Chapter 28: SONOGRAPHIC EXAMINATION OF THE UTERINE CERVIX

The uterine cervix plays a central role in the maintenance of normal pregnancy and in parturition. Thus, cervical disorders have been implicated in common obstetrical complications, such as "cervical insufficiency,"¹ preterm labor, and abnormal term parturition.² Yet, there is an incomplete understanding of the physiology and pathology of untimely cervical effacement and dilation during pregnancy. Mid-trimester cervical dilation is a major diagnostic and therapeutic challenge and a subject of intense debate among clinicians and researchers.

During most of normal pregnancy, the cervix remains firm and closed, despite a progressive increase in the size of the fetus and uterine distention. At the end of pregnancy and during labor, the cervix changes consistency (softens), shortens (effaces), and dilates to allow the expulsion of the conceptus. Labor, delivery, and the postpartum period are accompanied by dramatic changes in the uterine cervix.^{3,4,5,6,7,8,9, and 10} The term cervical ripening refers to the anatomic, biophysical, and biochemical processes that underlie the changes in cervical consistency, effacement, and dilation that generally precede the onset of spontaneous labor. Contrary to what was believed for many years, cervical ripening is an active metabolic process affecting the extracellular matrix components of the cervix. These changes increase cervical compliance. Untimely cervical ripening could result in complications of pregnancy. For example, failure of the cervix to ripen before myometrial activation at term (ie, onset of increased uterine contractility) may be the cause of a prolonged latent phase of labor; preterm premature cervical ripening may lead to mid-trimester spontaneous abortion or spontaneous preterm labor and delivery.

The uterine cervix is essentially a connective tissue organ, with smooth muscle cells accounting for only 10% to 15% of the distal part of the cervix.¹¹ The ability of the cervix to retain the conceptus during pregnancy is unlikely to depend upon a traditional sphincteric mechanism. Indeed, perfusion of strips of human cervix with vasopressin, a hormone that stimulates smooth muscle contraction, results in a very modest contractile response in comparison to that induced by this hormone in strips from the uterine isthmus and the fundus, which contain more muscle.¹²

It is now understood that the normal function of the uterine cervix during pregnancy depends largely upon the regulation of connective tissue metabolism. This tissue is formed by abundant extracellular matrix that surrounds individual cells. The major components of the extracellular matrix are collagen, proteoaminoglycans, elastin, and various glycoproteins such as fibronectins.¹¹ Collagen is considered the most important component of the extracellular matrix, determining the tensile strength of fibrous connective tissue. Changes in cervical characteristics during pregnancy have been attributed to changes in collagen content and metabolism.¹³ Proteoaminoglycans have also been implicated in cervical physiology. The proteoaminoglycan decorin has a high affinity for collagen and can cover the surface of the collagen fibrils, stabilizing them and promoting the formation of thicker collagen bundles or fibers. The predominant proteoglycan in the nonpregnant state is decorin and in the pregnant state is biglycan.¹⁴

Cervical Ripening

The biochemical events that have been implicated in cervical change during pregnancy and labor and delivery include (1) decreased collagen fiber alignment, (2) increased collagen solubility, (3) increased collagenolytic activity (both collagenase and leukocyte elastase), (4) increased interleukins (IL-1, IL-8, tumor necrosis factor α), and (5) increased matrix metalloproteinases [MMPs] (matrix degrading enzymes)—MMP-1 and MMP-8 when compared to nonripe cervices.^6,7,11,15

Petersen and Uldbjerg¹⁶ compared the cervical collagen content in nonpregnant patients with a history of "cervical incompetence" to normal nonpregnant women. Cervical incompetence was defined as "patients where no other cause of cervical incompetence could be deduced from the history and where the first pregnancy sustained beyond the 12th week of pregnancy ended with a late spontaneous abortion or preterm delivery." Patients with "cervical incompetence" by history had a markedly lower cervical hydroxyproline concentration and higher extractability (thought to represent collagen concentration solubility) when compared to normal parous patients. In addition, parous women had significantly lower hydroxyproline concentrations in the cervix when compared to nulligravidae and nulliparous women. The authors also identified a cutoff level of 13 mg of hydroxyproline/mg–wet weight, with a sensitivity of 0.9 and a specificity of 0.8, for the identification of women with "cervical incompetence."¹⁶

Liggins first likened cervical ripening to an inflammatory process.¹⁷ This concept has influenced research in cervical biology. However, recent observations have challenged the importance of inflammation in cervical ripening.^9,10,18,19 Sakamoto et al⁹ demonstrated that there was no correlation between the degree of clinical cervical ripening and IL-8 concentrations in cervical tissue (although IL-8 concentrations in the cervical tissue did increase after labor and delivery).

Word et al¹⁰ reported an animal model in which parturition does not take place despite uterine contractions because of a rigid, inelastic cervix at term. Cervical ripening did not occur despite infiltration with neutrophils and macrophages of this tissue (mice had a transgene insertion on chromosome 6).¹⁰ Timmons and Mahendroo¹⁸ have challenged the traditional paradigm that the influx of inflammatory cells is a major regulatory event in the initiation of cervical ripening using steroid 5 α-reductase type 1 null mice (Srd5a1^−/−). The investigators found that cervical ripening does not require activation of a typical inflammatory response, because macrophages, eosinophils, and myeloperoxidase activity did not increase during cervical ripening. Moreover, depletion of neutrophil numbers before birth has no effect on the timing or success of parturition. The authors proposed that normal cervical ripening does not require a typical inflammatory response and that inflammatory cells are recruited in preparation for the extensive cervical remodeling that occurs after birth when large amounts of hyaluronan and other extracellular matrix components (ie, collagen) must be removed.¹⁸ A recent microarray analysis of cervical tissue of women at term with a ripe and unripe cervix confirmed that inflammation-related genes did not emerge as differentially expressed with cervical ripening.¹⁹ Gene ontology analysis indicated that cervical ripening was associated with enrichment of specific biological processes (cell adhesion, regulation of locomotion) and metabolic functions (phosphate transfer). Furthermore, pathway analysis identified involvement of focal adhesion, extracellular matrix (ECM)-receptor interaction, cell communication, and cell adhesion molecule pathways.¹⁹

Several lines of evidence support the role of sex steroid hormones in cervical ripening, a concept that has important clinical implications. This evidence includes (1) intravenous administration of 17 β-estradiol induces cervical ripening,⁶ (2) estrogen stimulates collagen degradation in vitro,⁶ (3) progesterone and 17 β-estradiol inhibit collagenolysis in uterine cervical fibroblasts by increasing production of tissue inhibitor of metalloproteinases (TIMP) and suppressing production of collagenase and stromelysin;²⁰ and (4) administration of progesterone receptor antagonist induces cervical ripening in the first trimester pregnancy.²¹ However, recent experimental evidence indicates that the role of sex steroid hormones in cervical ripening is far more complex. For example, cervical ripening in the guinea pig begins long before a decrease in serum progesterone concentrations, suggesting that there must be a progesterone-independent mechanism capable of inducing cervical ripening in this species.²² With regard to the role of estrogens, the administration of estrogen to pregnant women does not consistently result in cervical ripening. Moreover, the administration of neither estradiol nor its precursor (androstenedione) induces cervical ripening in the presence of high concentrations of progesterone in guinea pigs.²² A puzzling observation is that the administration of estradiol to guinea pigs treated with onapristone (a progesterone receptor antagonist that induces cervical ripening) attenuates the cervical ripening normally induced by that compound. Therefore, further investigation is required to elucidate the precise role of sex steroids in cervical ripening.

Cervical Dilation

In contrast to cervical ripening, the processes of cervical dilation and postdelivery cervical remodeling are accompanied by invasion of inflammatory cells. Chemokines such as IL-8, CXCL1, CXCL2, CXCL3, CXCL5, and CCL20 have been implicated in the recruitment of neutrophils to the cervix.^{5,7,17,23,24,25,26,27,28,29,30,31,32, and 33}

Cervical modifications during pregnancy and parturition are not limited to physiological and biochemical changes (softening) but also involve anatomical changes (eg, effacement and dilation), which can be assessed with the use of ultrasound.

Sonographic imaging of the cervix is a less invasive, more precise and objective method of assessing the cervical status when compared to digital examination. Effacement (or cervical shortening), changes in the anatomy of the internal os (funneling), endocervical canal dilation, and spontaneous modifications or induced (transfundal pressure) can be determined by ultrasound examination.

Technique and Pitfalls

The cervix can be examined using a transabdominal, endovaginal, and transperineal approach. We prefer the endovaginal technique for optimal assessment of the cervix. The close proximity of the probe to the cervix and the use of a high-frequency transducer improve image quality. Transabdominal sonography requires a full bladder for adequate visualization of the cervix. Overdistension of the bladder compresses and artificially lengthens the cervix.^34,35 Andersen³⁵ demonstrated that cervical length measurements obtained in 186 pregnant women by transabdominal sonography were longer (5.2 mm on average) than those obtained using endovaginal sonography. In contrast to transabdominal ultrasound, endovaginal scanning does not require a distended bladder.

The transperineal technique can be used when an endovaginal transducer is not available, and does not require a full bladder.^{36,37,38,39,40, and 41} The cervical length can be properly visualized and measured adequately by transperineal sonography in about 80% of patients. However, the presence of air in the vagina may render the examination noninformative. In one study there was a high correlation between cervical length obtained by endovaginal and transperineal methods [interclass correlation coefficients = 0.83 (95% confidence interval 0.79–0.86)]. Of interest, a recent study by Meijer-Hoogeveen et al⁴¹ reported that transvaginal sonographic imaging was preferable to patients when compared to transperineal ultrasound. Hertzberg et al³⁹ prospectively evaluated 64 pregnant women at 14 to 38.2 weeks by transvaginal and transperineal ultrasound. The reviewers of the images preferred transvaginal measurements of the cervix when compared to those obtained using the transperineal approach (P < .001), with the most common reasons reported as the presence of bowel gas, which obscured the cervical external os and other landmarks not clearly depicted in the transperineal images. The mean transvaginal cervical length obtained by endovaginal scan at 14 to 20 weeks' gestation was significantly longer than that obtained by transperineal examination (34.1 mm vs 28.6 mm, P < .006).³⁹ Others have proposed that if the cervix is well visualized the transperineal technique offers a sensitivity of 77%, false-positive rate of 17%, and relative risk of 4.5 with a cervical length of 32.5 mm, measured at 24 weeks, for the prediction of preterm birth prior to 36 weeks.⁴⁰

Technique

Before conducting an endovaginal ultrasound, patients are asked to empty their bladder. During the procedure, the patient lies in the supine position with flexed knees and hips. Some have proposed the use of the upright position to obtain the cervical length measurement,⁴² whereas others have reported a good correlation between measurements obtained in the supine and standing positions.⁴³ The probe is covered with either a glove or an appropriate sheath. There is no risk for microbial inoculation as a result of the use of the endovaginal probe.⁴⁴ Gel is placed between the transducer and the cover as well as on the surface sheath. The operator introduces the vaginal probe into the anterior fornix until a midline sagittal view of the cervix and lower uterine segment are seen. The internal os, external os, cervical canal, and endocervical mucosa should be clearly identified (Figure 28-1). The endocervical mucosa is used to define the upper edge of the cervix. Otherwise, cervical length may erroneously include part of the lower uterine segment.

Excessive pressure with the probe may elongate the cervix. To avoid this pitfall, the probe is slowly withdrawn until the image blurs and is subsequently reapplied with an amount of pressure sufficient to restore the image. This can be avoided by confirming equal cervical widths and density of the anterior and posterior lips of the endocervical canal (Figure 28-2).^45,46

The cervical length is measured by freezing the screen 3 separate times. The reliability of measurements is increased if the variation between the measurements is not more than 2 to 3 mm; the optimal length of examination averages 5 to 10 minutes. For clinical purposes, the shortest cervical length is reported provided that the image is adequate. The examination is recorded on videotape and the presence of a funnel or dynamic cervical change is noted. If dynamic change is noted, the shortest cervical length measured during this dynamic change is recorded. A funnel is defined as dilation of the upper portion of the cervical canal (Figure 28-3). This can only be recognized by confirming that the walls of the funnel are formed by endocervical mucosa. Otherwise, the covering wall of the lower uterine segment can be erroneously considered as a funnel.

