Cervical Insufficiency

Cervical insufficiency has been traditionally defined as painless cervical dilation that leads to mid-trimester pregnancy loss in the absence of other causes (e.g., labor, trauma, major fetal anomaly, stillbirth). Using this definition, the incidence of cervical insufficiency in the general obstetric population is between 1/100 and 1/2000 births.^{1, 2, 3, 4} The most traditional definition requires that pregnancy losses be recurrent; unfortunately, the history-based definition of cervical insufficiency requires that a patient experience a pregnancy loss.

In recent years, the definition of cervical insufficiency has expanded to include women with a prior spontaneous preterm birth and evidence of cervical shortening (<25 mm) on transvaginal ultrasound. When this criterion is included in the definition, the incidence increases to 3–4% of the obstetric population.⁵

Clinical reports of mid-trimester pregnancy loss attributable to a primary cervical etiology appeared as early as the 17th century; the term “cervical incompetence” was first used in The Lancet in 1865.⁶ The contemporary concept of cervical insufficiency was not widely accepted until the middle of the 20th century after Palmer⁷ in 1948 and Lash⁸ in 1950 independently described interval repair of anatomic cervical defects associated with recurrent mid-trimester loss. Soon thereafter, Shirodkar⁹ in 1955, McDonald¹⁰ in 1957 and, later, Benson¹¹ in 1965 described the mid-trimester cerclage procedures now utilized in obstetric practice. Other investigators reported success with alternate surgical procedures including the Wurm bridging procedure by Hefner,¹² cone-electrocautery scarification of the external os by Barnes,¹³ and scarification with suture-closure of the external os by Baden.¹⁴

For the remainder of the 20th century, investigators focused primarily on pregnancy outcomes in cohorts managed with transvaginal (and less commonly, transabdominal) cerclage procedures. These cohorts were not referenced to appropriate control groups (i.e. women with a similar presentation randomized to no intervention), but rather only compared to the cohorts’ own obstetric histories or data from analogous reports. In short, the 20th century literature on women with a history of cervical insufficiency has been largely a chronicle of surgical methods to correct anatomic disruption of the internal os in women who had experienced recurrent painless dilation and mid-trimester birth.

The syndrome of spontaneous preterm birth

Spontaneous preterm birth may be best characterized as a syndrome comprising several anatomic and related functional components.¹⁵ These include the uterus and its myometrial contractile function (e.g., preterm labor), decidual activation and loss of chorioamnionic integrity (e.g., preterm membrane rupture) and, finally, diminished cervical competence, from either a primary anatomic defect or an early pathologic cervical ripening (e.g., cervical incompetence or cervical insufficiency). In a particular pregnancy, a single anatomic feature may appear to predominate, even though it is more likely that most cases of spontaneous preterm birth result from the interaction of multiple stimuli and pathways which culminate in the recognized clinical syndrome. Nevertheless, the relative importance of these (and possibly other as yet unidentified) components can vary, not only among different women, but also in successive pregnancies of an individual patient.

The normal cervix

The cervix plays an essential role in maintenance of a normal pregnancy; disruption of the normal anatomic structure or biologic function of the cervix may foretell preterm delivery. The cervix is composed primarily of fibrous connective tissue; this fibrous band is felt to be the chief mechanical barrier against the loss of the enlarging products of conception. The cervix and its mucous glands also play an important immunologic role in preventing organisms from ascending into the normally sterile intrauterine environment.

Early in pregnancy the cervix begins to undergo structural changes with increased vascularity, stromal hypertrophy, glandular hypertrophy and hyperplasia, and structural changes of the extracellular matrix.^{16, 17, 18, 19} Conformational changes in cervical collagen, with alterations in the number and types of cross-links between collagen triple-helices, lead to a steady increase in tissue compliance throughout pregnancy. Those who have alterations in the process, due heritable defects in collagen and elastin synthesis or assembly, have an increased risk of cervical insufficiency.^{20, 21, 22, 23, 24}

The biologic continuum of cervical insufficiency

As early as 1962 Danforth and Buckingham suggested that cervical insufficiency was not an all or none phenomenon.²⁵ Rather, it comprised degrees of insufficiency, and combinations of factors could cause “cervical failure”. This hypothesis never gained wide acceptance in spite of the obvious heterogeneity observed in clinical practice. Cervical insufficiency was generally viewed as dichotomous (possibly because available treatment strategies were similarly devised).

Based on analyses of cervical tissue for collagen/muscle ratios and recognizing the contribution of past traumatic disruption of cervical continuity, these investigators proposed a classification scheme to explain the clinical syndrome of cervical insufficiency.²⁵ In the proposed classification, one group of patients had ostensibly normal cervical tissue, whose integrity as an anatomical fibrous ring had been previously damaged. Such defects were usually the result of antecedent obstetric trauma or complications of dilation and curettage. These might even be concealed by a normal-appearing external os and ectocervix. The second group possessed an abnormally low collagen-muscle ratio that would compromise its mechanical function and lead to premature dilation. This so-called “muscular cervix” was probably a congenital problem and would manifest as a poor reproductive history in women with no prior cervical injury. The third group comprised women who had no history of antecedent trauma and who also had normal collagen-muscle ratios, but whose obstetric histories mimicked those of groups 1 and 2, presumably from stimuli which cause premature cervical ripening.

In a subsequent report these investigators analyzed cervical biopsies taken from postpartum women and compared them to hysterectomy specimens from non-pregnant patients.²⁶ Pregnancy was associated with increased water content, a marked decline in collagen and glycoprotein, and increased glycosaminoglycans. The cellular and biochemical changes suggested that cervical dilation in pregnancy is a dynamic process, and this might explain why a woman could have a pregnancy outcome consistent with cervical insufficiency in one pregnancy, but then without treatment, have a subsequent term birth. Presumably, the factor(s) inciting the pathologic cervical changes would vary among pregnancies. Women with a more muscular cervix might have an unusual susceptibility or lower threshold for the effects of the factors which precipitate the clinical syndrome of preterm birth.

These earlier observations were enlarged by Leppert and colleagues²⁷ who reported an absence of elastic fibers in the cervixes of women with clinically well-characterized cervical insufficiency on the basis of their reproductive history. Conversely, cervical biopsy specimens from women with normal pregnancies showed normal amounts and orientation of these elastic fibers. Supporting these microscopic observations was the biochemical finding of lower desmosine levels in the women with cervical insufficiency, indicating lower elastin content in the specimens. Rechberger and colleagues²⁸ also compared cervical biopsy specimens among non-pregnant controls, women in the mid-trimester with clinically defined cervical insufficiency and normal postpartum gravidas. Compared to normal postpartum patients, they found increased collagen extractability and collagenolytic activity in women with cervical insufficiency, suggesting a high collagen turnover characterized by higher proportions of newly synthesized collagen with lower mechanical strength. It is unknown whether these microstructural and biochemical phenomena were congenital, acquired from previous trauma or the result of other pregnancy-associated processes. Collectively, these biochemical and ultrastructural findings support the variable, and often unpredictable, clinical course of women with a history of cervical insufficiency.

Although the traditional paradigm has depicted the cervix as either competent or insufficient, recent evidence, including clinical data^{29, 30, 31, 32} and interpretative reviews^{33, 34, 35} suggest that, as with most other biologic processes, cervical competence is rarely an all or none phenomenon and most likely functions along a continuum of reproductive performance. Although some women have tangible anatomic evidence of poor cervical integrity, most women with a clinical diagnosis of cervical incompetence have ostensibly normal cervical anatomy. In a proposed model of cervical competence as a continuum, a poor obstetric history results from a process of premature cervical ripening, induced by a myriad of underlying factors, including subclinical infection, inflammation, local or systemic hormonal effects, or even genetic predisposition. If and when cervical integrity is compromised, other processes may be stimulated, appearing clinically as other components of the spontaneous preterm birth syndrome (i.e., premature membrane rupture, or preterm labor). A decision as to whether diminished cervical competence arises via primary endogenous mechanical deficiencies or exogenous factors, might define the optimal therapy. A more cogent approach to the concept of cervical competence is to view the cervix as an interdependent participant in the multifactorial model of the spontaneous preterm birth syndrome. The pragmatic clinical question that emerges from this paradigm is whether an individual patient has a significant (and presumably treatable) component of diminished cervical competence.

