The cervix as a biomechanical structure

Authors

  • M. House,

    Corresponding author
    1. Division of Maternal-Fetal Medicine, Tufts-New England Medical Center, Box 360, 750 Washington Street, Boston, Massachusetts, 02111, USA
    • Division of Maternal Fetal Medicine, Tufts-New England Medical Center, Box 360, 750 Washington Street, Boston, MA 02111, USA
    Search for more papers by this author
  • S. Socrate

    1. the Department of Mechanical Engineering, Massachusetts Institute of Technology, Boston, Massachusetts, USA
    Search for more papers by this author

Introduction

original image

Cervical insufficiency is a complex puzzle. Multiple etiologies lead to a final common pathway: undesired, painless cervical dilation. Most often, clinicians encounter the patient late in the pathologic process. In a patient with clinical features of cervical insufficiency, it is difficult to know if a weak cervix causes painless dilation or if dilation is caused by a different etiology1. A major obstacle is that the underlying assumptions of cervical insufficiency present a significant investigational challenge. For example, cerclage placement assumes that (a) the cervix is structurally weak and (b) a cerclage provides structural support. Currently, however, the assessment of cervical strength is based primarily on clinical history. An objective definition based on biomechanical evaluation is lacking. To further complicate the problem, there is no way to measure if a cerclage provides structural support.

Our engineering colleagues describe structural function in terms of the balance between two factors: structural strength and external loading. These concepts have straightforward application to the cervix. Cervical strength refers to the ability of the cervix to resist deformation, which is clinically referred to as ‘change’ (funneling, effacement or dilation). Loading refers to the forces acting to cause deformation. By definition, it is the balance between strength and loading that determines to what degree the cervix deforms. The Bishop score2 is an example of the utility of a biomechanical approach. The Bishop score is clinically useful because it incorporates both strength (consistency) and loading (dilation, effacement and station), thereby accounting for the underlying structural status of the cervix. In this editorial, we advocate a biomechanical approach to cervical function by discussing (1) the rationale for investigating the cervix as a biomechanical structure; (2) the structural features of the cervix; and (3) the clinical correlates to a structural paradigm of cervical function.

Rationale

Consider the following clinical hypothesis: a weak cervix causes spontaneous preterm birth. Although simple and intuitive, this hypothesis is currently impossible to investigate. Testing the hypothesis requires clinically acceptable, quantitative measurements of cervical strength and cervical loading. Under this hypothesis, a ‘strong’ cervix would easily resist physiologic loading. Conversely, physiologic loading would cause a ‘weak’ cervix to deform significantly. Clinically significant deformation would depend on the actual degree of deformation and the gestational age at which it occurs. A clinical study to test the hypothesis would predict that a strong cervix is associated with term delivery and a weak cervix with spontaneous preterm birth. If the hypothesis were proven, it would provide a rationale for interventions that strengthen the cervix or reduce the load, or both. If the hypothesis were disproved, other etiologies of preterm birth would gain favor.

Cervical structural features

Strength

Qualitative estimates of cervical strength are common in the clinical environment, but a quantitative definition is lacking and, without such a definition, objective assessment is not possible. Subjective terms, such as ‘soft,’ ‘ripe,’ ‘mature,’ or ‘favorable’ are often applied to a weak cervix and ‘firm’ or ‘unfavorable’ to its strong counterpart. Cervical strength depends on several factors, including its anatomic features (length of cervix, width of mucosa, width of stroma) and tissue properties. For example, consider the evolution of cervical tissue strength over time. An individual patient has a baseline tissue strength. As gestational age advances, her cervix becomes progressively weaker (‘softens’)3, 4. Close to labor, her cervix becomes weaker still (‘ripens’). It is known that an active inflammatory process has a role in the ‘ripening’ process5–7. What is not known is to what degree baseline strength, cervical softening and cervical ripening determine tissue properties at a particular moment in time.

One approach to quantifying the strength of the cervical stroma borrows from the engineering technique of characterizing the mechanical properties of materials through the use of constitutive models. A constitutive model is a mathematical model that relates the stress response in a material to its deformation history. In complex, multicomponent materials, the macroscopic response is controlled by the behavior of its constituent parts. During pregnancy, the cervical stroma is assumed to be the load bearing part of the cervix. Its constituent parts are predominantly extracellular matrix components (collagen, proteoglycans, elastin and water)7 and cellular components (fibroblasts, smooth muscle and vessels). The extracellular matrix predominates; the cellular components are not present in large quantities8 and are not thought to significantly contribute to the mechanical response9.

In the case of the cervix, only two notable constitutive models have been developed. The first model attempted to explain the reduction in stiffness of cervical tissue over the course of pregnancy using the theory of fiber-reinforced composites10, 11. It was based on the assumption that the tissue can be considered mechanically equivalent to a fibrous network (corresponding to the collagen fibers), embedded in a ground substance (mainly composed of hydrated glycosaminoglycans). The model achieved its objective in a semi-quantitative way by showing how decreased collagen concentration, decreased collagen alignment, changes in collagen–ground substance reinforcement and increased hydration combined to reduce tissue stiffness by a factor of 12 during pregnancy. However its underlying assumptions of linearity and small deformations make it restricted to simple estimates of the material properties.