When the cervical canal is curved, the cervical length can be determined by tracing along the canal or by adding the sum of 2 straight sections. To et al⁴⁷ concluded from a prospective study conducted in 301 women at 23 weeks' gestation that the disparity of measurement, taken as a straight line or along the cervical canal, might not have any clinical significance because a short cervix (less than 16 mm) is always straight.

If the duration of the examination is too short and the patient has dynamic cervical changes during the examination, the cervical length may not represent the true baseline status of the cervix. This may account for some observations in which patients seem to gain cervical length over time. The contours of the anterior and posterior lip of the ectocervix are usually clearly defined. However, in some instances the boundaries of the ectocervix cannot be discerned. Rarely, in selected patients whose cervix cannot be adequately imaged, instillation of saline into the vagina allows precise recognition of the ectocervix.⁴⁸

Standardized measurements of a funnel are difficult because a funnel can be obliterated by pressure from the transducer; it is often transient; and a distended bladder can obscure it. Since the longer the funnel, the shorter the remaining cervical length,^49,50 it has been argued that the endocervical length contains most of the information required for the prediction of preterm delivery. Most favor cervical length measurement because it is far more reproducible than the assessment of funneling. The value of detecting funneling in the prediction of preterm delivery is reviewed later in this chapter. A rare pitfall in the examination of the cervix is that caused by the presence of a large endocervical polyp, which can make identification of the correct plane difficult to image.⁵¹ Although cervical examination may appear simple to the experienced sonographer, certain patients may present significant challenges. For example, a cervix with an unusual orientation may be difficult to find. In a recent study, Yost et al⁵¹ reported that 27% of the scans performed in 60 women presented some anatomical or technical difficulty.

Therefore, several potential pitfalls should be avoided. These include (1) excessive probe pressure (falsely long; Figure 28-4A), (2) failure to observe cervical shortening for enough time (falsely long), (3) failure to recognize a poorly developed lower uterine segment, (4) unequal size and density of the anterior and posterior lips (Figure 28-4B), (5) full bladder (Figure 28-5), (6) endocervical canal not visualized, and (7) lack of amniotic fluid "sludge" recognition (Figure 28-6) (the clinical significance of amniotic fluid "sludge" is reviewed later in this chapter).

We have proposed that parturition has a common terminal pathway characterized by increased myometrial contractility, cervical ripening, and membrane/decidual activation.⁵² These processes are required for both term and preterm parturition. Normal labor at term is characterized by coordinated activation of these 3 mechanisms and thus, patients will present with increased uterine contractility, cervical effacement, and dilation and eventually rupture their membranes. However, premature activation of different components can occur in a coordinated or uncoordinated way. The premature coordinated activation of these mechanisms is observed in patients with classic preterm labor with progressive cervical dilation leading to preterm delivery. Preferential activation of the cervical ripening component in the mid-trimester can lead to premature cervical dilation. Later in pregnancy some patients will complain of vaginal pressure and will be recognized to have cervical dilation and effacement out of proportion to the frequency and intensity of myometrial contractility. Although these patients are traditionally considered to have preterm labor, clearly cervical ripening is the main component of the terminal pathway activated in these particular cases.⁵² This model of preterm parturition is important because it suggests that premature cervical ripening may result in a spectrum of disease ranging from mid-trimester abortion, some forms of preterm labor, and precipitous labor at term. Figure 28-7 describes the concepts outlined above.

Preterm birth is the leading cause of perinatal morbidity and mortality, accounting for 85% of neonatal deaths. One in eight babies was born preterm in 2005, and preterm birth accounts for over 500,000 newborns per year in the United States alone.⁵³ Moreover, the complications of preterm birth can be devastating, as prematurity is the leading identifiable cause of neurologic handicap. The annual cost to the United States as a result of preterm birth is in excess of $26 billion.⁵³

Currently, there are no accurate means of early diagnosis, prevention, or effective treatment of preterm birth. Many interventions have been proposed to reduce the rate of prematurity, without success. Uterine activity monitoring has been used to identify activation of the myometrium, fetal fibronectin to detect decidual-membrane activation, and cervical sonography to identify preterm cervical ripening. However, there is no evidence that the identification of maternal risk factors,^{54,55, and 56} the use of home uterine monitoring,^57,58 or antibiotic administration to patients with a positive cervico/vaginal fetal fibronectin test can reduce the rate of prematurity.⁵⁹ Similarly, pharmacologic inhibition of uterine contractility (tocolysis) has not reduced the rate of preterm birth or neonatal complications.

It is well established that a sonographic short cervix is a powerful predictor of spontaneous preterm birth.^{35,49,50,56,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75, and 76} Cervical sonography has been used most widely to assess the risk for spontaneous preterm birth in 3 groups: (1) asymptomatic patients, (2) patients at high risk for preterm delivery and/or mid-trimester loss, and (3) patients presenting with preterm labor.

Cervical Examination in Asymptomatic Patients

Cervical Length in the Prediction of Preterm Delivery in Asymptomatic Patients

Several studies have measured the cervical length of pregnant women using transabdominal, endovaginal, and transperineal ultrasound. In most cases, the cervical length is stable in the first 30 weeks of pregnancy both in nulliparous and in multiparous women who deliver at term, and a progressive, although not substantial, shortening of the cervix occurs in the third trimester of pregnancy.^60,61,76,77 Median or mean cervical lengths in low-risk populations in the mid-trimester are shown in Table 28-1.

One of the earliest studies to evaluate the relationship between a sonographic short cervix and the risk of spontaneous preterm birth was conducted by Andersen et al⁷⁶ in a cohort of 113 women in which patients were evaluated by transabdominal and transvaginal sonography, along with manual assessment of the cervix on one occasion before 30 weeks' gestation. The patient population, with an overall prematurity rate of 15%, was not separated into low-risk and high-risk categories. The authors reported that a cervical length of less than 39 mm is associated with a 25% risk of preterm delivery, while a long cervix (defined as a cervical length ≥39 mm) decreases the risk of preterm birth (6.7%).⁷⁶ Furthermore, the risk of spontaneous preterm birth was inversely related to the cervical length. These findings have been confirmed by other investigators, both in low-risk^{49,50,56,60,61,63,64,66,67, and 68,70,71,73,75,76,77,78,79,80, and 81} and high-risk asymptomatic patients.^{45,65,69,72,74,82,83, and 84} Table 28-2 describes the details of some of the studies with adequate information to allow calculation of diagnostic indices and predictive values. Our review focuses on the highlights of some of the studies that have significantly contributed to the understanding of the value of cervical sonography in screening for patients at risk for spontaneous preterm birth.

The Maternal-Fetal Medicine Units Network of the National Institute of Child Health and Human Development (NICHD) conducted a prospective cohort study entitled the "Preterm Prediction Study." The value of clinical, demographic, microbiologic, biochemical, and sonographic parameters in the prediction of preterm birth were examined. Iams et al⁴⁹ reported the cardinal observations of cervical sonography. A total of 2915 low-risk asymptomatic patients were examined at 24 weeks' gestation and at 28 weeks by transvaginal sonography to evaluate the cervix and calculate the risk of delivery prior to 35 weeks. The shorter the cervix, the greater the risk of spontaneous preterm birth (Figure 28-8). An exponential increase in the relative risk of delivering before 35 weeks was described (Figure 28-9). The diagnostic indices for different cutoff values of cervical length, funneling, and Bishop score are displayed in Table 28-3. This large study confirmed the results of Andersen et al⁷⁶ indicating that a short cervix increases the risk for preterm delivery while a long cervix decreases such risk, and extended the observations by allowing discernment of the comparative value of cervical length with other predictors of preterm delivery (demographic, biochemical, microbiologic, and clinical).⁵⁶

An important series of studies reported by Heath et al⁶⁶ and conducted at King's College Hospital in London examined the value of cervical sonography in the screening of preterm birth. Cervical length was measured by transvaginal sonography in a low-risk population of 2567 patients with known outcome. (For the distribution of cervical length at 23 weeks of gestation, see Figure 28-10). Patients with a history of preterm birth, of Afro-Caribbean origin, of low maternal age (<20 years) and low body mass index had a shorter cervix than those without such risk factors. However, when logistic regression analysis was used to examine the contribution of all these parameters to the prediction of preterm birth (≤32 weeks), a short cervix was the only predictor of outcome.⁶⁶ These findings suggest that clinical and demographic risk factors associated with preterm birth operate by inducing cervical ripening. In this study,⁶⁶ a cervix of 15 mm or less at 23 weeks of gestation (1.7% of the population) identified 60% of patients who subsequently had a spontaneous preterm birth at less than 32 weeks and 80% of those who had a spontaneous preterm birth at ≤30 weeks.

We conducted a retrospective cohort study⁷⁰ of 6877 women with cervical sonography performed between 14 and 24 weeks. Examinations were conducted transabdominally and, in cases of cervical length less than 30 mm or suboptimal visualization, transvaginally. A cervical length of 15 mm or less had a positive predictive value of 48%, a negative predictive value of 97%, a sensitivity of 8%, and a specificity of 99.7% for spontaneous preterm delivery at 32 weeks or less. A history of preterm delivery and African American ethnic group were also associated with the occurrence of spontaneous preterm birth, although the odds ratios were considerably lower than that of a short cervix. The low sensitivity of a short cervix in this study is similar to that reported by Taipale and Hiilesmaa⁶⁷ in a large study in Finland. The apparent discrepancy among studies could be explained by the different gestational age at which ultrasound examination was conducted. We have evidence that a cervical length obtained later in gestation is more predictive of preterm delivery than those performed earlier.⁷⁰ The study of Heath et al⁶⁶ includes exams conducted at 23 weeks, whereas the exams by Hassan et al⁷⁰ and Taipale and Hiilesmaa⁶⁷ were performed at earlier gestational ages (Figure 28-11^49,66,70,76).

In conclusion, several studies have confirmed and extended the observations of Andersen et al⁷⁶ in asymptomatic patients at low risk for preterm delivery.^{49,50,56,60,61,64,66,67,68,69,70, and 71,73,75,76,79,80, and 81,85,86,87,88, and 89}

A short cervix of 15 mm or less in asymptomatic patients at 23 weeks identifies a population at high risk for spontaneous preterm delivery. However, at least one-third of patients who deliver preterm (<32 weeks) will not have a short cervix in the mid-trimester, and therefore cervical length with ultrasound is not a screening tool but rather a method for risk assessment. The high positive predictive value of a short cervix (nearly 50% for spontaneous preterm birth <32 weeks) has justified trials of intervention (see below under cervical cerclage and progesterone to prevent preterm birth).