While the historic concept of the diagnosis and treatment of cervical incompetence often includes women with past cervical trauma from birth-associated lacerations, operative injury or cervical amputation, the prevalence of these antecedent events appears to be decreasing in contemporary practice. Obstetric care and surgical technique have improved, clinicians recognize the importance of gradual cervical dilation prior to curettage and cervical amputation is rarely performed. Cervical damage from trauma may be visible or palpable on a careful examination of the cervix before or during the early part of pregnancy, and if detected, should be managed with surgical repair if associated with a prior poor obstetric history or if the defect is believed to involve the internal cervical os.

Diethylstilbestrol (DES) exposure in utero

DES, a synthetic estrogen designed to supplement a woman’s natural estrogen and prevent miscarriage or preterm delivery, was used by an estimated 5−10 million people between 1938 and 1971.³⁶ As this medication has not been used since the early 1970s, women of reproductive age now have not been exposed in utero. However, approximately two-thirds of women exposed to DES in utero have reproductive tract anomalies and a significantly increased risk of cervical insufficiency.^{37, 38, 39}

Mechanical cervical dilation

Mechanical cervical dilation is commonly performed during mid-trimester dilation and evacuation procedures. Assessing the effect of mechanical cervical dilation on the risk of cervical insufficiency is difficult. First, an appropriate comparison group is difficult to determine. The most common comparison groups used are women with no prior pregnancies and women with a prior term birth. Perhaps the most appropriate comparison group would be women with a mid-trimester induction of labor with no associated surgical procedure; this cohort would most closely resemble women undergoing mid-trimester dilation and evacuation in both sociodemographics and pregnancy history. We were unable to identify a study utilizing this comparison group. Second, ascertainment of the outcome of cervical insufficiency and the exposure is difficult. This type of study requires a large cohort, typically large population-based databases. Cervical insufficiency may not be recorded in these databases as a diagnosis, and preterm delivery is frequently used as a surrogate marker. Additionally, many have used birth certificate databases, which may include only live births after 20–24 weeks, excluding those with a mid-trimester loss who did not reach this gestational age and may have a true diagnosis of cervical insufficiency. Large databases may only record gestational age and delivery, rather than type of procedure, making ascertainment of the exposure (mid-trimester mechanical cervical dilation) difficult.

In a well-performed review by Atrash et al.⁴⁰ that considered the above issues, dilation and evacuation procedures were associated with an increased risk of mid-trimester spontaneous loss. A more recent systematic review and meta-analysis suggested a slight increase (odds ratio 1.36, 95% CI 1.24–1.50) in the risk of preterm delivery after induced termination of pregnancy which was increased after multiple induced terminations (OR 1.93, 95% CI 1.28–2.71).⁴¹ In 2013, a study using the Scottish Morbidity Record and Scottish Stillbirth and Death Survey demonstrated that: (1) the risk of early spontaneous preterm delivery increased with the number of prior terminations of pregnancy and (2) this risk decreased over time in conjunction with any increases in medical abortions and cervical pretreatment (ripening) prior to surgical abortions.⁴² By the year 2000, the association of termination of pregnancy and early spontaneous preterm delivery in the following pregnancy was not apparent. Thus, while forced mechanical dilation of the cervix may be associated with a slight increase in the risk of preterm delivery and/or cervical insufficiency, newer operative techniques may mitigate that risk.

Loop electrosurgical excision procedures, cervical conization, and cervical dysplasia

Prior treatment for cervical dysplasia using loop electrosurgical excision procedures (LEEP), cold-knife conization, or laser conization to remove a portion of the cervix is plausibly a risk factor for cervical insufficiency. For similar reasons as examining the effect of mechanical dilation, estimating the effect of LEEP/conization is difficult: a randomized control trial is not possible, identifying a proper control group is difficult, details of the procedure are unlikely to be available in a large database, and preterm birth is frequently used as a surrogate marker for cervical insufficiency.

A multitude of retrospective studies have been published with varying results.^{43, 44, 45, 46, 47, 48, 49, 50, 51, 52} A systematic review and meta-analysis of 27 studies was performed in 2013.⁵³ All studies that compared women who had undergo an ablative or excisional conization for cervical intraepithelial neoplasia or stage IA1 cervical cancer to an untreated group and examined the outcome of preterm birth were included. In eight studies with a total of 3141 subjects, LEEP was associated with an increase preterm delivery <37 weeks in a relative risk of 1.70 (95% CI 1.24–2.35). In eight studies with 28,378 subjects, cold knife conization was associated with an increased risk of preterm delivery with a pooled relative risk of 2.59 (95% CI 1.80–3.72). Laser conization and ablation were not associated with an increased risk of preterm delivery (pooled RR 1.71, 95% CI 0.93–3.14 and RR 0.87, 95% CI 0.63–1.20). In another systematic review and meta-analysis, Kyrgiou examined the effect of excisional procedures on second trimester miscarriage rates; in this analysis of eight studies that included a total of 16,558 excisional procedures, second trimester miscarriage was increased after an excisional procedure (pooled RR 2.60, 95% CI 1.45–4.67).⁵⁴

As cervical excisional procedures purportedly increase the risk of cervical insufficiency through mechanical/structural alteration of the cervix, one would presume that either the quantity of tissue removed from the cervix would be associated directly with the risk of preterm delivery and/or that cervical length after excisional procedure would be predictive of preterm delivery risk. Indeed, the few published studies that have examined the amount of tissue excised at the time of the procedure and the risk of preterm delivery did find a correlation between the two.^{52, 55, 56}

Studies examining the mid-trimester cervical length consistently show that, compared to women with no prior surgical procedure, women with a prior excisional procedure had a shorter cervical length and in some studies a higher incidence of short cervix (<25 mm or <30 mm, depending on the study).^{57, 58, 59, 60, 61} Miller et al. performed a multivariable analysis to assess the independent contributions of a cervical excisional procedure and a shortened cervix; they concluded that both were independently associated with the risk of preterm birth and that the increase in risk of preterm birth after an excisional procedure is not solely through a shortened cervix mechanism.⁶⁰

Conner et al. proposed that the risk of preterm birth after excisional procedure may be associated with the biological mechanism of dysplasia.⁶² In a systematic review and meta-analysis, they considered the presence or absence of cervical dysplasia in the control group. When women with a LEEP were compared to women with either no history or an undocumented history of cervical dysplasia, LEEP was associated with an increased risk of preterm birth (pooled RR 1.86, 95% CI 1.58–2.21). When women with a history of a LEEP were compared to women with a history of cervical dysplasia, LEEP was not associated with an increased risk of preterm birth (pooled RR 1.08, 95% CI 0.88–1.33).

In sum, a history of a cervical excisional procedure is associated with a small increase in the risk of preterm birth, although the mechanism of the increased risk is unclear.

Prophylactic treatment based on risk factors

Cervical length screening and/or treatment for cervical insufficiency with cerclage on the basis of one or more of the above risk factors has not been proven effective and is not indicated.

Patient history

Traditionally, cervical insufficiency is a clinical diagnosis based on the history of a spontaneous midtrimester loss associated with painless cervical dilation. The burden of the diagnosis of cervical insufficiency is on the physician managing the patient at the time of the mid-trimester loss. Other causes of midtrimester loss should be excluded (labor, stillbirth, abruption, trauma, fetal anomaly, genetic causes, and uterine anomalies). As multifetal gestations are at high risk of mid-trimester loss, a diagnosis of cervical insufficiency should not be made after a midtrimester loss of a multifetal gestation. Careful documentation of patient presentation (presentation with hourglassing membranes, vaginal discharge, vaginal spotting, rupture of membranes, presence or absence of contractions, presence or absence of fever) is necessary. A diagnosis of cervical insufficiency should not be made simply in the presence of additional risk factors (including a current multifetal gestation or prior midtrimester loss of multifetal gestation). Additionally, as records from prior deliveries are not always available, informing the patient of her diagnosis and counseling her prior to discharge from the hospital that treatments may be available early in future pregnancies to prevent recurrent pregnancy loss is of the utmost importance.