More recently, a three-dimensional, constitutive model for the large-strain, time-dependent mechanical behavior of the cervical stroma in pregnancy was proposed by our group12. The model captured the global tissue response, which was controlled by the cooperative contributions of the major constituents. In terms of mechanical behavior, the dominant constituents were the collagen fibers and the hydrated network of glycosaminoglycans. The collagen network provided stiffness in tension. The glycosaminoglycans (proteoglycans and hyaluronic acid), owing to their high negative fixed charge density, drew water into the tissue and created a high osmotic pressure responsible for the compressive stiffness of the tissue. Free interstitial fluid flowed in the tissue according to pressure gradients within the porous network, and was a major contributor to the short-term properties of the tissue.

A constitutive model is evaluated on its ability to predict mechanical behavior under different modes of deformation (e.g. tension, compression). If the model predicts different types of mechanical behavior in vitro, presumably it reflects physical reality in vivo. Progress toward a robust constitutive model of cervical stroma, that allows a more complete understanding of its physical properties, will be an important step in quantifying the factors contributing to cervical function and thus dysfunction.

Loading

Loading refers to the forces that act to cause cervical deformation, and occurs in two ways: passive and active. Passive loading, the forces acting to cause cervical deformation in the absence of uterine contractions, arises when the expanding amniotic cavity comes into contact with the cervix, usually in the early second trimester. Contributing loading components include tensile stresses along the uterine wall associated with the growth of the baby and the amniotic membrane, hydrostatic pressure and intrauterine pressure. Superimposed on the passive load is active loading due to uterine contractions. The cervix must resist the total load (active + passive) in order to remain closed.

Significant challenges are inherent in estimating cervical loading. The cervix rests on the posterior wall of the vagina. It is attached circumferentially by the endopelvic fascia and lower uterine segment. The cervix is loaded superiorly by the decidua, which also has an adhesive function as it mediates mechanical contact between the fetal membranes and cervix13. In fact, loss of adhesion is thought to be related to the elevated fetal fibronectin in patients at risk for preterm delivery14. In summary, loading during human pregnancy is unique to the geometry and pelvic forces of the individual patient. Although progress in the study of cervical strength may be aided by animal models15, 16, it is difficult to envision an adequate animal model of human cervical loading.

The underlying assumption of cervical insufficiency is that the cervix is unable to resist a normal passive load. Yet this assumption begs the question: how much is a normal passive load? How is the load distributed? Although the cervix is passively loaded for the vast majority of pregnancy, most clinical studies focus exclusively on active loading. For example, the cause of preterm delivery in multiple gestation or polyhydramnios is suspected to be increased uterine distension leading to increased uterine contractility17. Yet increased uterine distension logically causes increased passive loading as well. Clinical studies of the total (passive + active) load may provide more insight than focusing on active loading alone.

Cervical deformation

The clinical case of cervical ‘change’, whether it is described as funneling, effacement or dilation, is a special case of a more general problem, namely that of solid body deformation. A more precise term for cervical ‘change’ is deformation. Cervical funneling, effacement and dilation are better described as clinically significant deformation of the cervical stroma, a description that incorporates both clinical significance and solid mechanics. Solid mechanics terminology is a highly technical field18. Investigation of clinically significant deformation requires a partnership between engineers and clinicians—the former because of their understanding of how to investigate soft material deformation and the latter because of their understanding of deformation that is clinically important.

Clinical correlates to a structural paradigm of cervical function

Cervical length

A cervix that shortens, under the loading associated with pregnancy, may be considered analogous to a structural member that is starting to fail. Just as a beam will bend before it breaks, a cervix will funnel and efface before it dilates. Just as the deformation of a beam can be predicted given well defined loading conditions and material properties, the pattern of deformation in the cervix can be predicted with realistic anatomic geometry, physiologic loading and material properties derived from biomechanical testing. For example, sonographic studies of the natural history of cervical deformation associated with insufficiency revealed that the cervix deforms in a ‘TYVU’ pattern19: a closed cervix corresponds to the ‘T,’ progressive funneling corresponds to ‘Y’ and ‘V,’ and end-stage funneling corresponds to ‘U.’ Using the methodology of structural mechanics, our group demonstrated that the mechanical stress distribution within the cervical stroma causes the cervix to deform in a TYVU pattern20, 21. Figure 1 shows an idealized, three-dimensional, finite element model of the pelvic region during pregnancy. Anatomic geometry was based, in part, on pelvic MRI22. TYVU deformation was demonstrated by varying the material properties of the constitutive model and applying representative pelvic loads (Figure 2). The key remaining question is whether TYVU deformation is caused by weak cervical structure or whether TYVU deformation reflects a different pathologic process, unrelated to cervical weakness.

Figure 1.

A three-dimensional, finite element model of the abdomen and pelvis at 20 weeks' gestation. The following anatomic structures are shown: uterine cavity (yellow), uterus and cervix (red), viscera of the abdomen and pelvis (gray) and limits of the abdominal cavity (green). The endopelvic fascia is hidden.

