorded. It has been more difficult to identify the causes of preterm birth and to ascertain whether and when parturition has begun. Often clinical presentation has been used to infer potential causes and to mark the onset of parturition as well. Etiologic priority is commonly assigned to the signs and symptoms reported at presentation, but the duration of the parturitional process before that time is not well studied.
This deficit in the literature influences research as well as clinical care. Classifying preterm birth according to presenting signs and symptoms (eg, preterm labor, preterm ruptured membranes, or cervical insufficiency) suggests that each of these presentations is the initial step in the parturitional sequence. Thus, distinctions made between preterm births preceded by preterm labor, preterm ruptured membranes, or cervical insufficiency have led to separate lines of investigation and treatment for each: contraction detection and suppression for preterm labor, maintenance of membrane integrity and infection prophylaxis for preterm rupture, and restoration of cervical tissue strength for women with cervical insufficiency.
These strategies have not reduced the rate of preterm birth, yet the categories upon which they are based continue to drive basic and epidemiological thinking about preterm birth, sometimes despite what is known about the sequence of events in term and preterm parturition.
The parturitional process is currently understood as having 3 components: myometrial contractility, decidual activation, and cervical ripening. These processes overlap and may occur in any sequence but in normal parturition at term are thought to begin with cervical softening and ripening, followed by decidual activation and myometrial contractions.
The time at which contractions become regular is typically recorded as the point when disorganized uterine activity has coalesced into true labor. The time of membrane rupture is similarly accepted as evidence of decidual activation. Both are easily documented by physical examination and standard tests described in obstetrical texts, but their actual time of occurrence is uncertain. The onset of cervical ripening has been even more difficult to detect and measure. Investigation of this parturitional step has improved with the advent of animal models and cervical sonography.
Animal and human studies suggest that cervical preparation for birth begins soon after conception. Two phases of cervical change during pregnancy before labor have been described: softening and ripening. Softening occurs slowly over many weeks and is characterized by an increase in compliance with maintenance of tissue competence in a high progesterone/low estrogen environment. Ripening occurs more rapidly, over weeks or days preceding the onset of labor and is accompanied by loss of tissue compliance, decreased tensile strength, and reversal of the progesterone/estrogen ratio, after which cervical dilation occurs in response to contractions.
Observation of the cervix over time with transvaginal sonography offers an opportunity to relate cervical softening and ripening (called funneling or shortening when seen on ultrasound) to the gestational age and clinical circumstances preceding preterm birth and to determine the presence and duration of cervical ripening prior to clinical symptoms. An observational study in which cervical length measurements are masked would be ideal but difficult because clinicians today may feel compelled to react to the identification of short cervix by recommending an intervention. We therefore accessed clinical and sonographic data from such a study performed in 1992-1994 to compare the rate of cervical shortening in women who delivered at term and preterm, with the latter analyzed according to their clinical presentation after preterm labor, preterm membrane rupture, or a medical/obstetrical indication for preterm birth.
In this secondary analysis, our primary hypotheses were that the rate of change (ie, slope) in cervical length in women presenting with spontaneous preterm labor would not differ from the rate observed in women who delivered after preterm ruptured membranes and that both would differ from the slope observed in women with a medical or obstetrical indication for preterm birth or who delivered at term.
Accurate characterization of the sequence of events that precede spontaneous preterm birth might define 1 or more preterm birth phenotypes more accurately than is possible using only the terminal signs and symptoms.
Materials and Methods
The Preterm Prediction Study was a prospective multicenter observational study of risk factors for preterm birth conducted by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) Maternal-Fetal Medicine Units Research Network between 1992 and 1994. The goal of the study was to identify new markers of preterm birth. The study design, methods of analysis, and results have been previously reported. The de-identified data set is now available on request from the NICHD. The current analyses were conducted after the Biomedical Institutional Review Board at the Ohio State University concluded that this analysis of completely de-identified data did not require further review.
The Preterm Prediction Study enrolled women with singleton pregnancies who were selected to reflect the parity and race of women receiving prenatal care at each of 10 participating sites. Women with pregnancies complicated by placenta previa, cervical insufficiency, or fetal anomalies were not enrolled. Except for fetal death, major anomalies, membrane prolapse, regular painful contractions, advanced cervical dilation, and oligo- or polyhydramnios, all data collected for the study were not revealed to care providers or enrollees.
Subjects were followed up from enrollment before 24 weeks of gestation through delivery to gather clinical and laboratory data at a series of 4 study visits scheduled at 2 week intervals. Cervical sonography was performed by sonographers who were centrally certified by a single reviewer (J.D.I.). The cervical length was measured according to a protocol described previously at the initial visit (visit 1) at 22 0/7th to 24 6/7th weeks’ gestation and again 4 weeks later, between 26 0/7th and 28 6/7th weeks’ gestation at visit 3.
The primary outcome of the Preterm Prediction Study was a birth that followed spontaneous preterm labor (PTL) or preterm premature ruptured membranes (PPROM) before 35 0/7th weeks’ gestation. Spontaneous preterm births before 32 and 37 weeks were secondary outcomes. Labor was defined in the original protocol as progressive cervical dilation and effacement.
Preterm labor was defined as 6 or more documented uterine contractions per hour during the admission and 1 or more of the following: (1) ruptured membranes within 1 hour of onset of contractions; (2) documented cervical change; (3) cervical dilation 2 cm or greater internal os; and (4) cervical length of 1 cm or less or 50% or greater effacement, before 37 weeks’ gestation. Women who delivered before 37 weeks after presenting with spontaneous preterm labor or PPROM were classified as having a spontaneous preterm birth, including those with PPROM in whom labor was induced. Preterm births that followed the induction of labor for other reasons or cesarean section without labor that were performed for maternal or fetal benefit were recorded as indicated preterm births.
The current analysis was limited to women who completed both cervical ultrasound examinations (visits 1 and 3) within the protocol-specified gestational age intervals and for whom delivery outcome was available ( Figure 1 ). In this analysis, births after 37 0/7th weeks are defined as term, and births before 37 0/7th weeks are defined as preterm and, as in the original study, categorized as occurring after spontaneous preterm labor, after spontaneous preterm PROM, or because of a medical indication for preterm birth.