Sonographic diagnosis

The difficulty of diagnosing cervical insufficiency by patient history alone has led physicians to seek other, more objective diagnostic measures. Various sonographic findings including cervical length, funneling at the internal os, and dynamic response to provocative maneuvers (e.g., fundal pressure) have been utilized to select women for treatment, generally cerclage. In most of these earlier reports, the sonographic evaluations were not blinded, leading to uncontrolled interventions and difficulty determining their value. In many instances the sonographic criteria for cervical incompetence were only qualitatively described and thus, may not be reproducible. Table 1 depicts published reports linking the findings from cervical sonography to a diagnosis of incompetence. Note that the diagnostic criteria are disparate and in many cases, not described in a quantitative fashion.

Table 1. Published reports of the sonographic diagnosis of cervical incompetence

GA, gestational age; DES, diethystilbestrol; N/S, not stated; MT, midtrimester; CL, cervical length; CI, cervical incompetence

Subsequently large, blinded, and reproducible observational studies investigating the relationship between mid-trimester cervical length and the risk of preterm birth were published.^{29, 79, 80, 81, 82, 83} Iams and his colleagues in the NICHD Maternal-fetal Medicine Units Network completed a study of 2915 unselected women with a singleton pregnancy that underwent a blinded cervical sonographic evaluation at 22–24 weeks’ gestation.²⁹ The relative risk of spontaneous preterm birth increased as cervical length decreased. In spite of this significant relationship, as a test for predicting spontaneous preterm birth <35 weeks at a cervical length cutoff of <26 mm (the population 10th percentile), the sensitivity was low (37%), and the positive predictive value was poor (18%). These findings were confirmed by Taipale and colleagues in an unselected, but generally low-risk Finnish population.⁸⁰ Both groups concluded that cervical sonography was not a viable screening test for spontaneous preterm birth in a low-risk population.

Owen and colleagues examined the utility of cervical ultrasound as a predictor of spontaneous preterm birth at <35 weeks in high-risk women.⁸³ Participants included gravidas with a prior spontaneous preterm birth at <32 weeks. Women believed to have cervical insufficiency (based on a clinical history) were not eligible. Beginning at 16–18 weeks of gestation, 163 gravidas underwent serial, biweekly sonographic evaluations until the 23rd week of gestation. Fundal pressure was also utilized as a provocative measure to determine whether dynamic cervical shortening could be induced. This study design permitted analysis of the shortest observed cervical length over time which also included both spontaneously occurring and fundal pressure-induced cervical length shortening. As in the previous studies by Iams²⁹ and Taipale⁸⁰the risk of spontaneous birth increased as cervical length shortened. However, in this high-risk population, the sensitivity increased to 69% and the positive predictive value to 55% at a cervical length cutoff of <25 mm (the 10th centile at the 16–18 weeks, intake evaluation). In other words, 55% of women with a prior preterm birth <32 weeks and a cervical length <25 mm at their first evaluation delivered prior to 35 weeks.

In sum, cervical sonography performs poorly as a screening test in low-risk women but it appears to have significant clinical utility in high-risk women, defined as a prior early spontaneous preterm birth. Whether cervical ultrasound has similar predictive values in other populations of at-risk women (e.g., DES, prior cervical surgery, uterine anomalies, multiple induced abortions, etc.) remains speculative, as it has not been well-studied. Some investigators have included women with these risk factors in study populations comprised primarily of women with prior spontaneous preterm birth, but the results could not been subcategorized because of small sample sizes.⁸¹

Use of cervical ultrasound in twin gestations has also been reported; however, the test characteristics, especially sensitivity and positive predictive value (<40%), appear to be generally lower than for women with a prior early spontaneous preterm birth.^{84, 85} Other case series have documented low (<10%) incidences of significant mid-trimester cervical length shortening in populations of multiple gestations.^{84, 86, 87, 88}

Therefore, a diagnosis of cervical insufficiency may be made in a woman with a singleton pregnancy with a history of prior preterm birth (of a singleton gestation) and a mid-trimester (<24 weeks) cervical length <25 mm.

Physical exam

Occasionally, a woman will present with symptoms and physical findings that support an antepartum diagnosis of acute cervical insufficiency. This typically consists of vague symptoms of pelvic pressure, discharge, spotting, and/or increased urinary frequency. Physical exam demonstrates a cervical dilation of 1–4 cm and fetal membranes visible at or beyond the external os. These patients warrant monitoring for intrauterine infection (fever, uterine tenderness, leukocytosis) and preterm contractions/labor.

Diagnostic tests

Because the diagnosis of cervical insufficiency has been determined primarily from past reproductive performance and physical examination findings, the obvious limitations associated with the clinical diagnosis have prompted the search for sensitive and specific tests that could be applied in a prospective manner to women deemed at risk for cervical insufficiency, thus obviating recurrent pregnancy loss as a diagnostic criterion. Such a test would ideally provide a timely diagnosis and the potential for optimal therapeutic intervention.

To be clinically useful, a scoring system, or other test for cervical insufficiency should be able to accurately discriminate between women who have a condition amenable to a proven effective form of therapy and women with other causes of poor obstetric outcome. Even in the absence of effective therapies, such an assessment would be useful to target at‑risk populations, whose optimal treatments could be determined from controlled intervention trials.

Most of the earlier reported tests for cervical insufficiency were based on the functional anatomy of the interval os in the non-pregnant state and are of historical interest. In 1953, Rubovits and colleagues utilized an intrauterine balloon filled contrast media and imaged with a radiograph to discriminate the normal from the incompetent lower uterine segment and internal os.⁸⁹ Mann, in 1959 reported a new method for the diagnosis of cervical incompetence which used an intrauterine balloon with a contrast agent and subjected to filling pressures of 200–260 mmHg.⁹⁰ This, he believed, overcame the limitations of earlier diagnostic studies that utilized plain hysterosalpingography by measuring the internal os dimensions under pressure. As this evaluation yielded an internal os measurement, a continuous variable, he chose what appeared to be an arbitrary cutoff point to define insufficiency.

In 1973 Peterson and Keifer proposed another improvement after concluding that Mann’s technique described above would artificially cause cervical and uterine dilation using pressures was “fraught with danger.”⁹¹ Their apparatus was similar to Mann’s, but the intrauterine balloon had a second port open to the atmosphere and limited the intrauterine pressure. Studying 55 women with suggestive histories, these investigators excluded the diagnosis of insufficiency in 16 women when the maximum diameter of the cervical canal was less than 10 mm. They followed these women in successive pregnancies and observed 22 neonatal survivors, whereas their prior neonatal loss rate had been 52%; by inference, the etiology of these prior losses was not cervical insufficiency. In the women with cervical canal diameters ≥10 mm, prophylactic cerclage, either before or after conception, was performed, and these 26 women had 31 successful pregnancies. Since the subsequent loss rates could not be determined (the denominators were not reported), it was difficult to understand the magnitude of the benefit attributable to the test results or the cerclage. Other attempts at objective assessments included passage of a #8 Hegar dilator into the non-pregnant cervical canal without resistance⁹² and traction forces required to dislodge a Foley catheter whose balloon was placed above the internal os and filled with 2–3 mm of water.⁹³ Subjectively effortless passage of the dilator or removal of the Foley balloon with <600 g of force would confirm an objective diagnosis.

More recently, in 1988, Kiwi and colleagues estimated the elastic properties of the non-pregnant cervix in two cohorts of women: 247 women with a poor obstetric history (defined as ≥3 first trimester abortions, ≥1 second trimester loss, ≥1 prior preterm birth, or some combination thereof) and 42 controls who may have had no more than two first trimester abortions, and no prior preterm births.⁹⁴ Cervical elastance was derived from the pressure–volume slope of a 4 mm balloon placed in the endocervical canal and inflated to pressures as high as 150 mmHg. Although women in the poor obstetric history group had significantly lower elastance values than the controls, there was significant overlap between the two groups. They reported no subsequent pregnancy outcomes, proposed no clinically useful cutoff for their evaluation and could only suggest that such objective evaluation might ultimately prove to be clinically useful to accurately select patients for prophylactic cerclage.