Longitudinal Study of Cervical Ultrasound in Asymptomatic Patients

Zorzoli et al⁸⁵ evaluated changes in the cervical dimension of 154 pregnant women at a mean gestational age of 12, 16, 20, 25, and 31 weeks in a population with a prevalence of prematurity of 1.9% (defined as <35 weeks). They reported that cervical length did not change significantly (P = .06) with gestational age, whereas the anteroposterior diameter at mid-portion (cervical width) of the cervix shortened with advancing gestational age. Multiparous women had longer and thicker cervices than primigravidae or women with previous cesarean sections or first trimester abortions. Similarly, Cook and Ellwood⁹⁰ studied 41 patients longitudinally from 18 to 30 weeks' gestation. The cervical length and cervical diameter were followed every 2 weeks. Cervical length and diameter were constant in both nulliparous (n = 21) and primiparous (n = 20) women throughout the studied period. The mean cervical length in primiparous women was longer than that in nulliparous women (44.4 ± 5.1 mm vs 40.6 ± 4.7 mm; P <.001).⁹⁰

In contrast to the earlier studies, two more recent longitudinal studies in nulliparous and parous women reported a decrease in cervical length with advancing gestation.Bergelin et al⁹¹ studied 19 healthy nulliparous women every 2 weeks from 22 weeks until delivery at term. In all but 1 patient, cervical length decreased and cervical width increased with advancing gestation. Three patterns of change in cervical length were observed: (1) a continuous decrease in 53% (10 of 19), (2) an accelerated shortening rate after approximately 30 weeks in 26% (5 of 19), and (3) a sudden shortening after 36 weeks in 16% (3 of 19) of patients. A similar study was conducted in 21 parous women. Patients were examined every 2 weeks from 24 weeks until delivery. Cervical length was unchanged in 3 women, but decreased in 18. As in the study with nulliparous patients, 3 patterns of cervical change were noted: (1) steady, continuous rate of decrease (n = 12; median 1.1 mm/wk, range 0.6 to 2.4); (2) accelerated rate of decrease towards the end of pregnancy (n = 4; median 3.0 mm/wk, range 1.5 to 4.8); (3) sudden drop in cervical length at term (n = 2).

Conflicting results have also been reported regarding whether multiparous women have a longer cervix than nulliparous women. Nulliparous women had a significantly longer cervix when compared to multiparous women in two studies,^49,77 and in another, cervical length was comparable between multiparous and nulliparous women at 17 to 32 weeks,⁹² but at 33 weeks the cervix was longer in parous women.

Cervical Length in Predicting Preterm Delivery in High-Risk Singleton Gestations

Cook and Ellwood⁹³ conducted a prospective cohort study in 120 women considered to be at risk for preterm delivery. The patients at risk included patients with a history of recurrent first trimester abortions, second trimester loss, previous preterm delivery (<34 weeks), previous cervical surgery, and uterine anomaly. Initial cervical assessment was performed between 9 and 29 weeks' gestation. Further assessment varied from weekly to monthly according to cervical findings and obstetric history. The cervical parameters measured were cervical length and diameter, and internal os dilation. Twenty-four patients (20%) delivered before 34 weeks of gestation. Cervical length was the only factor found to be of value in the prediction of preterm delivery. A cervical length of 20 mm or less before 20 weeks was associated with delivery before 34 weeks in 95% of women.

Guzman et al⁶⁵ enrolled 469 high-risk patients between 15 and 24 weeks of gestation. High risk was defined as the presence of a history of spontaneous preterm birth before 37 weeks' gestation, prior mid-trimester loss, more than 2 terminations of pregnancy, cone biopsy, uterine malformation, previous cerclage, and diethylstilbestrol exposure. Transvaginal cervical sonography and transfundal pressure were performed and the shortest cervical length, funnel width, funnel length, and cervical index were recorded serially. A cervical length of 25 mm or less had a sensitivity, specificity, positive predictive value, and negative predictive value of 76%, 68%, 20%, and 96%, respectively, to identify preterm birth at less than 34 weeks of gestation. Cervical length was the best parameter in the prediction of preterm birth in women with prior mid-trimester losses. The authors suggested that using a cervical length of 15 mm or less had a sensitivity, specificity, positive predictive value, and negative predictive value of 81%, 72%, 29%, and 96%, respectively, in predicting spontaneous preterm delivery at less than 34 weeks. Cervical length was better at predicting earlier forms of prematurity (at <28 and <30 weeks) than later forms (<32 and <34 weeks).

Owen et al⁷² performed an observational study including 183 patients who had a history of spontaneous preterm birth before 32 weeks of gestation. The patients were enrolled between 16 and 19 weeks of gestation and followed every 2 weeks until 24 weeks of gestation. Forty-eight (26%) women delivered before 35 weeks of gestation. A short cervix (<25 mm) at the first scan was associated with a relative risk of 3.3 (95% CI = 2.1 to 5.0) for spontaneous preterm birth (<35 weeks). The sensitivity, specificity, and positive predictive values were 19%, 98%, and 75%, respectively. After controlling for cervical length, neither funneling nor dynamic shortening were independent predictors of spontaneous preterm birth. However, using the shortest cervical length on serial evaluations, after any dynamic shortening, the relative risk of a cervical length of less than 25 mm increased to 4.5 (95% CI = 2.7 to 7.6) with a sensitivity, specificity, and positive predictive value of 69%, 80%, and 55%, respectively. Thus, a serial measurement at up to 24 weeks significantly improved the sensitivity but lowered the positive predictive value of the test.

Subsequently, 2 secondary analyses^82,94 were performed of the study by Owen et al.⁷² The first investigation evaluated the effect of gestational age at delivery of a prior preterm birth upon the predictive value of a sonographic short cervix. Patients were divided into 2 groups according to gestational age at delivery of a prior preterm birth: (1) before 23 weeks and (2) 23 to 31 weeks. There was no difference in the cervical length between groups (initial or shortest over study period, median 38 mm vs 37 mm, P = .54, and 30 mm vs 30 mm, P = .97), and a short cervix (<25 mm) had a similar positive predictive value in both groups (80% vs 71%, P > .99) for delivery before 35 weeks.⁹⁴ Owen et al further noted that patients diagnosed with a short cervix prior to 22 weeks were more likely to deliver prior to 26 weeks than later (26 to 34 weeks).⁸²

Durnwald et al⁷⁴ performed a retrospective analysis of 188 women with at least 1 prior spontaneous preterm birth who underwent transvaginal ultrasound at 22 to 24 6/7 weeks. The authors reported that in patients with a short cervix (<25 mm) there was no difference in the rates of preterm birth at less than 32 or less than 35 weeks, irrespective of the number of prior preterm births (rate of delivery <32 weeks: patients with 1 prior preterm birth = 12.5% vs 2 or more prior preterm births = 21.5%, P = .47).

Crane and Hutchens⁹⁵ performed a systematic review of 14 studies that evaluated the use of sonographic cervical length to predict preterm birth in patients at high risk for preterm birth. Patients with singleton pregnancies with a prior history of spontaneous preterm birth, uterine anomalies, or prior cervical excision procedures were included from this review. The primary outcome was delivery at less than 35 weeks' gestation. Patients with a history of spontaneous preterm delivery and a cervical length of less than 25 mm at less than 20 weeks had a positive likelihood ratio (LR) of 11.30 (95% CI = 3.59 to 35.57) and at 20 to 24 weeks an LR of 2.86 (95% CI = 2.12 to 3.87), respectively for delivery at less than 35 weeks.

Recently, Szychowski et al⁹⁶ conducted an analysis of pre-randomization data from a multicenter trial evaluating the efficacy of cervical cerclage in women with a short cervix. Transvaginal sonographic cervical length was measured in 1014 high-risk women at 16 to 22 6/7 weeks gestation. The risk of cervical shortening in women who had a history of the earliest prior preterm birth at less than 24 weeks was significantly higher when compared to those who had a prior preterm birth at 24 to 33 weeks (relative risk [RR] 1.8, P < .0001). In addition, the time to cervical length shortening was significantly shorter in the early prior preterm birth group (hazard ratio = 2.2, P < .0001).

A sonographic short cervix is also predictive of preterm birth in other high-risk populations such as patients with uterine anomalies⁸³ or those with multiple prior induced abortions.⁸⁴

However, the increased risk for preterm birth in patients with a short cervix is not confined to patients with an a priori high risk. In a study⁹⁷ of 109 asymptomatic patients with a short cervix (≤15 mm) diagnosed using transvaginal ultrasound between 14 and 24 weeks' gestation, it was found that the risk of very early preterm delivery (<24 weeks and <28 weeks) is not significantly different between high-risk (n = 42) and low-risk patients (n = 67) and the impact of an a priori risk (history of preterm delivery, late spontaneous abortion, or a cervical surgery) for preterm delivery begins to be significant only for later preterm deliveries (at 32 weeks).

Short Cervix and the Risk of Preterm Premature Rupture of Membranes and Subsequent Preterm Delivery

The Maternal-Fetal Medicine Units Network examined the relationship between a short cervix and preterm birth caused by preterm prelabor rupture of membranes (PROM).⁹⁸ A total of 2929 patients were evaluated in 10 centers between 23 and 24 weeks of gestation. The frequency of preterm birth at less than 37 weeks of gestation was 14.4%. Preterm PROM at less than 35 weeks' gestation and at less than 37 weeks' gestation occurred in 2% and 4.5% of the patients, respectively, and this accounted for 32.6% of all preterm deliveries (<37 weeks). A short cervix, previous preterm birth caused by preterm PROM, and positive fetal fibronectin test were strong predictors for preterm birth caused by preterm PROM at both less than 35 and 37 weeks' gestation. Multivariate analysis indicated that a short cervical length at 23 to 24 weeks' gestation was consistently associated with preterm PROM among both nulliparous and multiparous women [at <35 weeks—nulliparous, odds ratio (OR) 9.9 (95% CI = 3.3 to 25.9) vs multiparous, OR 4.2 (2.0 to 8.9); and at <37 weeks—nulliparous, OR 3.7 (1.8 to 7.7) vs multiparous, OR 2.5 (1.4 to 4.5)].

Odibo et al⁹⁹ demonstrated the relationship between a short cervix and preterm PROM leading to preterm birth in a high-risk population. The authors studied 69 women at high risk for preterm birth (by obstetric history and transvaginal cervical length <25 mm) between 14 and 24 weeks of gestation. The incidence of preterm PROM was 39% (27 of 69). Cervical length of less than 10 mm had a sensitivity, specificity, and positive and negative predictive values of 33%, 90%, 69%, and 68%, respectively, in predicting preterm PROM at less than 35 weeks of gestation.⁹⁹

Later, Kishida et al¹⁰⁰ evaluated 72 singleton pregnancies between 20 and 33 weeks. The authors reported that the presence of an IL-6 concentration of 240 pg/mL or more in cervical secretions and a cervical length of 28 mm or less conferred a probability of 58.1% for the development of preterm PROM.¹⁰⁰

Cervical Length in Twin Pregnancies

Twin gestations occur in 1% of all pregnancies, and they are at increased risk of preterm birth. Several studies have examined the value of endovaginal sonography for the prediction of preterm delivery in twin pregnancies, and reported an increased risk for preterm birth in twin pregnancies with a sonographic short cervix.^{101,102,103,104,105,106,107,108,109,110,111,112,113,114,115, and 116}

The Preterm Prediction Study of the NICHD Network of Maternal-Fetal Medicine Units examined risk factors for preterm delivery in twin gestations. Goldenberg et al¹⁰³ reported that a short cervix, defined as a length of 25 mm or less, was more common in twin than in singleton gestations at 24 and 28 weeks. Moreover, at 24 weeks a short cervix was the only factor predictive of preterm birth. At 28 weeks, a positive fetal fibronectin test was significantly associated with spontaneous preterm birth at less than 32 weeks.

Souka et al¹⁰⁶ studied the sonographic cervical length at 23 weeks' gestation in 215 twin pregnancies. The sensitivity of a short cervix defined as a length of 25 mm or less in the prediction of spontaneous preterm delivery at 28, 30, 32, and 34 weeks was 100%, 80%, 47%, and 35%, respectively (Table 28-4). The rate of spontaneous delivery at or before 32 weeks increased exponentially with decreasing cervical length measured at 23 weeks. An interesting observation was that the risk of preterm delivery for patients with a cervical length of 25 mm or less in twin pregnancies was similar to the risk in singleton pregnancies with a cervical length of 15 mm or less (52%). This has been interpreted as indicating that the cervical length required in twin gestations to confer protection against preterm delivery is greater than that of singleton gestations.

Guzman et al¹⁰⁷ reported a prospective longitudinal study of 131 twin pregnancies between 15 and 28 weeks of gestation. A short cervix (≤20 mm), regardless of gestational age, predicted preterm delivery as well as funnel width, funnel length, percentage of funneling, and cervical index.