Figure 2.

The model in Figure 1 was loaded using finite element techniques. The uterine cavity is shown in yellow, the uterus and cervix in red and the endopelvic fascia in orange. For the sake of simplicity, the abdominal contents are not shown and the cervix is magnified. The material strength of the cervix was decreased by varying the material parameters of the constitutive model. The simulation demonstrates ‘TYVU’ deformation with membrane prolapse (a to d).

Cerclage

Two assumptions underlie cerclage placement: the cervix is weak and cerclage provides structural support. When a rescue cerclage is placed in the case of bulging, visible membranes in the midtrimester and the patient delivers in the third trimester, it seems clear that the two assumptions are valid23. However these cases are not the norm. Many prophylactic cerclages are placed in the early second trimester, based on clinical suspicion of a weak cervix, before any cervical deformation is noted. Thus it is possible that the cervix is not weak; no quantitative biomechanical assessment has addressed this assumption. In addition, the second assumption may be invalid: a cerclage may not provide mechanical support. The rationale for cerclage placement is to reinforce the tissue with a strong, load-bearing suture. However, in many cases, the cerclage is probably not load-bearing24. Most of the cervical load is concentrated at the internal os while many cerclages are placed in the middle of the cervix, which may not be close enough to support the load. Moreover, a cerclage is tightened with the expectation that compression will prevent cervical dilation and membrane prolapse. Yet, in the case of a prophylactic cerclage placed at 14 weeks' gestation, compressive forces probably do not persist. With increasing gestational age, the compressive forces decline as the tissue remodels and stress relaxation occurs. In summary, a cerclage may be ineffective because either it does not strengthen the cervix or it is placed in a strong cervix. These issues probably explain the marginal efficacy of cerclage in randomized trials25.

Uterus–cervix interaction

Although cervical dilation is often associated with uterine contractions, this is not always true. Pregnant women clearly experience contractions without associated cervical dilation; sometimes, they experience cervical dilation in the absence of contractions. It is intuitive that both the strength of the contractions and the weakness of the cervix play important roles. Yet a quantitative method to describe the way uterine contractions lead to cervical deformation does not exist. What is lacking is an understanding of the physical process of deformation as it relates to the interaction of the cervix and uterus. Objective investigation of the connection between the uterus and the cervix is a promising area of investigation26 and may help to explain the physical difference between harmless preterm contractions and preterm labor and delivery.

The role of ultrasonography

Transvaginal ultrasonography of the cervix has changed our conception of cervical function. Historically, the cervix has been described as either competent or incompetent. Ultrasound demonstrated that the cervical function is ‘continuous’ rather than binary27. Subsequent studies confirmed that the risk of preterm birth is progressively increased with a progressively shorter cervix28. Again, the question that remains unclear is whether a short cervix is caused by structural weakness.

Ultrasound may prove useful for assessing cervical strength. Quantitative sonographic measurement of material properties in vivo is an active area of investigation. Ultrasound elastography is one promising technique that uses sonographic information to determine the mechanical properties of soft tissue29, 30. Briefly, the transducer is used to compress the tissue and image processing techniques quantify pre- and postcompression data. The technique is based on the fact that soft tissue deforms differently from firm tissue, and the sonographic images reflect this difference. In biological tissues, which have changing and varied mechanical properties, tissue stiffness as a function of position can be measured. Clinical studies are ongoing in cardiology31, urology32 and gastroenterology33, and pilot studies have begun in obstetrics34.

Analogy with osteoporosis

The biomechanical paradigm of cervical function is most analogous to the paradigm of osteoporosis. Well-designed studies have proven that a weak hip causes hip fracture, because bone-strengthening medications prevent hip fracture. These studies were possible because an accurate estimate of bone strength (bone mineral density) was available. Measurement of bone strength allowed therapies to be targeted to the most vulnerable patients. To apply this strategy to the cervix, it is clear that two quantities need to be estimated: cervical strength and cervical loading. Cervical weakness is thought to be caused by collagen fibrous network disruption within the cervical stroma11, 35, 36. The central hypothesis of a biomechanical paradigm is the following: an intervention that inhibits collagen network disruption and/or decreases cervical loading will prevent spontaneous preterm birth. Although some progress has been made in estimating cervical strength4, 37, 38 there is no widely accepted method to measure it. Even less is known about cervical loading. Future progress in estimating strength and loading is critical to investigating whether a ‘weak’ cervix is clinically important.

Summary

Ours is a time of breathtaking advances in understanding molecular antecedents of medical disease. Yet the wizardry of modern molecular biology cannot directly address what obstetricians suspect to be important: cervical deformation. Deformation, or the lack thereof, is a question of solid mechanics. Investigation of cervical deformation requires a concerted, collaborative effort between engineers, biologists and clinicians. A multidisciplinary effort may allow a better understanding of the questions that both obstetricians and parents know to be important: why does preterm birth occur and what can be done to prevent it?

Acknowledgements

We thank Helen Feltovich for her editorial assistance.

Ancillary