In 1993 Zlatnik and Burmeister reported their experience with a cervical compliance score (analogous to an Apgar score) derived from the results of three other tests: hysterosalpingography, passage of a #8 Hegar dilator, and intrauterine balloon traction, performed in 138 non-pregnant women.⁹⁵ Their histories included prior delivery <34 weeks following clinically diagnosed preterm labor or preterm membrane rupture, with or without antecedent bleeding. A small portion of their cohort had a questionable history of cervical insufficiency. Scores could range from 0 to 5, and women with low scores ≤2 were more likely to have been delivered at 27–34 weeks as compared to women with higher scores who were more likely to have been delivered in the mid-trimester at 14–26 weeks (p <0.01). In subsequent pregnancies, whose management and outcomes were ascertained by a mail-out questionnaire (83/138 responded), cerclage was recommended in most women with a score of 3 and in all women with a high score >3; 21/27 underwent surgery. Surprisingly, more women with scores ≥3, 78% of whom received cerclage, were delivered at 14–29 weeks' gestation (p = 0.07) casting doubt on its clinical utility to appropriately select patients for cerclage (under the presumption that cerclage is an effective therapy).

Others evaluated the utility of scoring systems to aid in patient selection for cerclage procedures, intending to increase the specificity of the clinical diagnosis. In 1976 Block and Rahal reported 31 pregnancies in 25 women who received prophylactic cerclage for a diagnosis of cervical incompetence.⁹⁶ The score increased based on a greater number of clinical or historical indications for cerclage, and by inference, a more specific diagnosis of cervical insufficiency. Women with high scores had a longer cerclage-to-delivery interval as compared to women with lower scores. Women with higher scores also had higher overall pregnancy success rates with cerclage than did the cohort of women who had fewer indications and lower scores. In 1991 Ger and colleagues performed an analogous study of 47 women who received a McDonald cerclage based on their obstetric histories and physical examinations, but not necessarily for a clinical history of cervical insufficiency.⁹⁷ Based on their obstetric histories and examination findings, scores could range from 1 to 5. Similar to Block and Rahal, they observed longer cerclage-to-delivery intervals in women with high scores >3 as compared to women with lower scores. While these observations support the multifactorial model of spontaneous preterm birth described earlier, these investigators could provide no practical recommendations.

All such attempts at providing an objective diagnosis of cervical insufficiency failed for a number of reasons, and none of these tests are in common use today. These reports generally suffered from a failure to evaluate standard test characteristics (i.e., sensitivity, specificity) against some reference standard for the diagnosis or another clinically relevant endpoint. They often utilized subjective criteria, caused patient discomfort, required ionizing radiation, were cumbersome and were generally only applicable in the non-pregnant state; many women at risk for cervical insufficiency present for care after a subsequent conception. Moreover, none of these tests could reasonably predict pregnancy-associated conditions that would lead to premature ripening and cervical dilation. Finally, since there is no universally applicable standard for the diagnosis of incompetence, and since the results of such tests were never evaluated and linked to a proved effective treatment, their clinical utility was, at best, theoretical.

The goal of screening is to identify those at risk of a disease before the onset of symptoms; the goal of screening for cervical insufficiency therefore is to identify women at risk of mid-trimester pregnancy loss or preterm delivery prior to the onset of contractions, rupture of membranes, or delivery. An effective screen program should (1) screen for a prevalent condition, (2) accurately identify those at risk for the condition, and (3) offer an effective treatment for the condition identified. The accepted method of screening for cervical insufficiency is transvaginal ultrasound;^{98, 99} who should be screened for cervical insufficiency using sonography remains controversial.

Transvaginal ultrasound

Transvaginal ultrasound is a screening test for cervical insufficiency that is safe, acceptable, and reliable. Those responsible for measuring transvaginal cervical lengths should undergo the Cervical Length Education and Review (CLEAR)¹⁰⁰ program sponsored by the Society of Maternal Fetal Medicine through its Perinatal Quality Foundation and follow the standards for measurement in this program. When measured appropriately, transvaginal cervical length has an inter- and intra-observer variability of less than 10%.¹⁰¹

Transabdominal ultrasound

Transabdominal ultrasound is not considered an acceptable screening method for cervical insufficiency. Transabdominal ultrasound is less sensitive than transvaginal ultrasound, and image quality may be poorer than transvaginal ultrasound.¹⁰²

Who should be screened?

In an unselected population of singleton pregnancies, a short cervical length (≤25 mm) has a positive predictive value for preterm birth prior to 35 weeks of only 17.8%; in other words, only 17.8% of women with a cervical length ≤25 mm at 22–24 weeks' gestation will deliver prior to 35 weeks.²⁹ When women with singleton pregnancies and a prior preterm birth <32 weeks are considered, the positive predictive value of a cervical length <25 mm for preterm birth prior to 35 weeks increases to 75%, although the sensitivity remains low at 19%.⁸³ In other words, in women with a singleton gestation and a prior spontaneous preterm birth <32 weeks, a cervical length <25 mm will only detect 19% of women who will go on to delivery at <35 weeks but of those who have a cervical length <25 mm, 75% of them will go on to deliver preterm (Table 2).

Table 2 Test characteristics of cervical length to predict preterm birth <35 weeks

*This study reported the risk of preterm birth <34 weeks

Of note is the number of women needed to screen to identify one woman with a short cervix. In cohorts reporting screening of a general, low-risk population (singletons, no prior preterm birth) the incidence of a cervical length ≤25 mm was 0.89–2.0% and of cervical length ≤20 mm was 0.47–1.2%.^{105, 106} The incidence is higher in those with a prior preterm birth, approximately 12.5%.⁸³

We are not aware of a randomized trial of screening for cervical insufficiency (i.e. women randomized to routine transvaginal cervical length measurement versus no screening that offers interventions and then measures rates of preterm birth between those randomized to routine screening versus no screening). In a retrospective cohort study examining the incidence of preterm birth before and after instituting a universal cervical length screening program, a universal cervical length screening program was associated with a decreased frequency of preterm birth prior to 37 weeks.¹⁰⁶ In another retrospective study at a clinical center where universal screening was offered, women who declined screening were compared to those who accepted screening. Those who declined screening had a higher incidence of early preterm birth.¹⁰⁵ Cost-effectiveness analysis suggests that universal screening of singleton fetuses is cost-effective when the treatment used for short cervix is vaginal progesterone.¹⁰⁷

The low incidence of a short cervix and the controversy over appropriate treatment methods have led to debate over who should be screened (Table 3). The American College of Obstetricians and Gynecologists does not mandate universal screening; however, should screening be implemented, they recommend a protocol validated in a clinical trial to prevent preterm birth.⁹⁸

Table 3 Summary of screening for cervical insufficiency by population

The mainstay of cervical insufficiency treatment is the cerclage, which may be placed via vaginal or abdominal (cervico-isthmic) routes. History- and ultrasound-indicated cerclage are not an appropriate treatment for suspected cervical insufficiency in nulliparous women or in those with multifetal gestation. Alternative treatments for cervical insufficiency include vaginal progesterone and pessary.

Indications for cerclage

History indicated                       

A history-indicated cerclage may be performed in singleton pregnancies in women with either (1) a history of mid-trimester loss characterized by painless cervical dilation without labor or abruption or (2) a history of a cerclage placement in a prior pregnancy for painless cervical dilation. Initially offered to women only with recurrent mid-trimester losses, a cerclage is now frequently offered after a single mid-trimester loss.¹⁰⁸ A history-indicated cerclage is typically performed at 12–15 weeks of gestation after most spontaneous abortions have occurred so as to avoid complications of early spontaneous loss. It also permits a thorough sonographic evaluation of the fetus to rule out many severe or life-threatening anomalies. In selected cases, prenatal diagnosis using chorionic villus sampling can also be performed prior to the surgery.