Yang et al¹⁰⁸ studied 65 twin pregnancies between 18 and 26 weeks' gestation. Transvaginal or translabial cervical sonography was used to evaluate cervical length and the presence of a funnel. The prevalence of preterm delivery (<35 weeks' gestation) was 23% (15 of 65). A cervical length of 25 mm or less and 30 mm or less was associated with a sensitivity of 27% and 53%, respectively, in predicting preterm delivery. The positive predictive value was 67% and 62%, respectively, for each cutoff (RR = 4.6 [95% CI = 2.0 to 10.3] and RR = 3.6 [1.6 to 7.8], respectively).

A long cervix in twin gestations is reassuring. Imseis et al¹⁰⁵ reported that 97% of twin gestations with a cervical length of 35 mm or more delivered after 34 weeks of gestation.

Skentou et al¹⁰⁹ evaluated 464 twin pregnancies at 23 weeks' gestation. The median cervical length was similar to singletons (36 mm). However, a greater proportion of twin pregnancies had a cervical length of 25 mm or less (12.9%¹⁰⁹) when compared to singletons (8.4%¹¹⁷). The same held true for patients with a cervical length of 15 mm or less (4.5% in twins¹⁰⁹ vs 1.5% in singletons¹¹⁷). Forty percent of patients with a twin gestation who delivered prior to 33 weeks had a cervical length of 20 mm or less.¹⁰⁹

A prospective multicenter study included 251 twin pregnancies at 22 weeks of gestation and 215 at 27 weeks of gestation; of the population that was included at 22 weeks of gestation, for spontaneous delivery before 32 and 35 weeks of gestation, the sensitivity of cervical length ≤30 mm was 40% and 27%, respectively; the specificity was 89% and 90%, respectively. For the population that was included at 27 weeks, a cervical length of 25 mm or less had a sensitivity of 100% and 54%, and a specificity of 84% and 87% for the prediction of spontaneous preterm delivery prior to 32 and 35 weeks, respectively.¹¹¹

A reference range for cervical length in 144 twin pregnancies that delivered after 34 weeks was reported by Fujita et al.¹¹⁸ An arbitrary cervical length measurement obtained at 13 to 32 weeks was used in the analysis. The mean cervical length decreased 0.8 mm per week (95% CI = −1.02 to −0.49) from 47 mm at 13 weeks to 32 mm at 32 weeks.¹¹⁸ Furthermore, the cervical shortening of those twin pregnancies that deliver preterm is more rapid than those who delivered at 36 weeks or later.¹¹²

Sonographic cervical length is also predictive of preterm birth in twin gestations in acute preterm labor and can distinguish those who will deliver within 7 days from those who will not.¹¹³ In a study of 87 twin pregnancies presenting in preterm labor, 80% (4 of 5) of patients with a cervical length of 1 to 5 mm delivered within 7 days, in contrast to 0% (0 of 21) of patients with a cervical length of greater than 25 mm.¹¹³ A cutoff of 25 mm (≤25) has been proposed for the risk assessment of preterm birth in twin pregnancies.^113,115

Robyr et al,¹¹⁴ in an observational study, examined the utility of preoperative sonographic cervical length in predicting preterm birth in 137 cases of twin-to-twin transfusion syndrome (TTTS) undergoing laser coagulation of placental anastomoses before 26 weeks' gestation. The risk of delivery prior to 34 weeks was 74% in patients with a cervical length less than 30 mm. In addition, logistic regression analysis identified cervical length less than 30 mm (OR 3.53 [1.55 to 8.03]) and multiparous (OR 2.27 [1.09 to 4.74]) as independent risk factors for preterm birth at less than 34 weeks.¹¹⁴

Sonographic Evaluation of Cervical Length in Triplet Pregnancies

One study examined the value of cervical ultrasound in 32 triplet gestations. Progressive shortening of the cervix occurred with advancing gestational age. Cervical length in patients who delivered before 33 weeks was significantly shorter at 20, 29, and 31 weeks than that of patients who delivered at or after 33 weeks.¹¹⁹

Guzman et al¹²⁰ conducted a prospective cohort study including 51 triplet gestations evaluated longitudinally between 15 and 28 weeks of gestation. Cervical assessment included cervical length, funnel width and length, percentage of funneling, and cervical index at rest and with transfundal pressure. A cervical length of 25 mm or less between 15 and 20 weeks' gestation had both a specificity and positive predictive value of 100% and a sensitivity of 50% in predicting delivery at less than 28 weeks of gestation. The sensitivity, specificity, positive predictive value, and negative predictive value of a short cervical length measured between 21 and 24 weeks and between 25 and 28 weeks were 86%, 79%, 40%, 97%, and 100%, 57%, 18%, 100%, respectively. The authors suggested that a cervical length of 25 mm or less between 15 and 24 weeks' gestation and 20 mm or less between 25 and 28 weeks' gestation were at least as good as other ultrasonographic cervical parameters for the prediction of spontaneous preterm birth in triplet gestation.

To et al¹²¹ measured the cervical length at 23 weeks of gestation in 38 triplet pregnancies. The rate of spontaneous preterm birth at less than 33 weeks was 16% (6 of 38). The shorter the cervix (at 23 weeks), the higher the rate of preterm delivery. Cervical length of 25 mm or less was present in 16% (6 of 38) of the patients. The sensitivity and positive predictive value for this cutoff were both 50% (3 of 6). The corresponding figures for cervical length of 15 mm or less were 8%, 33%, and 67%, respectively.

Maymon et al¹²² evaluated 45 triplet pregnant women longitudinally from 26 weeks of gestation. The prevalence of preterm delivery was 50% (spontaneous and indicated). Cervical length at 26 weeks' gestation was found to be a risk factor for preterm delivery (<33 weeks). A cervical length of 25 mm or less had a sensitivity, specificity, and positive and negative predictive values of 94%, 45%, 91%, and 70%, respectively. This study also indicated that the later the examination, the higher the sensitivity and positive predictive value. Maslovitz et al¹²³ reported a sensitivity of 75%, specificity of 90%, positive predictive value of 83%, and negative predictive value of 81% for a sonographic cervical length of less than 25 mm at 14 to 20 weeks for the prediction of preterm birth prior to 32 weeks among 36 triplet pregnancies.

Cervical Examination in Patients Presenting with Preterm Labor

A meta-analysis of controlled trials in which patients with preterm labor were treated with either a placebo or β-adrenergic agents indicated that 47% of women treated with placebo deliver at term.¹²⁴ This has been interpreted as indicating that many patients are falsely diagnosed to have preterm labor. Assessment of the likelihood of preterm delivery is of interest because it may influence important clinical decisions such as administration of tocolysis and steroids, as well as transfer to a tertiary care center and/or discharge from the hospital.

Importantly, Iams et al⁶² conducted a study of 60 singleton and twin pregnant women presenting with preterm labor. All patients with a cervix longer than 30 mm delivered at term. There is now compelling evidence that sonographic examination of the cervix in patients presenting with preterm labor can assist in the risk assessment for preterm delivery. The high negative predictive value for preterm birth associated with a long cervix has important clinical implications in symptomatic patients.^{62,113,125,126,127,128,129,130,131,132,133,134,135,136,137,138, and 139} Therefore, a sonographic short cervix is a powerful predictor of preterm birth in women in preterm labor. Gomez et al¹⁴⁰ reported the use of sonographic of cervical length and fetal fibronectin test to predict spontaneous preterm delivery within 48 hours, 7 days, and 14 days of admission as well as delivery at 32 weeks or less and 35 weeks or less. A cervical length of less than 15 mm was the most powerful predictor of preterm birth within 48 hours (OR 9.7, P < .05). Both cervical length and fetal fibronectin were predictive of preterm birth within 7 and 14 days. When fetal fibronectin test results were added to those of a cervical length cutoff of less than 30 mm, there was significant improvement in the prediction of preterm delivery.¹⁴⁰ Tsoi et al¹³⁵ reported no significant contribution to preterm birth prediction by the use of fetal fibronectin, ethnicity, gestational age, body mass index, prior preterm birth, maternal age, parity smoking, or use of tocolytics. Similarly, in a study of 62 pregnant women presenting in preterm labor, only plasma proMMP-9 and cervical length (cutoff 15 mm) were significant predictors of preterm delivery within 7 days.¹³⁴

Table 28-5 summarizes the results of some of the studies published to date that evaluated the use of sonographic cervical length in patients presenting in preterm labor. Although the studies have utilized different cervical length cutoffs, chosen based upon gestational age and prior pregnancy history, there is agreement that the shorter the cervix, the higher the risk for preterm delivery. In addition, in general, the sensitivity and positive predictive value for the prediction of preterm birth are relatively high.

A systematic review of 7 studies evaluating the use of transvaginal ultrasound and cervical dilation in patients in preterm labor was conducted by Leitich et al.¹⁴¹ The authors subdivided patients into 3 groups: (1) patients with preterm labor, and (2) low-risk asymptomatic patients with early (20 to 24 weeks) or (3) late (27 to 32 weeks) sonographic cervical length. The optimal cutoff values for cervical length in patients in preterm labor ranged between 18 and 30 mm, with sensitivity between 68% and 100%, and specificity were between 44% and 79%.¹⁴¹

Several studies have demonstrated that the use of sonographic cervical length in labor and delivery can improve the management of patients in preterm labor by decreasing the number of unnecessary interventions. In addition, by applying specific management schemes that utilize a cervical length cutoff, both hospital stay and health care costs have been reduced.^{130,133,134,137,138,140,142,143,144, and 145} However, the utility of hospitalization in patients with a short cervix¹⁴⁶ or repeat sonographic evaluation of the cervix after tocolysis has been completed has been questioned.¹⁴⁷

A randomized controlled trial¹³⁶ was conducted in 41 women with threatened preterm labor. Patients underwent a sonographic evaluation of the cervix or received the planned treatment (tocolytics and steroids). In the ultrasound group, patients with a cervical length greater than 15 mm had treatment withheld. A higher proportion of patients in the control group received steroid treatment despite remaining undelivered for more than a week when compared to those in the ultrasound group (14% [3 of 21] vs 90% [18 of 20], RR 0.16; 95% CI = 0.05 to 0.39). In addition, there was a significant difference in the duration of hospital stay between the 2 groups (median [range]; 0 hours [0 to 24] in the ultrasound group vs 24 hours [0 to 67], P < .0001).¹³⁶ Larger studies are needed to assess the impact of the use of this management scheme on neonatal outcome.

Patients with preterm labor and a short cervix are at an increased risk for intrauterine infection/inflammation. Gomez et al¹⁴⁸ conducted a study in 401 patients presenting with preterm labor and cervical dilation of less than 3 cm (as assessed by digital examination). All patients underwent amniocentesis. The prevalence of microbial invasion of the amniotic cavity (MIAC) was 7% (28 of 401). Patients with a short cervix (<15 mm) had a higher rate of microbial invasion of the amniotic cavity and preterm delivery at less than 35 weeks compared with those with a long cervix. (MIAC: short cervix, 26.3% [15 of 57] vs long cervix 3.8% [13 of 344], P < .05; and for preterm delivery <35 weeks: short cervix, 66.7% [38 of 57] vs long cervix, 13.5% [44 of 327], P < .01). Moreover, the rate of infection was higher in patients presenting at 30 weeks or less with a cervical length less than 15 mm when compared to those 15 mm or greater (43% [9 of 21] vs 3.9% [3 of 76], P < .05). The authors provided an estimated risk of MIAC in patients in preterm labor according to gestational age and cervical length (Table 28-6).

An abnormal cervical index and the presence of funneling in patients in preterm labor also increases the risk of preterm delivery, as demonstrated by Gomez et al.¹²⁶ Timor-Tritsch et al¹⁴⁹ have studied the clinical significance of funneling for the prediction of preterm delivery in a population with symptoms of preterm labor. In 70 patients admitted to the hospital for threatened preterm labor, wedging of the internal os was associated with preterm delivery with a sensitivity of 100%, a specificity of 74%, a positive predictive value of 59%, and a negative predictive value of 100%.