Ultrasound indicated

An ultrasound-indicated cerclage may be placed in a woman with a prior preterm birth with a cervical length ≤25 mm. This type of cerclage is typically placed at 16–22 weeks, the typical time period for cervical length screening. An ultrasound-indicated cerclage is not indicated in women with a short cervix but no prior preterm delivery.⁹⁸ An ultrasound-indicated cerclage may be more difficult to perform than a history-indicated cerclage due to cervical shortening.

Physical exam indicated

A physical exam-indicated cerclage may be placed when a woman presents with painless cervical dilation up to 23−24 weeks' gestation. Cerclage may be placed up to 4-cm dilation and may be considered with more advanced dilation; techniques to enhance visualization of the cervix in the presence of bulging membranes include amnioreduction, Trendelenburg position, use of a Foley catheter to replace the membranes into the lower uterine segment, and instillation of normal saline (up to 1 L) into the urinary bladder to displace the lower uterine segment. Prior to placement of an exam-indicated cerclage, the patient should be evaluated for evidence of active labor (manifested by contractions and progressive cervical change), intrauterine infection (fever, fundal tenderness), and abruption (vaginal bleeding).

Cervicoisthmic (abdominal) cerclage

Rarely a patient presents who is a viable candidate for a cervicoisthmic cerclage. This procedure is reserved for women who have failed a transvaginal cerclage; however, it is conceivable that some women failed transvaginal cerclage (with or without subsequent cervical damage) because they did not have cervical incompetence as the cause for their prior reproductive adversity. Because of the associated morbidity, a patient should be selected for cervicoisthmic cerclage only after a careful review of her history and a thorough clinical assessment indicates that the prior (failed) transvaginal cerclage was clinically indicated and that placement of another transvaginal cerclage is technically not feasible because of the magnitude of the anatomic deficiency. Primary cervicoisthmic cerclage might be considered in a patient with a typical history of incompetence and no prior transvaginal cerclage if the cervix is surgically or congenitally absent.

Contraindications to cerclage

A cerclage is not recommended in pregnancies complicated by known lethal fetal anomalies or stillbirth as no fetal benefit is expected to be gained by extending the length gestation in these pregnancies. Cerclage placement in pregnancies complicated by twins may increase the risk of preterm birth, even in the presence of a short cervix. Therefore, cerclage is not recommended in twin pregnancies.^{108, 109, 110} Cerclage may be harmful in the setting of intramniotic infection, active labor, preterm ruptured membranes, or active bleeding and therefore is not recommended.

Efficacy of cerclage placement

History indicated

The efficacy of history-indicated cerclage is difficult to evaluate due to the above described difficulties in diagnosing cervical insufficiency. Three randomized trials of singleton gestations have evaluated the efficacy of cerclage in women with a possible history cervical insufficiency.^{111, 112, 113} The largest trial (n = 1292) was conducted by the Royal College of Obstetricians and Gynaecologists from 1981 to 1988 in 12 countries.¹¹³ The patient population was heterogeneous with at least six risk-factor subgroups. The risk of preterm birth prior to 33 weeks was significantly lower in those that were assigned to cerclage compared to expectant management (13% versus 17%, p = 0.03), although women in the cerclage group also received more tocolytics, were hospitalized more frequently, and suffered from puerperal fever more often. A subgroup analysis of this study suggests that cerclage was only beneficial when placed after three or more spontaneous preterm births. However, most obstetricians are loath to wait for three losses and intervene after only one, despite a lack of evidence suggesting benefit.¹¹⁴

A meta-analysis of randomized trials in women with a prior spontaneous preterm birth assigned to either history-indicated cerclage or cervical length screening with cerclage if the cervix shortens determined that history-indicated cerclage did not perform better than cervical length screening.¹¹⁵ Additionally, the majority of women who underwent cervical length screening did not develop a short cervix. Therefore, the authors concluded that women with a history of prior preterm birth can be monitored with cervical lengths rather than undergo history-indicated cerclage.

In summary, history-indicated cerclage may not benefit most women. The preferred candidate for a history-indicated cerclage is a patient who has had recurrent mid-trimester loss consistent with cervical insufficiency. Those who had an early preterm delivery after cerclage for cervical shortening or in those whom a cerclage could not be placed after cervical shortening may also be candidates for a history-indicated cerclage.

Ultrasound-indicated, prior preterm birth

In women with a prior preterm birth, cerclage for a cervical length <25 mm is efficacious. Owen and colleagues performed the largest US trial, in which 1014 women with a prior spontaneous preterm birth between 16 and 34 weeks underwent serial cervical screening.⁵ Of the women screened, 318 women developed a cervical length <25 mm and 302 were randomized: 149 to cerclage and 153 to no cerclage. Women randomized to cerclage had a decreased risk of preterm birth prior to 37 weeks (45% versus 60%, p = 0.01), previable birth (<24 weeks, 6.1% versus 14%, p = 0.03), and perinatal death (8.8% versus 16%, p = 0.046).

Berghella et al. performed a meta-analysis of cerclage for short cervical length. Only trials that randomized women with a singleton gestation, prior spontaneous preterm birth, and short cervix to either cerclage or no cerclage were evaluated. Although several trials included other women (i.e. no prior preterm birth or twins), only data for women meeting the above inclusion criteria were evaluated. Four of the studies only included women with a preterm birth prior to 32–34 weeks; only one study included women with a prior preterm birth up to 36 weeks. The meta-analysis demonstrated a lower risk of preterm birth at any gestational age (<37, <32, <28, <24 weeks) in women receiving a cerclage with a cervical length <25 mm. Cerclage in this population was also associated with decreased perinatal mortality (pooled risk ratio 0.64, 95% CI 0.45–0.91). The authors calculated that 20 cerclage procedures are needed in this patient population to prevent one perinatal death.

Intramuscular 17-alpha-hydroxyprogesterone (17-OHP) also decreases the risk of preterm birth in women with a prior preterm birth.¹¹⁶ There does not appear to be an additive benefit to 17-OHP and cerclage,^{117, 118} nor is it known which intervention is superior for women with prior preterm birth. As neither intervention is 100% effective and physicians are also currently unable to predict who will respond to cerclage or 17-OHP, many women with a history of prior preterm birth will receive both interventions. Some physicians reserve cerclage for those with a prior preterm birth <34 weeks, as the majority of randomized trials were limited to this patient population.

Ultrasound-indicated, no prior preterm birth

Placement of a cerclage is not recommended in women with a short cervix and no history of preterm birth.⁹⁸ In an individual patient meta-analysis of women with asymptomatic cervical shortening, cerclage was not beneficial in reducing the risk of preterm birth prior to 35 weeks in the subgroup of women with no prior preterm birth (RR 0.84, 95% CI 0.60–1.17).¹⁰⁹

Physical-exam indicated

Women who present with acute cervical insufficiency, typically defined as a mid-trimester cervical dilation of at least 2 cm and no other predisposing cause (labor, infection, bleeding, rupture of membranes) are often considered for a physical-exam indicated cerclage. Data on the efficacy of this procedure are limited, and typically are reported as case series or cohorts.

Aarts and colleagues reviewed eight series published between 1980 and 1992 comprising 249 patients who received a physical-exam indicated mid-trimester cerclage and estimated a mean neonatal survival rate of 64% (reported range 22–100%).¹¹⁹ Novy and colleagues published a series of 35 cases of acute cervical insufficiency (cervical dilation 2–5 cm); 19 women received cerclage, and 16 were managed with bed rest.¹²⁰ Neonatal survival was 80% in the cerclage cohort versus 75% in the bed rest group. In another report using a similar study design, Olatunbosun and colleagues documented a longer mean gestational age at delivery compared to bed rest (p=0.001), although neonatal survival was not significantly different (p = 0.3).¹²¹ The Global Network for Perinatal and Reproductive Health reported a cohort of women who received either cerclage or expectant management for acute cervical insufficiency presenting between 14^0/7 and 25^6/7 weeks.¹²² Women who received a cerclage tended to present earlier than those who were expectantly managed (19.1 versus 23.1 weeks); despite this difference, women receiving a cerclage were more likely to deliver at a later gestational age (33 versus 25.9 weeks, p = 0.003) and have a longer time from presentation to delivery (12.4 versus 1.6 weeks, p <0.001).

A small randomized clinical trial evaluated emergency (i.e. physical exam-indicated) cerclage plus bed rest versus bed rest alone in 23 women who presented with cervical dilation and membranes prolapsing to or beyond the external os prior to 27 weeks of gestation.¹²³ Both singleton and twin gestations were eligible; however, no information on the amount of cervical dilation was reported, and so it is not known whether the groups were comparable in this important aspect. They observed a longer mean interval from presentation to delivery in the cerclage group (54 days versus 20 days, p = 0.046). The neonatal survival rate was 9 of 16 with cerclage and 4 of 14 in the bed-rest group. Although the survival differences were not statistically significant, there were significantly lower rates of neonatal composite morbidity (which included death) in the cerclage group (10 of 16 versus 14 of 14, p = 0.02). Their sample size was not large enough to assess cerclage efficacy in singletons versus twins, but in this rare circumstance, it may be reasonable to consider a physical exam-indicated cerclage for women with twins who present with acute cervical insufficiency.

Mays and colleagues performed amniocentesis in 18 women who presented with this syndrome and analyzed the amniotic fluid for glucose, LDH, Gram stain, and culture.¹²⁴ An additional seven women were managed with cerclage after refusing amniocentesis. Clinical decisions were made based on amniotic fluid biochemistries, and low glucose and elevated LDH, which were immediately available, were sufficient for a diagnosis of subclinical infection. Of 11 women who underwent cerclage with no evidence of subclinical infection, the neonatal survival was 100%, and the mean latent phase duration from presentation to delivery was 93 days. Of the seven women with abnormal biochemistries in whom cerclage was withheld, no neonatal survivors were observed, and the mean latent phase was 4 days. Recognizing that at least a portion of the seven women who declined amniocentesis, but who received emergent cerclage, also had subclinical infection, it was predictable that the mean latent phase in this cohort was intermediate (17 days) as compared to the groups with amniotic fluid analyses. These investigators suggested that amniocentesis could aid in selecting candidates for emergent therapeutic cerclage. Although these findings were provocative, the uncontrolled study design did not substantiate that cerclage is effective in these cases, but it demonstrated that acute cervical insufficiency is frequently associated with intrauterine infection.

Alternatives to cerclage

As not all patients are candidates for cerclage, alternative, non-surgical approaches have been sought.

Progesterone

All women with a prior spontaneous preterm birth between 16 and 36^6/7 weeks should have recommended weekly intramuscular 17-alpha-hydroxyprogesterone 250 mg from 16 to 36 weeks of gestation.⁹⁸ This intervention has a demonstrated 30% reduction in the risk of recurrent preterm birth.¹¹⁶

Progesterone may also be given vaginally in women with a short cervix. One large, multicenter randomized, placebo-controlled trial of 458 singletons with a short cervix on ultrasound (10–20 mm) demonstrated a decreased risk of preterm before 33 weeks (RR 0.55, 95% CI 0.33–0.92).¹²⁵ This study also demonstrated a decreased risk of neonatal morbidity and mortality (RR 0.57, 95% CI 0.33–0.99). A meta-analysis that included five randomized trials including women with a short cervix on ultrasound demonstrated that vaginal progesterone for these women decreased the risk of preterm birth prior to 34 weeks or perinatal death (RR 0.66, 95% CI 0.52-–0.83); the number needed to treat to prevent one case of preterm birth prior to 34 weeks or perinatal death was 11.¹²⁶ In women with a short cervix (≤20-mm) who are not a candidate for cerclage (i.e. no prior spontaneous preterm birth) or decline cerclage, vaginal progesterone is a treatment option.

Pessary           

Investigators in Europe and the United States have studied vaginal pessaries for the treatment of the incompetent cervix.¹²⁷ In 1961, Vitsky proposed the mechanism whereby a lever pessary (Smith, Hodge or Risser design) might be an effective treatment for the incompetent cervix.¹²⁸ A vaginal pessary would displace the cervix posteriorly and shift the gravitational effects of the expanding uterine contents off the internal os and onto the anterior lower uterine segment.

Data around pessary use are conflicting. A large randomized trial conducted in Spain for singletons with a cervical length <25 mm in the midtrimester compared Arabin pessary to expectant management. This study suggested a dramatic reduction in the risk of preterm birth prior to 34 weeks (27% versus 6%, p <0.001) and prior to 28 weeks (8% versus 2%).¹²⁹ However, another large multicenter, multinational trial found no benefit of pessary for preterm birth prevention <34 weeks (OR 1.12, 95% CI 0.75–1.69) or <28 weeks (OR1.71, 95% CI 0.89–3.29).¹³⁰

Pessary has also been investigated for use in twin pregnancies. A subgroup analysis of a randomized trial of pessary for preterm birth prevention in twins suggested potential benefit for those with a cervix <38 mm.¹³¹ This treatment should be studied further prior to universal implementation of this intervention; studies are ongoing in the US and in Europe (www.clinicaltrials.gov).

Technique

Cerclage procedures still in common use today include those originally described by Shirodkar,⁹ McDonald,¹⁰ and Benson,¹¹ although various modifications to each have gained wide acceptance.

McDonald cerclage

Although McDonald¹⁰ and Shirodkar⁹ originally reported their surgical techniques in temporal proximity, most clinicians now favor the McDonald method for several reasons. Because it does not require mucosal dissection and elevation to expose the cervix at the level of the internal os, it is easier to perform and is associated with less blood loss. The circumferential suture is tied on the ectocervix and easily accessible, which facilitates later cerclage removal. Although both methods have their proponents, there are no controlled comparative data on treatment efficacy. The findings of uncontrolled cohort series suggest that success rates (usually defined as perinatal survival) are similar and neither technique has an obvious advantage.^{132, 133}

To place a prophylactic McDonald cerclage, the anesthetized patient is placed in dorsal lithotomy position. At least one assistant is required to provide exposure using right angle or medium-sized Deaver retractors. While some surgeons prefer the use of a weighted speculum, we have found that individually placed retractors permit better exposure around the entire circumference, and especially for the posterior bites. After an antiseptic vaginal prep, the anterior ectocervix is grasped with a ring forceps, or similar non-traumatic instrument such as a Babcock clamp, used to provide counter traction. The urinary bladder is generally emptied prior to the procedure, although some surgeons recommend leaving some urine in the bladder to better define the position of the bladder as it reflects onto the ectocervix.

For right-handed surgeons the first tissue bite is taken at the 11–12:00 position on the cervix, exiting at around the 10:00 position. When placing the anterior stitch, the surgeon must avoid the bladder mucosa that can be identified by moving the cervix in and out, and noting where the vaginal mucosa folds in as it reflects off the ectocervix. A uterine sound bent in a 180° curve and introduced through the urethra can also be utilized to identify the most inferior aspect of the bladder reflection. Intraoperative cervical ultrasound may also be used to identify the bladder reflection.

The circumferential suture continues with successive bites placed as close to the vaginal fornix as technically feasible. The posterior stitches are the most technically demanding and require coordinated exposure by the assistant and appropriately directed counter traction by the surgeon. To facilitate suture placement, the forceps used for counter-traction should be replaced prior to each bite, just distal to and centrally located between the planned needle entry and exit sites. Jennings proposed a modification whereby each successive tissue bite reenters the cervix through the exit hole left by the preceding stitch.² This effectively buries most of the suture beneath the mucosa and has a theoretic advantage of minimizing the exposure of the submucosal stroma to vaginal pathogens.

Since the size of the cervix and needle can vary, there is no predetermined number of tissue bites required to completely encircle the cervix, but 4–6 are generally necessary; however, the exact number of bites is not important. The last bite should exit in close proximity to the original entry site.

Another variation of the original procedure uses two sutures placed several millimeters apart.¹³⁴ This has the theoretic advantage of spreading the suture tension over a larger area and may help prevent the more cephalic stitch from becoming displaced. A retrospective cohort comparing two stitches to one stitch did not find any benefit in preterm birth reduction with two stitches.¹³⁵ We utilize a second stitch if we believe that the first suture was not optimally placed at the bladder reflection anteriorly or as high as possible in the posterior vaginal fornix. It is necessary to record how many stitches were placed and where the knots were tied to facilitate their later removal.