In summary, cervical sonography is a powerful method to assess the risk of preterm delivery in patients presenting with preterm labor. In patients found to have a long cervix, it may be beneficial to avoid aggressive intervention. Patients in preterm labor who are diagnosed with a short cervix have an increased risk of intrauterine infection/ inflammation as well as a higher rate of preterm delivery, and thus may benefit from targeted interventions (ie, antibiotic and/or steroid administration and transfer to a center with a newborn special care unit). Further investigation into the most efficacious intervention for these patients is needed.

The causes of a sonographic short cervix are largely unknown. A congenital short cervix has been reported in patients exposed to diethylstilbestrol (DES),¹⁵⁰ though now this is an extremely rare occurrence. Previous studies suggest that maternal age, body mass index, ethnicity, and obstetrical history are independent explanatory variables of sonographically determined cervical length.¹⁵¹ One interpretation of these observations is that the combination of genetic and environmental factors may play a role in determining cervical length. For instance, pregnant adolescents, undernourished women (low body mass index), individuals of African Caribbean or African American origin, and those with a previous preterm delivery have a higher risk of spontaneous preterm birth.^151,152 It is possible that these factors confer risk by altering, among other biological determinants of preterm birth, cervical length.

Since the uterine cervix is a connective tissue structure,¹¹ lack of appropriate nutrition during development (in utero or during critical periods of reproductive development, such as adolescence) conceivably may lead to inadequate extracellular matrix deposition in the cervix, and this may predispose to a short cervix. For example, a low body mass index (below 19.8) is an independent risk factor for preterm delivery (OR 3.6, 95% CI = 1.6 to 8.0).¹⁵³ However, this association may operate through a short cervix. If the low body mass index is due to environmental factors, the short cervix could be an acquired phenomenon. The possibility that maternal under-nutrition may predispose to a short cervix needs to be explored.

Several authors have documented a relationship between a previous obstetrical history of preterm birth and cervical length.^63,154 Iams et al⁶³ reported a study of cervical length in patients with (1) a previous history of "cervical incompetence", (2) a previous preterm delivery at less than 26 weeks, (3) a previous preterm delivery at 27 to 32 weeks, (4) a previous preterm delivery at 33 to 35 weeks, and (5) a control group of women with previous term delivery. A strong relationship was established between cervical length in the index pregnancy and previous obstetric history. This relationship appears to be linear, and patients considered to have a typical history of an incompetent cervix did not constitute a unique group.⁶³ Similar results have been reported by Guzman et al¹⁵⁴ who described a strong relationship between previous obstetric history and cervical length in the subsequent pregnancy. Specifically, the authors observed that the frequency of a short cervix (cervical length <20 mm) or progressive shortening of the cervix to a length of less than 20 mm was associated with the gestational age at delivery in the previous pregnancy.¹⁵⁴ Collectively, these studies suggest a relationship between a history of preterm delivery and the cervical length in a subsequent pregnancy. A short cervical length in the subsequent pregnancy may reflect the result of genetic and environmental factors predisposing to preterm delivery, which can express themselves in changes in cervical structure.

Another cause for a short cervix is premature cervical ripening in response to a pathologic stimulus such as infection. Previous studies have indicated that up to 50% of patients presenting with painless cervical dilation before 24 weeks of gestation have a positive amniotic fluid culture for microorganisms.^155,156 In these cases, infection can be a cause of premature cervical dilation, or may be secondary to prolonged exposure of chorioamniotic membranes to the microbial flora present in the lower genital tract.¹⁵⁵

The Syndromic Nature of a Sonographic Short Cervix

Most women, as they progress in the third trimester of pregnancy toward the spontaneous onset of labor, undergo cervical ripening and, as part of this process, their sonographic cervical length is shortened. Therefore, shortening of the cervix is part of the common terminal pathway of parturition, which includes increase in myometrial contractility, cervical ripening, and membrane/decidual activation.^{52,157,158,159, and 160}

Unscheduled cervical ripening in the preterm gestation may occur in combination with activation of other components of the common terminal pathway or may be asynchronous.^{52,157,158,159, and 160} Thus, the phenotype of the preterm parturition syndrome (preterm labor with intact membranes, premature cervical dilation, preterm PROM, or the combination of these presentations) represents the recruitment of the different components of the uterine pathway of parturition.¹⁵⁹

A short cervix often occurs when there is cervical ripening; however, not all short cervices are ripened, hence the observation that some women with a short cervix deliver at term. We proposed that a short cervix is syndromic in nature and can be caused by multiple etiologies,^{1,159,160, and 161} such as (1) the loss of connective tissue after a cervical operation such as conization^{162,163,164,165, and 166} or LEEP procedure¹⁶³, (2) a congenital disorder such as cervical hypoplasia after DES exposure^{164,167,168, and 169}, (3) intrauterine infection^148,155,170, (4) a suspension or decline of progesterone action¹⁷¹, and (5) a cervical disorder that manifests itself with the clinical presentation of premature cervical dilation. Each of these different causes of the syndrome could be affected by genetic or environmental factors. Moreover, more than one mechanism of disease may be operative in a specific patient. The possibility of novel and yet undiscovered mechanisms of disease playing a role must also be considered.

A Sonographic Short Cervix as the Only Clinical Manifestation of Intra-amniotic Infection and/or Inflammation

Gray et al¹⁷² reported that all women with a positive amniotic fluid culture for Ureaplasma urealyticum at the time of genetic amniocentesis delivered a preterm neonate with histologic evidence of chorioamnionitis. This study, however, did not contain information about cervical length. We have established that 9% (5 of 57) of asymptomatic women in the mid-trimester with a cervical length less than 25 mm without cervical dilation had microbiologically proven intra-amniotic infection.¹⁷⁰ The isolated microorganisms included Ureaplasma urealyticum and Fusobacterium. Patients with amniotic fluid cultures positive for Ureaplasma received intravenous antibiotics (azithromycin) for 7 days and underwent a test-of-cure amniocentesis. Three of the 4 patients had a negative amniocentesis, thus demonstrating eradication of the Ureaplasma, and subsequently delivered at term. Recently, an association between a short cervix in the mid-trimester and elevated concentrations of inflammatory cytokines in amniotic fluid has been proposed. Kiefer et al¹⁷³ reported an association between a cervical length ≤5 mm in the mid-trimester and increased amniotic fluid concentrations of inflammatory cytokines such as monocyte chemotactic protein-1 (MCP-1) and IL-6. Keeler et al¹⁷⁴ in a subsequent study of 44 patients in the mid-trimester with a cervical length ≤25 mm, in which those with intra-amniotic infection and/or preterm labor were not excluded, reported that amniotic fluid MCP-1 was predictive for spontaneous preterm delivery. Vaisbuch et al.¹⁷⁵ demonstrated intra-amniotic inflammation (defined as an amniotic fluid matrix metalloproteinase-8 concentration >23 ng/mL) in 22% of asymptomatic patients (10 of 45) with a cervical length ≤15 mm in the mid-trimester of pregnancy; an occurrence that was associated with adverse pregnancy outcomes. Indeed, women with intra-amniotic inflammation had a shorter median diagnosis-to-delivery interval than those without this condition. Furthermore, 40% of patients with intra-amniotic inflammation delivered within one week of the amniocentesis.

Cervical Funneling

Funneling is the protrusion of no less than 3 mm of amniotic membranes into the internal os, and the size of the funnel frequently changes during the course of the scan. Funneling can be considered as effacement in progress.⁴⁷ Two types of funneling have been described: "V" and "U" shaped. In the V-shaped pattern, the membranes protrude into the cervical canal to form a triangular-shaped funnel. In the U-shaped pattern, the membranes protruding into the endocervical canal form a curvilinear image (see Figure 28-3; Figure 28-12). Zilianti et al³⁷ described the normal progression of the morphology of the upper cervix during the course of labor at term with transperineal sonography and coined the acronym of "TYVU" to describe the morphologic changes. Figure 28-12 displays such an evolution. Similar findings can be demonstrated with endovaginal sonography. However, it should be noted that the work was conducted in term labor, and that it remains to be determined whether preterm labor has a similar progression.

The endocervical canal, the funnel length, and the funnel width are used to calculate the cervical index. The cervical index is equal to (funnel length +1)/endocervical length.¹²⁶ This parameter was derived to take into account both remaining endocervical length and the length of the funnel, as they are both part of the original endocervical canal before effacement began. Several authors have demonstrated that the cervical index and funneling are strong predictors of the risk of preterm delivery in patients with and without preterm labor (Tables 28-7, and 28-8). Some authors have described the dimensions of the funnel as a percentage of the endocervical length. This is fundamentally a similar concept as the cervical index. For example, a funnel representing 40% to 50% or more of the total cervical length has been associated with an increased risk of preterm birth when noted before 30 weeks in a population with a high prevalence of preterm delivery (42%).⁵⁰

In the NICHD Prematurity Prediction Study, Iams et al¹⁷⁶ reported that the value of funneling (defined as 3 mm in width) as a predictor of delivery was similar to the value of cervical length, but the data showed substantial variations among centers. The relative risk of funneling for preterm delivery before 35 weeks was 5.0 at 24 weeks and 4.78 at 28 weeks.⁴⁹ It is of interest that in the study of Taipale and Hiilesmaa,⁶⁷ funneling, defined as dilation of the internal os of 5 mm or greater, was a stronger predictor of preterm delivery before 35 weeks than endocervical cervical length. This study enrolled 3694 patients between 18 and 22 weeks of gestation. A dilated internal os (funneling) had a relative risk of 28 for preterm delivery before 35 weeks while a short cervix (defined as <30 mm) had a relative risk of only 8. Logistic regression analysis demonstrated that the adjusted odds ratio for delivery before 35 weeks for funneling was 20 and for a short cervix was 6.5. We believe that the most likely explanation for the better performance of funneling compared to cervical length in this study is that the definition of a short cervix was less than 30 mm. Most studies indicate that such cervical length has a low positive predictive value for preterm delivery. Results may be different if a cervical length of 15 mm or 20 mm was used for analysis. In patients with a short cervix Owen et al,⁷² To et al,⁷³ and Guzman et al¹⁷⁷ have concluded that funneling does not provide any additional diagnostic information over that provided by cervical length in the prediction of preterm birth.