Choice of suture and needle for the McDonald procedure is somewhat operator dependent. Most surgeons now utilize a permanent synthetic material such as #1 or #2 nylon, Prolene, or Mersilene. Mersilene 5mm tape has also been proposed, but, compared to suture, it is more difficult to pull the tape through the stroma and requires more tissue traction and manipulation. Monofilament suture has the advantage of offering less tissue resistance, which should facilitate both placement and removal. However, monofilament suture is prone to slippage and may be more difficult to tie down firmly. If the desired suture is not available swaged on the proper size needle, the suture can be threaded into a Mayo needle for the procedure. Since the descending branches of the uterine artery are found at 3:00 and 9:00, this area should be avoided when placing the stitches.

After the cerclage stitch has been placed, it is important to take up any slack introduced with the multiple tissue bites, and we utilize a “laundry bag” technique whereby traction is applied to each side of the exiting suture while holding counter traction at the exit site with two fingers of the opposite hand. Once this is accomplished the suture is tied down firmly but should not cause visible blanching of the surrounding tissue. In order to facilitate later identification and removal, long tag should be left above the knot. After placement, a digital examination will confirm a closed endocervical canal that is not overly constricted. However, it should not admit a gloved finger. Cervical ultrasound has also been used to assess suture placement, however, its role as a useful adjunct has not been determined.

Shirodkar cerclage

Shirodkar⁹ was the first to report a true cerclage operation, and with some modifications, this procedure is still utilized in contemporary obstetric practice. Although some surgeons empirically prefer this technique over the McDonald, the choice seems more dependent on training and the importance ascribed to its theoretic benefits, i.e. the ability to place mechanical support close to or at the level of the internal os; this goal is generally not attainable using the McDonald approach. To perform a Shirodkar cerclage, the patient’s positioning and prep are performed as previously described. With good exposure by assistants, the surgeon places an atraumatic clamp on the ectocervix, first on the posterior side, which is used to displace the cervix anteriorly, exposing the posterior aspect of the cervical mucosa. The mucosa is incised in a transverse direction near the fornix, allowing the posterior vaginal mucosa to be detached and elevated to the level of the internal os. The pouch of Douglas should not be entered in this step. Although a linear incision on the posterior side was originally described, a transverse incision facilitates the lateral exposure required to place the suture correctly in the stroma without entering the endocervical canal.

Once the surgeon is satisfied with the posterior incision and dissection, the anterior aspect of the cervix is exposed by moving the grasping tenaculum or forceps to the anterior cervix. The bladder reflection is identified similar to the initial step in a vaginal hysterectomy. To aid in the identification of the incision site and improve the plane of dissection, 5–10 ml of sterile saline can be injected into the submucosal tissue. A 1–2 cm transverse incision is made, and once the correct tissue plane is reached, the bladder is pushed off the cervix with blunt dissection. The dissection continues until the uterine corpus is felt ballooning outward from the cervix. Once this has been accomplished, a curved Allis clamp is used to simultaneously grasp the lateral edges of the anterior and posterior incisions. With downward traction on the ectocervix the lateral cervical stroma is exposed as far cephalad as possible. This creates an optimal path for the needle and offers sufficient tissue stability to direct its path through the stroma, avoiding the descending branches of the uterine arteries.

The needle and suture, generally 5 mm Mersilene tape, is passed from anterior to posterior on the patient’s right (assuming a right-handed operator). Some surgeons use an aneurysm needle threaded with the Mersilene tape. After the anterior-posterior bite is completed, a similar technique is used to stabilize the left side, where the needle is passed from posterior to anterior. The tape is then securely tied without excessively constricting the cervical tissue. The ends of the knot are left approximately 5 mm long, so that the mucosa can be closed around them using an absorbable suture, leaving the cut ends exposed to facilitate later identification and removal. The posterior incision is also closed over the tape that should complete hemostasis. Some surgeons have suggested reversing the insertion procedure to place the knot posteriorly. Theoretically this should reduce the risk of chronic bladder irritation.¹³⁶ However, the choice appears to be operator preference, as no comparative data have been published confirming the superiority of one versus the alternate method.

Cervicoisthmic cerclage

Cervicoisthmic cerclage (or abdominal cerclage), traditionally placed at laparotomy, presents the most surgically challenging and morbid cerclage procedure. The decision to place this cerclage means that a patient must generally undergo two abdominal operations, because a safe vaginal birth is impossible, and cesarean is required. Originally described by Benson and Durfee,¹¹ a circumferential ligature of 5 mm Mersilene tape is placed near the insertion of the uterosacral and cardinal ligaments at the bifurcation of the uterine artery, between its ascending and descending branches. In order to gain exposure and better visualize the vascular supply, the bladder is dissected off the lower uterine segment as is done during hysterectomy, and the incision is extended laterally on both sides to include a portion of the anterior peritoneal reflection of the broad ligament. While an assistant provides fundal traction, the surgeon grasps the uterine vessels and with gentle lateral traction exposes a triangular-shaped avascular space between the bifurcation of the artery and the uterine isthmus. Once identified, a right angle clamp is passed from the anterior to posterior, avoiding the vessels and ultimately tenting the posterior leaf of the broad ligament. The tissue overlying the tips of the clamp is incised, exposing the tips, by which a 15 cm length of Mersilene tape is brought back through the newly formed space. An identical procedure is performed on the contralateral side and the tape is then tied and secured anteriorly. The cut ends are sutured down flush with the encircling tape with non-absorbable suture. The bladder reflection is then reapproximated.

Novy¹³⁷ reported a modification of the original technique whereby a blunt-tip needle and swaged on Mersilene tape is passed through the avascular space from anterior to posterior on one side and from posterior to anterior on the other, thus, obviating the need for any blunt dissection. A similar modification was also reported by Topping.¹³⁸ Mahran¹³⁹ reported a different modification: after the bladder is taken down, a Mayo needle threaded with 6 mm nylon tape is passed from anterior to posterior within the cervical isthmic tissue, but near its lateral edge, at the level of the aforementioned arterial bifurcation. The opposite end is also passed anterior to posterior in an analogous fashion on the contralateral side, and the ends are tied posteriorly. On each side the needle passes through the tissue of the uterosacral ligaments. Placing the knot on the posterior aspect may make the cerclage easier to remove. None of these reported modifications appears to offer appreciable advantage over the originally described technique, and so the choice is operator dependent.

The cervicoisthmic cerclage is generally placed between 10 and 16 weeks’ gestation, although it may be placed prior to pregnancy. Placement at the earlier range is technically easier, as the expanding uterine volume impairs exposure and increases the risk of bleeding complications. Coordination among the surgeon and assistants is crucial to maintaining the optimal surgical exposure. As with the transvaginal cerclage, individual hand-held retractors provides superior exposure compared to a self-retaining retractor for the critical steps.

Recently, attention has turned to minimizing the morbidity of cervicoisthmic cerclage by placement either laparoscopically or robotically. Case reports, case series, and prospective cohorts of laparoscopic placement and removal of a cervicoisthmic cerclage have been published.^{140, 141, 142, 143, 144, 145, 146} In a systematic review of laparoscopic cervicoisthmic cerclage placement (n = 135) compared to those placed by laparotomy (n = 1116), the overall fetal survival rate was 84.1% with the laparoscopic approach and 90.8% with the laparotomy approach.¹⁴⁷ Complications rates did not appear to be significantly different.

It is unclear whether cervicoisthmic cerclage must be removed after childbearing is completed. The most opportune time to remove the cerclage is at the cesarean birth, generally performed at 36–37 weeks’ gestation. However, some women will desire more children, and, reasonably, would choose to maintain the cerclage, even though they might never conceive again. If conception does recur with the cerclage in place, there is still a risk of spontaneous abortion. In general a first trimester miscarriage can be managed in the usual manner, since the cerclage should not interfere with the passage of a small 8–10 mm curette.

Complications of cerclage

The most commonly reported complications associated with cerclage are membrane rupture and intrauterine infection. Bleeding may occur, but serious hemorrhage is generally limited to cervicoisthmic procedures. Harger¹⁴⁸ tabulated cerclage-associated complications reported in the past 40 years. Chorioamnionitis complicated 0.8–8% of elective cerclage procedures and 9–37% of “urgent” or emergent procedures. Membrane rupture attributed to elective cerclage was observed in 1–18% and was associated with up to 65% of emergent cases, although the reported range with emergent cerclage is extremely wide. Whether cerclage alone can precipitate overt preterm labor seems doubtful. However, the foreign body might lower the threshold for uterine activity because of local inflammatory effects. Uterine activity often occurs in proximity to cerclage placement and women who have undergone cerclage are more likely to receive tocolytic agents during their gestation.¹¹³

Since preterm labor, chorioamnionitis and PROM can complicate any gestation, and have all been associated with preterm birth, the attributable risk of these events from cerclage is difficult to ascertain, particularly in the case of acute cervical insufficiency where preexisting (subclinical) infection is commonly seen. PPROM and preterm labor may similarly complicate those pregnancies managed expectantly without cerclage. Nevertheless, iatrogenic membrane rupture does occur during emergent cerclage placement and probably shortens the gestation as a result.

In the largest randomized trial of prophylactic cerclage yet published the investigators noted a (possibly or probably attributable) adverse event rate of 15% of women who received cerclage.¹¹³ Rust and colleagues reported a 16% incidence of postoperative chorioamnionitis in a randomized trial of cerclage for cervical sonographic indications.¹⁴⁹ Owen and colleagues reported only a 1.3% incidence of surgical complications in a randomized trial of cerclage in women with a transvaginal cervical length <25 mm and a prior preterm birth (1 PROM, 1 hemorrhage in 153 cerclage placement).⁵

Complications may also occur at the time of delivery. Cervical lacerations occur in 1–13% of women with a cerclage.¹³² Scarring of the cervix may lead to an inability of the cervix to dilate and lead to cesarean delivery.¹³²

Some cerclage complications are occult and their effects not realized in the index pregnancy. Whether a cerclage is placed for a carefully scrutinized history or for questionable indications, the observation that it may have been associated with a desirable outcome is generally used to justify future cerclage therapy. Cerclage may lead to an increased cesarean rate, if removal is deemed to be difficult or the decision was made to leave it for successive gestations.

Perioperative management

Preoperative considerations

Many clinicians recommend obtaining cervicovaginal cultures for common pathogens and treating positive cultures prior to placing a cerclage. Active cervicitis should be considered a contraindication to prophylactic cerclage placement, and this must be successfully treated before surgery. Other contraindications to cerclage include ruptured membranes, lethal fetal anomaly, suspected chorioamnionitis, vaginal bleeding and labor.

Antibiotics

Antibiotics are not routinely indicated in history-indicated or ultrasound-indicated cerclage.

For women undergoing physical-exam indicated cerclage, some physicians consider giving antibiotics given the high incidence of subclinical intra-amniotic infection.¹⁵⁰ We identified one randomized trial of antibiotics at the time of physical-exam indicated cerclage.¹⁵¹ This study randomized women to indomethacin (50 mg at 0, 8, and 16 h postoperatively) and cefazolin (weight-based dosing, at 0, 8, and 16 h postoperatively). A greater proportion of the indomethacin/cefazolin group had their pregnancies prolonged by at least 28 days (p = 0.01), although gestational age was similar between groups.

Anesthesia

Regional anesthesia is typically sufficient for placement of cerclage, although general anesthetic may also be used. General anesthesia is typically associated with higher demand for opioid and non-opioid pain medication post-operatively.¹⁵²

Postoperative management

Patient activity

Pelvic rest and sexual abstinence for the remainder of gestation are widely prescribed, but resumption of intercourse may be reasonable in selected cases.¹⁵³ Breast stimulation is known to cause uterine contractions and should also be avoided. Since the use of bed rest in pregnancy as an effective therapy has been seriously questioned,¹⁵⁴ it seems reasonable to individualize this recommendation based on a patient’s symptoms and physical findings.

Indications for removal

Gestational age  In the absence of complications such as spontaneous labor or rupture of membranes, the cervical cerclage should be removed at 36–37 weeks. As the likelihood of going into spontaneous labor increases beyond 36 weeks, cerclage removal at this gestational age is recommended to avoid cervical lacerations. Cerclage removal typically can be performed in the office, although rarely patients do not tolerate the exam and regional anesthesia may be required; removal of the Shirodkar stitch is much more challenging than the McDonald. Expectant management of a cervical cerclage is not recommended beyond 37 weeks. After cerclage removal, the patient may await spontaneous onset of labor. Cerclage removal alone is not an indication for delivery.

Labor  A cervical cerclage should be removed in the setting of active labor or suspected active labor. Because the cerclage may prevent initial cervical dilation, labor should be suspected in the setting of painful, regular contractions or vaginal bleeding. On physical exam, the cervix may feel effaced and even dilated, the stitch may feel as though it is on tension, or the lower uterine segment may feel as though it is bulging. While many physicians and patients are hesitant to remove a cerclage prematurely out of a desire to prolong the pregnancy as long as possible, the harm of leaving the cerclage in place and risking significant cervical lacerations must also be considered. In the setting of ongoing labor, the risk of leaving the cerclage in place likely outweighs the benefits and the cerclage should be removed.

Rupture of membranes  The cerclage should be removed and the patient delivered when the membranes rupture at 34 weeks or beyond. There is significant debate regarding the timing of cerclage removal after preterm, premature rupture of membranes (PPROM). Some argue that the cerclage should remain in place in order to prevent cervical dilation while others argue that the cerclage should be removed as it is a potential nidus of infection and chorioamnionitis. A randomized trial of 56 women with PPROM and a cervical cerclage in place randomized women to either immediate removal of cerclage or retention.¹⁵⁵ In this study, retention of the suture was not associated with increased risk of infection nor was it associated with prolongation of the pregnancy. Due to the unpredictable nature or PPROM with the possibility of onset of labor, we typically remove the cerclage at the time of PPROM. At early gestational ages (<30 weeks), we may leave the cerclage in place for 48 hours of betamethasone administration and then remove the stitch.

Abdominal cerclage  Women with an abdominal cerclage should be delivered between 36 and 37 weeks to avoid the risks of labor. It is unclear whether cervicoisthmic cerclage must be removed after childbearing is completed. The most opportune time to remove the cerclage is at the cesarean birth, generally performed at 36–37 weeks’ gestation. However, some women will desire more children, and, reasonably, would choose to maintain the cerclage, even though they might never conceive again.

Contemporary lines of evidence indicate that cervical insufficiency is rarely a distinct and well-defined clinical entity, but only one component of a larger and more complex syndrome of spontaneous preterm birth. The original paradigm of obstetric and gynecologic trauma as a common antecedent of cervical insufficiency, has been replaced by the recognition of functional, as opposed to anatomic deficits as the more prevalent etiology. Cervical competence functions along a continuum, influenced by both endogenous and exogenous factors which interact through various pathways with other recognized components of the preterm birth syndrome: uterine contractions and decidual/membrane activation. Thus, the convenient term, cervical insufficiency, may actually represent an oversimplified, anachronistic and incomplete version of the broader pathophysiologic process. Consequently, the empiric use of traditional therapies, unsubstantiated by results of clinical trials, must be questioned. Effective, evidenced-based management guidelines will be based on a more complete understanding of the preterm birth syndrome. This will improve patient selection and permit specifically tailored treatment regimens to be devised and then confirmed by the results of well-designed intervention trials. It should be recognized that even effective therapies will appear ineffective when applied to an unselected or poorly selected patient population.

This chapter has focused on the cervix and its contribution to the problem of prematurity. The authors are confident that continued research into the causes and prevention of spontaneous preterm birth will slowly, but ultimately identify strategies to effectively address the most important and persistent problem in obstetrics. Meanwhile, our incomplete understanding of this syndrome and the evolving controversies over its optimal management will continue to frustrate both patients and clinicians. Nevertheless, we should remain mindful of the importance of evidenced-based solutions and not let enthusiasm for unproven therapies impede our thoughtful scientific investigations.