Amniotic Fluid "Sludge"

Particulate matter in the amniotic fluid is present in about 4% of pregnancies during transvaginal ultrasound in the first and early second trimester.¹⁷⁸ The prevalence of this sonographic finding increases with gestational age, reaching 88% by 35 weeks.¹⁷⁹ Particulate matter in the first 2 trimesters of pregnancy has been associated with intra-amniotic bleeding^180,181 and the acrania-anencephaly sequence¹⁸² and has been observed in women with high concentrations of maternal serum α-fetoprotein.¹⁸³ In contrast, in the last trimester of pregnancy, particulate matter and "echogenic amniotic fluid" have been attributed to the presence of vernix caseosa and/or meconium,^{184,185,186, and 187} and with a lecithin-sphingomyelin ratio indicative of lung maturity.^188,189 Congenital anomalies associated with particulate matter in the amniotic fluid include harlequin ichtiosis¹⁹⁰ and epidermolysis bullosa letalis.¹⁹¹

Amniotic fluid "sludge" is defined as particulate matter seen in proximity of the internal cervical os during a transvaginal sonographic examination of the cervix, and occurs in 1% of uncomplicated pregnancies¹⁹² (see Figure 28-6). The first description of amniotic fluid "sludge" was in patients in preterm labor. Espinoza et al¹⁹² conducted a retrospective study of 84 patients in preterm labor with intact membranes. The prevalence of amniotic fluid "sludge" was 22.6% (19 of 84) in patients with preterm labor. Patients with amniotic fluid "sludge" had a higher frequency of positive amniotic fluid cultures (33.3% [6 of 18] vs 2.5% [1/40], P = .003) and histological chorioamnionitis (77.8% [14 of 18] vs 19% [11 of 58], P < .001), and a higher rate of spontaneous preterm delivery within 48 hours (42.9% [6 of 14] vs 4.4% [2 of 45], P = .001), within 7 days (71.4% [10 of 14] vs 115.6% [7 of 45], P < .001), less than 32 weeks (75% [9 of 12] vs 25.8% [8 of 31], P = .005), and less than 35 weeks (92.9% [13 of 14] vs 37.8% [17 of 45], P < .001) than those without amniotic fluid "sludge". Stepwise logistic regression analysis indicated that the presence of "sludge" was an independent factor associated with the likelihood of spontaneous delivery within 48 hours and 7 days, but not less than 32 weeks or less than 35 weeks. The diagnostic indices of the amniotic fluid "sludge" are displayed in Table 28-9. Survival analysis demonstrated that patients with amniotic fluid "sludge" had a shorter examination-to-delivery interval compared to those without "sludge" (sludge median: 1 day [interquartile range: 1 to 5 days] vs no sludge, median: 33 days [interquartile range: 18 to 58 days]; P < .001) (Figure 28-13). These results indicate that amniotic fluid "sludge" during transvaginal examination of the cervix is a risk factor for MIAC, histological chorioamnionitis, and impending preterm delivery.¹⁹²

The presence of amniotic fluid "sludge" on transvaginal examination in asymptomatic patients is also associated with delivery within 14 days of ultrasound and preterm delivery at less than 32 and less than 34 weeks.^193,194 Bujold et al¹⁹³ conducted a retrospective study of 89 women at 18 to 32 weeks who were at high risk for preterm delivery. Logistic regression analyses revealed that a cervical length less than 25 mm and light or dense amniotic fluid "sludge" were independent predictors of delivery within 14 days and less than 34 weeks.¹⁹³

In a retrospective case control study Kusanovic et al¹⁹⁴ included 281 patients with (n = 66) or without (n = 215) amniotic fluid "sludge" at 13 to 29 weeks' gestation. Patients were at high risk for preterm birth based upon the presence of one or more of the following: a history of spontaneous preterm delivery, previous mid-trimester loss, short cervix (<25 mm), Müllerian duct anomalies, and a history of cone biopsy. The prevalence of amniotic fluid "sludge" was 23.5% in this population. Amniotic fluid "sludge" was identified as an independent risk factor for spontaneous preterm delivery at less than 28 weeks', less than 32 weeks', and less than 35 weeks' gestation; preterm PROM; MIAC; and histological chorioamnionitis. Furthermore, asymptomatic patients with amniotic fluid "sludge" had shorter ultrasound-to-delivery and ultrasound-to-preterm PROM intervals than those without sludge.

Recently, amniotic fluid "sludge" has been demonstrated to represent a biofilm.^195,196 A sample of sludge was retrieved by transvaginal amniotomy under ultrasound guidance from a patient at 28 weeks with spontaneous labor and clinical chorioamnionitis. Grossly, the sludge had a "pus-like" appearance and the Gram stain showed gram-positive bacteria. The amniotic fluid culture was positive for Streptococcus mutans, Mycoplasma hominis, and Aspergillus flavus. Of interest, the results of the amniocentesis performed at the time of admission for preterm labor were negative for intra-amniotic infection.^191,192 The aspirated sludge was then analyzed further. Scanning electron microscopy showed flocs of amniotic fluid "sludge" that consisted of bacterial cells and the exopolymeric matrix material that are typical of a biofilm (Figure 28-14).¹⁹⁶ The evidence that sludge represents a biofilm in this case includes the following: (1) bacteria detected with fluorescence in situ hybridization (FISH), with the use of a probe against the conserved sequence of prokaryotes; (2) bacterial aggregates were separated by material that resembled a matrix; and (3) lectin-based identification of exopolymeric matrix that stained with wheat germ agglutinin.¹⁹⁶

The clinical significance of this relates to the challenges encountered with diagnosis and treatment of this condition. First, the diagnosis of microbial invasion in the presence of biofilms is extremely challenging, and current cultivation techniques are inadequate to detect such infections. The consequences are that the frequency of infection of the amniotic cavity may be underestimated, and that molecular microbiologic techniques will be required to improve diagnosis. Second, the optimal treatment of biofilm-related infections represents a challenge in clinical medicine. Antimicrobial agents appear to be inactivated or fail to reach bacteria within a biofilm. Interestingly, bacteria within biofilms have increased resistance to antimicrobial compounds even though the same bacteria can be sensitive to the same agent if grown under standard conditions.^{197,198,199, and 200} Thus, the difficulties in treating intra-amniotic infection may be due to the refractoriness of biofilms to conventional antibiotic treatment. Third, biofilms in the amniotic fluid may represent a unique form of these structures, which can be dislodged by fetal movement, resulting in seeding of planktonic bacteria and the eliciting of an inflammatory response (see Figure 28-14).

Sonographic cervical length has been utilized to identify the patients who may benefit from cerclage placement or progesterone treatment. Yet, despite numerous trials conducted to determine a treatment for a sonographic short cervix, no standard, effective intervention is available.

Cervical Cerclage in Patients with a Sonographic Short Cervix

Randomized Clinical Trials of Cervical Cerclage before the Cervical Ultrasound Era

The clinical value of a prophylactic cervical cerclage has been a subject of debate in the obstetric literature for many years. Three randomized clinical trials have been conducted so far in singleton pregnancies. Table 28-10 summarizes the details of the clinical trials. All studies were conducted before sonographic evaluation of the cervix became widely available.

Rush et al²⁰¹ reported a study conducted in 194 women at high risk for late spontaneous abortion or preterm delivery. Patients were randomly allocated to either cerclage (n = 96) or expectant management (n = 98) between 15 and 21 weeks. The overall rate of preterm delivery was 33%, but there was no difference in the rate of preterm delivery between patients treated with a McDonald cerclage or expectant management (preterm delivery <37 weeks; expectant management 32%, vs cerclage 34%). The duration of pregnancy was slightly longer in patients managed expectantly than in the cerclage group, but this difference was attributed to days gained after the 37th week. Rupture of membranes was more frequent in patients who had a cerclage than in the control group (18.7% [18 of 96] vs 12.3% [12 of 98]), but this difference did not reach statistical significance. There was no difference in neonatal outcome. However, fever during the puerperium was more common in patients managed with cerclage (cerclage, 9.6% [10 of 96] vs expectant, 3% [3 of 98]; P = .07).

Lazar et al²⁰² reported a trial of 506 women considered to be at moderate risk for preterm delivery based upon a scoring system which considered obstetrical history, state of the cervix (ie, open internal os, a short cervix at digital examination, etc). Two hundred sixty-eight patients were allocated to McDonald cerclage placement with number 3 nylon and 238 to expectant management. The rate of preterm delivery was 6.7% (18 of 268) in the cerclage group and 5.5% (13 of 238) in the control group (overall rate was 6.1%). Patients in the cerclage group had more admissions to the hospital (excluding the hospitalization for cerclage or birth) and received oral tocolysis more frequently than those in the control group (P < .01 for each). However, there were no differences in perinatal mortality between the 2 groups.

The largest trial conducted to date was organized by the Medical Research Council of the United Kingdom and the Royal College of Obstetricians and Gynecologists.²⁰³ The criterion for enrollment was uncertainty on the part of the obstetrician as to whether to recommend a cerclage. The main reason for enrollment was a past history of 1 or more second trimester spontaneous abortions or preterm deliveries (71% of patients) and a history of cervical operation. The type of operation and surgical material was left to the discretion of the caregiver, and the study was conducted in multiple countries. Of patients allocated to cerclage, only 92% had the suture inserted, while 8% of those in the control group had a cerclage. The mean gestational age at cerclage was 15.9 weeks (interquartile range: 14.3 to 18.4). The primary outcome was delivery at less than 33 completed weeks. The rate of delivery before 33 weeks (range 20 to 33 weeks) was significantly lower in the cerclage group than in the control group (cerclage 13% vs control group 17%; OR 0.72; 95% CI = 0.53 to 0.97; P = .03). Fever attributed to uterine infection was more common in patients allocated to the cerclage group (6% vs 3%; OR 2.12; 95% CI = 1.08 to 4.16; P = .03). The authors estimated that 25 cerclages would need to be inserted to prevent 1 preterm delivery at less than 33 weeks. The overall rate of preterm delivery was 28% (<37 weeks). A subgroup analysis suggested that patients with a history of 3 spontaneous mid-trimester abortions or preterm deliveries seemed to benefit the most from the procedure. The authors called for additional research in methods to identify the patients who may benefit from this operation.

One small study has examined the value of prophylactic cervical cerclage in twin gestation. There was no evidence of benefit in this condition.²⁰⁴

In conclusion, results of randomized clinical trials suggest that cerclage either had a modest effect on reducing the rate of preterm delivery or no effect whatsoever. We believe that the fundamental problem is that preterm labor is a syndrome with multiple etiologies and the operation can only benefit those patients who have a primary or important cervical process responsible for preterm delivery. It is clear that obstetrical history is insufficient to identify such a patient. It is possible that future studies in which patient selection for cerclage is based upon results of cervical ultrasound or other noninvasive diagnostic tests may uncover a role for cerclage in modern obstetrics.

Cervical Cerclage after Cervical Ultrasound Assessment

Observational Studies

The prediction of spontaneous preterm birth with cervical sonography has clinical relevance, given the hope that an intervention would reduce the rate of prematurity in patients with a short cervix. The nature of the intervention could encompass medical treatment with agents preventing cervical ripening, mechanical devices such as pessaries, or a cervical cerclage.

Heath et al²⁰⁵ reported that patients with a short cervix who underwent placement of a cerclage had a 10-fold reduction in the rate of preterm birth. Clinicians responsible for the care of the patients with a short cervix (defined as 15 mm or less) were informed of the results and allowed to use their preference for further management. Of the 43 patients with a short cervix, 22 were treated with the placement of a Shirodkar cerclage and 21 were managed expectantly. Although there were no differences in the clinical characteristics of the 2 groups, the rate of preterm delivery (<32 weeks) was 52% (11 of 21) in the group managed expectantly and only 5% (1 of 22) in the group managed with a cervical cerclage.

Guzman et al²⁰⁶ also reported that cerclage improved pregnancy outcome in patients with a history of a spontaneous abortion or preterm delivery who shortened their cervix (defined as 20 mm or less) between 20 and 24 weeks of gestation. The decision to treat or not to treat was also left at the discretion of the treating physician. Seventy-nine percent (22 of 28) of women treated with cerclage delivered at term (≥37 weeks) while only 35% (6 of 17) of those in the bed rest group delivered at term (P = .04). Similarly, Hibbard et al²⁰⁷ reported that patients with a short cervix undergoing cervical cerclage delivered later in gestation than patients managed expectantly. Eighty-five patients with a cervical length of less than 30 mm were included. Some women had a history of a previous preterm birth. A cervical cerclage was placed in 43 patients at the discretion of the treating physician. Although the mean cervical length measurements were similar in both groups (cerclage 22 mm vs noncerclage 20 mm; P = .32), the mean gestational age at delivery and birth weight in the cerclage group were higher than in the group managed without the placement of a cerclage (34 weeks vs 32 weeks, P = .04; and 2530 g vs 2084 g, P < .04).

In contrast to the observational studies suggesting that cerclage may be beneficial, 2 other studies did not find cerclage to be beneficial. Berghella et al²⁰⁸ studied 168 women at risk for preterm delivery who underwent serial sonography between 14 and 23 weeks. Patients with a cervical length of less than 25 mm or funneling greater than 25% of the cervical length were offered either a McDonald cerclage or expectant management. Sixty-three patients had an abnormal cervix; 39 underwent cerclage and 24 were managed expectantly. Patients in the 2 groups were similar in clinical characteristics, cervical length, and funneling, but those who were treated with a cerclage had a lower gestational age when the cervical changes were identified (18 weeks vs 21 weeks) than those in the control group. Placement of cervical cerclage did not reduce the rate of preterm delivery (cerclage 27% vs expectant management 23%) or the duration of pregnancy after the identification of abnormal cervical findings. Similarly, Hassan et al²⁰⁹ found cerclage to have no demonstrable beneficial effect in an observational study of 70 patients who had cervical length of less than 15 mm between 14 and 24 weeks' gestation. Fifteen (21%) patients had a previous history of preterm delivery. McDonald cerclage was placed in 25 patients (36%). The rate of preterm delivery (<34 weeks) was not different between groups (cerclage 68% [17 of 25] vs no cerclage 53% [24 of 45]; P = .2). Patients with cerclage placement had a higher rate of preterm PROM than those without cerclage (65% [15 of 23] vs 36% [16 of 44]; P < .05).

Newman²¹⁰ conducted a prospective cohort study in 128 twin pregnancies at 18 to 26 weeks with a sonographic cervical length of 25 mm or less. Only 16% (21 of 128) underwent cerclage placement. Although decreasing cervical length was associated with delivery at less than 34 weeks, preterm PROM, lower combined birth weight, and very low birth weight, cerclage placement did not improve these outcomes.²¹⁰

Randomized Clinical Trials

At present, there are 5 randomized controlled trials^{211,212,213,214,215,216, and 217} to assess the clinical value of cerclage by using the information from cervical ultrasound examination. The results of these trials are summarized in Table 28-11.

In the Cervical Incompetence Prevention Randomized Cerclage Trial (CIPRACT) study, Althuisius et al^211,212 randomized 73 pregnant women at less than 15 weeks of gestation with risk factors for "cervical incompetence" to have a prophylactic cerclage (n = 23) or be observed (n = 44). Four patients had a spontaneous abortion during the first trimester and 2 were lost for follow-up. The risk factors for "cervical incompetence" included the history of preterm delivery before 34 weeks' gestation, previous preterm PROM before 32 weeks' gestation, history of cold knife conization, DES exposure, and a Müllerian duct abnormality. Prophylactic cerclages (ie, McDonald) were generally placed between 10 and 12 weeks or later if enrolled at a later gestational age using a braided polyester thread. In both groups, the cervical length was evaluated every 2 weeks after randomization. The rate of preterm delivery (<34 weeks) was similar in both groups (prophylactic cerclage 13% [3 of 23] vs observation 14% [6 of 44]; P > .05) and neonatal survival (prophylactic cerclage 91% [21 of 23] vs observation 93% [41 of 44]; P > .05). Patients allocated to the observation group were followed with serial sonography. If the cervical length shortened (<25 mm) before the 27th week of gestation, they were randomized to either therapeutic cerclage with bed rest (n = 20) or bed rest (n = 16). Placement of a therapeutic cerclage reduced the rate of preterm delivery at less than 34 weeks and the composite neonatal morbidity (neonatal intensive care unit [NICU] admission or neonatal death) (0% vs 44% [7 of 16], P = .002; and 5% [1 of 19] vs 50% [8 of 16], P = .005).

In contrast, Rust et al^213,214 randomized 113 patients with a short cervix (<25 mm) or funneling (≥25%) between 16 and 24 weeks into the therapeutic cerclage group (n = 55) and the no cerclage group (n = 58). The population included patients with and without risk factors for preterm birth. All patients underwent amniocentesis to exclude intra-amniotic infection and received 48 hours of therapy with indomethacin and antibiotics. There were no significant differences between the 2 groups with respect to the rate of preterm delivery at less than 34 weeks' gestation (35% vs 36.2%), readmission for preterm labor (52% vs 53%), placental abruption (11% vs 14%), chorioamnionitis (20% vs 10%), and the perinatal death rate (13% vs 12%).

The CIPRACT study^211,212 enrolled only patients at risk for preterm delivery, while in the trial of Rust et al,^213,214 13% of patients were at low risk. The positive predictive value for short cervix to predict preterm delivery tends to be higher in patients with a previous history of preterm birth. The fact that the rate of preterm delivery in the control group of the CIPRACT study is higher than that of the study of Rust et al (43.8% vs 36.2%) may explain, at least in part, the different results between these 2 studies.

Berghella et al²¹⁵ conducted a randomized clinical trial of the use of McDonald cerclage in women with 1 or more risk factors for preterm birth (1 or more prior deliveries at less than 35 weeks, 2 or more curettages, history of DES exposure, cone biopsy, Müllerian anomaly, or twin gestation). Sixty-one patients with a cervical length less than 25 mm or funneling greater than 25% were randomized to cerclage or bed rest. The authors reported that 47 pregnancies (77%) were high-risk singleton gestations.

There was no significant difference in the rate of preterm birth prior to 35 weeks in women who underwent cerclage (14 of 31; 45%) compared to those in the bed rest group (14 of 30; 47%) (RR 0.94; 95% CI = 0.34 to 2.58). Similarly, patients with a singleton gestation and a prior preterm birth at less than 35 weeks of gestation and a cervical length less than 25 mm (n = 31 women) also had no benefit from cerclage placement (40% vs 56%; RR 0.52; 95% CI = 0.12 to 2.17).²¹⁵

The largest randomized clinical trial in patients with a cervical length ≤15 mm reported so far was conducted by the Fetal Medicine Foundation of the United Kingdom.²¹⁶ Cervical length was determined in low-risk patients at median gestational age of 23 weeks, and those with a cervix ≤15 mm were randomized to either expectant management (n = 126) or cerclage group (n = 127). The rate of preterm delivery at less than 33 weeks' gestation was not significantly different (expectant management group 26% [33 of 126] vs cerclage group 22% [28 of 127]). The conclusion of this study is that cerclage placement in patients with a short cervix but without risk factors for preterm delivery does not reduce the rate of spontaneous preterm birth or the rate of perinatal death.²¹⁶ Recently, Owen et al. conducted a randomized clinical trial including 302 women (148 assigned to cerclage) with a cervical length <25 mm and prior spontaneous preterm birth at <34 weeks. The authors reported that cerclage did not show a reduction in the rate of preterm birth at <35 weeks, the study's primary outcome. However, a secondary analysis in 64 women with a cervical length ≤15 mm suggested a beneficial effect to cerclage in this group.²¹⁷

The hypothesis that cervical competence or sufficiency represents a spectrum was studied by Parikh and Mehta,²¹⁸ who used digital examination of the cervix to assess sufficiency. The authors, however, concluded that degrees of cervical competence did not exist. Iams et al,⁶³ using sonographic examination of the cervix, suggested that cervical sufficiency/insufficiency is a continuum. The authors reported a strong relationship between cervical length in pregnancy and previous obstetrical history. This relationship was nearly linear, and patients with a typical history of an incompetent cervix did not constitute a separate group from those who delivered preterm.⁶³ Similar results have been reported by Guzman et al.¹⁵⁴ Collectively, these studies suggest that there is a relationship between a history of preterm delivery and the cervical length in a subsequent pregnancy. Inasmuch as patients with a short cervix are at increased risk for a mid-trimester pregnancy loss (clinically referred to as "cervical insufficiency" by some physicians) or spontaneous preterm delivery with intact or rupture of membranes, ^{35,49,60,61,62,63,64,65,66, and 67,69,70,71,72,73,74,75, and 76,79,80,81, and 82,154} a short cervix could be considered as the expression of a spectrum of cervical diseases or functions. However, it is noteworthy that some women with a short cervix have an adverse pregnancy outcome, while others have an uncomplicated term delivery.^{35,49,60,61,63,65,73,76,151,154,220,221, and 222} Indeed, approximately 50% of women with a cervix of 15 mm or less deliver after 32 weeks.^66,70 This indicates that cervical length may be only one of the factors determining the degree of cervical strength, and that a short cervix should not be equated with "cervical insufficiency."

The clinical value of cervical cerclage has been the subject of many observational and randomized clinical trials,^{201,202,203,204,205,206,207,208,209,210,211,212,213,214,215,216, and 217,223,225,226,227,228,229,230,231,232,233,234,235,236,237,238,239, and 240} and the studies have been the topic of several systematic reviews. ^{241,242, and 243} The evidence suggests the following conclusions:

1. Cervical cerclage in women with a sonographic short cervix (15 mm or less) and a low risk for preterm delivery (by history) does not reduce the rate of spontaneous preterm birth.²¹⁶

2. The effectiveness of cervical cerclage in women with a sonographic short cervix and at high risk (by history) for the prevention of preterm delivery remains controversial.^{212,213,214, and 215}

3. The role of prophylactic cerclage in high-risk patients without a sonographic short cervix for the prevention of preterm delivery/mid-trimester abortion (by history) is unclear.^{197,198, and 199} While the largest trial conducted prior to the introduction of ultrasound evaluation of the cervix suggested a modest beneficial effect, ²⁰³ other trials'^201,202 and systematic reviews²⁴³ prior to the use of ultrasound have indicated that the evidence of effectiveness is either weak or nonexistent.

4. In patients at risk for preterm delivery, serial sonographic examination of the cervix followed by cerclage in those who shortened the cervix is a reasonable alternative to prophylactic placement of a cerclage based on uncontrolled studies.^{20,212,237,244}

Progesterone Use in Patients with a Sonographic Short Cervix

Facchinetti et al²⁴⁵ reported the results of a randomized prospective clinical trial in which women with preterm labor and intact membranes (25 to 33 6/7 weeks) were allocated to either observation or intramuscular administration of 341 mg of 17 α-hydroxyprogesterone caproate twice a week until 36 weeks of gestation or delivery, and sonographic cervical length was measured at discharge as well as 7 (7 to 9) to 21 (18 to 21) days later. Patients allocated to receive 17 α-hydroxyprogesterone caproate had a longer sonographic cervical length than those in the observation group. These findings, coupled with experimental data, suggest that progesterone may have major effects in the uterine cervix.

Two clinical lines of evidence support that cervical status may identify the patient who could benefit from progestogen administration.^246,247 First, a randomized clinical trial reported by Fonseca et al²⁴⁶ indicates that vaginal progesterone reduces the rate of preterm birth by 44% (from 34.4% in the placebo group to 19.2% in the progesterone treatment group) in women with a sonographic short cervix. Next, a secondary analysis²⁴⁷ suggested that patients with a short cervix may benefit from vaginal progesterone administration.

Fonseca et al,²⁴⁶ working with the Fetal Medicine Foundation Second Trimester Screening Group, conducted a randomized, double blind, placebo-controlled trial in which women with a short cervix (≤15 mm by transvaginal ultrasound), between 20 and 25 weeks of gestation, were allocated to daily vaginal administration of 200 mg of micronized progesterone or placebo (safflower oil) from 24 to 34 weeks. The frequency of spontaneous preterm delivery at less than 34 weeks (primary end point for the trial) was significantly lower in the progesterone group than that in the placebo group (19.2% [24 of 125] vs 34.4% [43 of 125]; P = .007). A secondary analysis of this trial indicated that among women without a history of delivery before 34 weeks, the incidence of preterm birth was significantly lower in women receiving progesterone than in those allocated to placebo (17.9% [20 of 112] vs 31.2% [34 of 109]; RR 0.57; 95% CI = 0.35 to 0.93; P = .03). The trial was not designed to test whether progesterone administration could reduce neonatal morbidity and such a reduction was not observed.²⁴⁶

DeFranco et al²⁴⁷ reported a retrospective analysis of the effect of vaginal progesterone on pregnancy outcome (preterm birth and infant outcome) as a function of cervical length. The hypothesis for this study was that the effect of prophylactic vaginal progesterone may vary according to cervical length. To test this concept, all patients eligible to participate in the trial of O'Brien et al²⁴⁸ had a sonographic measurement of the cervix at the time of enrollment. Patients without a previous preterm birth and a cervical length of 25 mm or less were subjected to a separate randomization procedure (to vaginal progesterone or placebo) and excluded from the main trial. The rationale for this was that the main trial tested the effect of vaginal progesterone on patients with a history of preterm delivery regardless of cervical length. However, after the completion of the main trial, only 9 patients had been enrolled in the "short cervix only" arm of the trial. Analysis of the effect of progesterone on 9 patients with a short cervix was not meaningful. Therefore, the investigators modified their initial plan so that the analysis included women enrolled in the main trial (and therefore with a history of previous preterm birth). The authors divided the patient population into quartiles according to cervical length at enrollment and tested whether outcomes were different as a function of cervical length (the lowest quartile was a cervical length of ≤32 mm). Although there was a delay in delivery, progesterone administration did not result in a significant difference in outcome in patients in the lowest quartile.

The authors²⁴⁷ then explored the effect of vaginal progesterone as a function of cervical length using 2 new cutoffs: 30 mm or less and less than 28 mm. For the cutoff of 30 mm, there was a trend for a longer randomization-to-delivery interval in women allocated to progesterone than those allocated to placebo (Wilcoxon P = .043, log-rank P = .057). However, there was no difference in the frequency of preterm delivery at 32 weeks or less.

With the second cutoff (<28 mm), patients who had received vaginal progesterone had a lower rate of spontaneous preterm delivery at 32 weeks or less of gestation, though this was not observed for preterm delivery defined as 35 weeks or less or 37 weeks or less. It is noteworthy that the frequency of newborn intensive care unit admission was lower in women with a cervical length of 30 mm or less and less than 28 mm and who had received progesterone treatment than in those allocated to the placebo group. The same was the case for the duration of stay in the newborn intensive care unit.²⁴⁷ This analysis provides the first hint that vaginal progesterone administration may improve infant outcome in properly selected patients. It is important to stress, however, that these conclusions must be considered tentative because they derive from a secondary analysis, which is intended to be hypothesis generating.²⁴⁹ Therefore, further investigation is necessary. Specifically, randomized clinical trials designed to test the effect of progestins in women with a short cervix are justified.

Cervical sonography is the most objective and reliable method to assess cervical length. The shorter the sonographic cervical length in the mid-trimester, the higher the risk of spontaneous preterm labor/delivery.^{45,49,50,56,60,61,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83, and 84} There is no agreement as to what is a sonographic short cervix. For example, Iams et al⁴⁹ proposed that a cervix of 26 mm or shorter at 24 weeks of gestation increases the risk for spontaneous preterm delivery (RR 6.19; 95% CI = 3.84 to 9.97). The prevalence of spontaneous preterm delivery (defined as <35 weeks) in this study was 4.3%, and the positive predictive value was 17.8% for a cervical length 25 mm or less at 24 weeks of gestation.⁴⁹ Thus, most women with a short cervix (defined as ≤25 mm) and no history of previous preterm birth will not deliver a preterm neonate. Other investigators have proposed a cutoff of 15 mm,⁶⁶ because a cervical length of 15 mm or less is associated with nearly a 50% risk of spontaneous preterm delivery at 32 weeks of gestation or less, when neonatal morbidity is substantial.^{70,117,136,148,216,237,250,251} To et al²⁵¹ have developed a method to assess the risk of preterm delivery for individual patients using sonographic cervical length and other maternal risk factors such as maternal age, ethnic group, body mass index, cigarette smoking, and previous cervical surgery.

Celik et al²⁵² have proposed the use of an individualized risk assessment for the prediction of preterm birth. The authors conducted a large, prospective observational study of over 58,000 patients. Transvaginal sonographic cervical length screening was performed at 20 to 24 6/7 weeks. All patients were receiving routine antenatal care and were given the option to participate in one of 2 randomized clinical trials of intervention (cerclage or progesterone) if they had a sonographic cervical length of 15 mm or less. Patients who had received the active intervention of either trial were excluded from the risk assessment study. Obstetrical history was subdivided into 5 groups: (1) primigravida or only losses prior to 16 weeks; previous preterm delivery (defined by earliest preterm birth) at (2) 16 to 23 weeks; (3) 24 to 33 weeks; (4) 34 to 36 weeks; or (5) term (>36 weeks). Adjusted odds ratios were calculated for spontaneous delivery at less than 28 weeks (extreme), 28 to 30 weeks (early), 31 to 33 weeks (moderate), and 34 to 36 weeks (mild) using logistic regression analysis. The authors concluded that cervical length was the best predictor of spontaneous preterm birth. The risk estimation was improved by the addition of obstetric history, but not any maternal characteristics or other interactions. This important study includes tables that provide likelihood ratios to allow for the risk assessment of spontaneous preterm birth. The detection rates of extreme, early, moderate, and mild spontaneous preterm birth by combining the obstetric history and cervical length were 80.6%, 58.5%, 53.0%, and 28.6%, respectively, with a 10% screen-positive rate.²⁵² This model provides for a higher detection rate than obstetric history or cervical length alone, and allows for population-based screening, which, in a setting of potential effective interventions, may have great utility.

Importantly, sonographic cervical length is the single most powerful predictor for preterm birth in the index pregnancy and is far more informative than a history of previous preterm birth.^66,70,252 This has implications in the selection of patients for future trials and the interpretation of past trials. Specifically, all trials conducted to date^{243,248,249,250, and 251} but one,²⁴⁵ have identified patients for study based upon a history of previous preterm birth. In as much as preterm birth is a syndrome, it is likely that the risk conferred by a previous preterm delivery may vary within and between populations. This may explain the apparent contradictory results between trials (ie, conflicting reports of trials of vaginal progesterone and the wide range of preterm birth among control groups).^248,255,256

Sonographic cervical length is not a screening test for spontaneous preterm delivery, because only some of the patients who will have a spontaneous preterm birth have a short cervix in the mid-trimester. However, sonographic cervical length is a method for risk assessment for spontaneous preterm delivery. Its importance derives from the observation that only a fraction of patients with a history of previous preterm birth who shorten their cervix during the index pregnancy may benefit from treatment.^212,247

In most circumstances, cervical assessment can be adequately performed by two-dimensional ultrasound. However, there are limited scan planes that can be obtained due to the anatomic limitations of the transvaginal approach. Three-dimensional (3D) ultrasound can provide innumerable planes that cannot be obtained by conventional 2D ultrasound. From a stored data set, it is possible to reconstruct any plane required for examination.

Hoesli et al²⁵⁷ compared the 3D volume assessment of the cervix with the two-dimensional cervical length measurement in low- and high-risk patients. The high-risk patients were 27 in-patients with preterm contraction and/or PROM between 22 and 34 weeks of gestation. The low-risk patients included 28 out-patients with an uncomplicated pregnancy without contraction, vaginal infection, and history of preterm contractions. Patients in the high-risk group delivered earlier than patients in the low-risk group (mean, 37.9 ± 3.6 vs 39.2 ± 1.7; P = .047). Using the two-dimensional (2D) technique, the mean cervical length in high-risk patients was shorter than that in low-risk patients, but 3D volume measurement did not show a statistically significant difference. In contrast, Bega et al²⁵⁸ studied 21 multiparous patients at risk for preterm delivery with 37 cervical assessments by 2D- and 3D ultrasound between 11 and 32 weeks of gestation. The author reported that in 10 instances (27%) the cervical length measurement was different between the 2 techniques. Interestingly, of 21 examinations showing funneling, 6 (29%) of these were detected only by 3D ultrasound. The mean funnel width as measured in the coronal 3D ultrasound plane was wider than that measured by 2D ultrasound (15.9 mm vs 13.0 mm; P < .05) and that obtained by sagittal plane (15.9 mm vs 12.9 mm; P < .05).

Severi et al²⁵⁹ prospectively evaluated 103 singleton pregnancies at 18 to 34 weeks. Three-dimensional cervical length was significantly correlated with interval between ultrasound examination and delivery as well as gestational age at delivery. A 3D cervical length of 35 mm in the second trimester was noted to be the optimal cutoff for preterm birth prediction (sensitivity 100%, specificity 88%).²⁵⁴

Others have confirmed the use of a higher cervical length as the threshold for optimal prediction of spontaneous preterm birth when using 3D ultrasound. Dilek et al²⁶⁰ evaluated the use of cervical volume measurements in 250 women at 22 weeks' gestation. The authors reported a cutoff value of 33.15 mm using cervical volume measurements with a sensitivity of 80% and 12.7% false-positive rate. There was no difference in the sensitivity and positive predictive value in the use of cervical volume when compared to cervical length.²⁶⁰

A comprehensive reference range for cervical volume and power Doppler parameters in normal pregnancy was reported by Rovas et al.²⁶¹ This large study included 352 nulliparous and 291 parous patients at 17 to 41 weeks. A representative sample was obtained by examining 25 patients at each gestational age. All patients delivered at 37 or more weeks. The authors noted no change in the cervical volume over the gestational age range of 17 to 40 weeks and established reference values for cervical volume and 3D power Doppler parameters.²⁶¹

Recently, Basgul et al²⁶² compared nulliparous and multiparous cervices by the use of 3D ultrasound and the virtual organ computer-aided analysis (VOCAL™) tool. The authors calculated the volume and the power Doppler vascularization index, flow index, and vascularization flow index. Although there were no differences between nulliparous and multiparous patients in any of the characteristics, the application of this technology to predict pregnancy outcome requires further study.

A decrease in total collagen content and an increase in collagen solubility and collagenolytic activity characterize cervical change during pregnancy and labor and delivery. A quantitative method to determine cervical ripening analyzing the collagen content of the cervix through fluorescent spectroscopy has been proposed recently.²⁶³ The collascope is an optical device that determines cervical collagen content by means of a fluorescent signal generated whenever collagen crosslinks are excited by a light whose wavelength is about 340 nm.²⁶³ The system has been used in rats and in humans at different stages of pregnancy, and has shown that cervical ripening occurs progressively in the last trimester of pregnancy. Further studies are required to determine the value of this technology in obstetrics and whether it adds information not provided by ultrasound.

Tekesin et al²⁶⁴ reported the use of quantitative ultrasonic tissue characterization in 46 pregnant women to assess the texture features of the uterine cervix. In patients with premature contraction and shortening of cervix (n = 16), the average gray-scale values were lower than those obtained from asymptomatic patients (n = 30) at the same gestational age.

The most important challenge in the field of sonographic evaluation of the uterine cervix is the development of a model to refine the current method for the prediction of spontaneous preterm birth. It has yet to be determined if there are some factors that can be used in conjunction with sonographic cervical length to improve preterm birth prediction. These adjunct tools may include the following:

1. Is individualized risk assessment possible with the use of additional parameters to cervical length (ie, serum markers, previous pregnancy history, etc)?

2. Is universal transvaginal sonographic screening warranted?

3. Is noninvasive assessment of collagen content able to provide prognostic information?

   • - Fluorescence spectroscopy (collascope)

   • - Quantitative ultrasonic tissue characterization of the cervix

4. Does the use of cervical artery Doppler velocimetry and/or cervical gland evaluation²⁶⁶ add predictive value?

5. Is the knowledge of uterine electrical activity at the time of sonographic evaluation a useful tool for preterm birth prediction?²⁶⁷

6. Does the use of VOCAL™ improve the sonographic evaluation of a uterine cervix?

In summary, a large body of literature has examined the nature of cervical change during pregnancy in both human and animal studies. Cervical disease is just one of the etiologies of the preterm parturition syndrome. The success of the treatment of cervical disease in pregnancy, with specific interventions such as a cervical cerclage or progesterone therapy, requires an accurate diagnosis. Cervical cerclage or progesterone may benefit some patients who have a primary or important cervical process and are in the reversible phase of parturition. It is clear that obstetrical history is insufficient to identify such a patient. Cervical sonography represents an important advance toward the identification of the patient with premature cervical ripening. Additional studies such as the biochemical definition of the short cervix may contribute to the ability to discriminate the patient with a short cervix who will deliver a preterm neonate from the one who will deliver a term infant, as well as the patient who may benefit from a cerclage or progesterone therapy.

Acknowledgments

This work was supported (in part) by the Perinatology Research Branch, Division of Intramural Research, Eunice Kennedy Shriver National Institute of Child Health and Human Development, NIH, DHHS.