From Molecular to Macro: The Key Role of the Apical Ligaments in Uterovaginal Support

Abstract

In order to explain the pathophysiology of pelvic organ prolapse, we must first understand the complexities of the normal support structures of the uterus and vagina. In this review, we focus on the apical ligaments, which include the cardinal and uterosacral ligaments. The aims of this review are the following: 1) to provide an overview of the anatomy and histology of the ligaments; 2) to summarize the imaging and biomechanical studies of the ligament properties and the way they relate to anterior and posterior vaginal wall prolapse; and 3) to synthesize these findings into a conceptual model for the progression of prolapse.

Introduction

The cardinal and uterosacral ligaments are the primary structures that provide apical support to the uterus and upper vagina (Figure 1). These suspensory structures have long been studied to better understand the development of pelvic organ prolapse. Loss of apical support is not only associated with prolapse of the apex, but it also has a close relationship with anterior vaginal prolapse and, to a lesser extent, posterior prolapse.^1–3 Our understanding of apical support has evolved over time with the advent of new technology that allows for more detailed study of the tissue composition and its biomechanical properties. Magnetic resonance imaging (MRI) studies in living women have clarified the importance of understanding apical support geometry and mechanics. In addition, there have been important insights into how in vivo measurements of the ligament properties greatly differ from in vitro testing. Moreover, MRI studies have allowed comparison between women with and without pelvic organ prolapse, revealing the precise details regarding the geometry of the ligaments and the way that geometry changes in cases of prolapse. Given the importance of the apical supports to pelvic organ prolapse, we conducted a literature review to summarize what is known about the nature of the apical ligaments and synthesize what is known to propose a theory for an overall disease model of prolapse.

Figure 1. Apical support.

This drawing shows the location of the cardinal and uterosacral ligaments in relation to the pelvic sidewall and vagina after removal of the bladder and uterine corpus. ©DeLancey

Anatomy, Histology, and Biochemistry

An understanding of apical support must begin with an anatomically and biomechanically accurate picture of the apical structures. A comprehensive review article on the anatomy and histology of the cardinal and uterosacral ligaments was published previously.⁴ We will summarize the main points here in Table 1 because they are necessary for providing the background for the other sections of this review. It is important to realize that, despite their names, the cardinal and uterosacral “ligaments” are visceral ligaments similar to mesenteries and are made up of completely different tissue than skeletal ligaments that connect bones. The apical ligaments consist of blood and lymphatic vessels, nerves, adipose tissue, and loose areolar connective tissue (Figure 2). Notably absent from these “ligaments” are any regular bands of dense connective tissue, comprised of type I collagen, that are seen in histologic sections of skeletal ligaments.⁵ As a reflection of this fact, the nomenclature of these structures has included “parametrium,” “matrix,” and various other terms. In this review, we will refer to the apical supports as “ligaments,” since this terminology is currently widespread, but we will keep in mind the fact that their histologic composition is completely different from that of skeletal ligaments.

Table 1.

Basic anatomy and histology of apical ligaments

Feature	Cardinal Ligamenta	Uterosacral Ligamenta
Origin	• Anterior trunk of internal iliac artery (one-third)	• Cervix alone (33%)
	• Upper border of greater sciatic foramen (two-thirds)	• Cervix and vagina (63%)
		• Vagina alone (4%)
Insertion	• Cervix and upper third of vagina	• Overlying sacrospinous ligament/coccygeus muscle complex (82%)
	• Some fibers go to the bladder (one-third)	• Sacrum (7%)
		• Piriformis muscle, sciatic foramen, or ischial spine (11%)
Histology	• Contains uterine vessels, adipose, and inferior hypogastric plexus that conveys autonomic nerves to pelvic organs	• Classical/superficial portion: visible edge beside the cul-de-sac composed of smooth muscle continuous with uterine musculature; appearance can be affected by position, pneumoperitoneum, and anesthesia
	• Nerves located in intermediate and distal sections	• Deep portion (i.e. rectal pillars/pararectal fascia): likely contains splanchnic nerves and surrounding connective tissue; visible on MRI; extends from sacrum to upper vagina; establishes support of posterior cul-de-sac (point D of POP-Q)

^aData in parentheses indicates proportion of women affected.

POP-Q: pelvic organ prolapse quantification system

Figure 2. Apical ligament histology.

A. Histology after trichrome staining of biopsy specimen of the deep uterosacral ligament showing mainly nerve fibers (n), adipose tissue (ad), and a few vessels (v). B. Histology of the cardinal ligament showing mainly vessels (v). Inset shows the histologic structure of dense connective tissue for contrast. Reproduced from Ramanah 2012.⁴

The histological composition of the apical ligaments has been compared between women with and without prolapse. Interestingly, the smooth muscle content of the uterosacral and cardinal ligaments does not differ between women with and without prolapse, but the collagen content does.⁶ There are two types of collagen fibers that are commonly studied in the context of prolapse—type I collagen provides the majority of tissue resistance to tension, while type III collagen provides flexibility and distention of tissue. When the uterosacral ligaments of women with prolapse are analyzed, they are found to have higher levels of type III collagen and lower levels of type I collagen compared to those of women without prolapse.^7,8 In addition to the quantity of collagen fiber types, the qualitative characteristics of collagen fibers appear to have a role in prolapse as well. When the cardinal ligaments of women with prolapse were examined with electron microscopy, the collagen fibers appeared thicker and more loosely packed, with increased distances between fibers, compared to women without prolapse.⁹

Furthermore, matrix metalloproteinases (MMPs), which are enzymes that participate in the degradation of the extracellular matrix, may have a role in prolapse. Increased expression of MMPs 1 and 9, which preferentially cleave collagen, has been identified in vaginal and uterosacral biopsies of women with prolapse.^10,11 These findings support the theory that increased collagen degradation in the uterosacral ligament could cause weakening of the tissue and eventually lead to prolapse. These tissue-level studies are limited because only a small portion of the ligaments were studied, and they include only those portions adjacent to the uterus due to the need to harvest tissue at surgery.

Relaxin and fibulin are additional proteins that have been studied in the uterosacral and cardinal ligaments. Compared to controls, women with prolapse appear to have uterosacral ligaments with higher levels of relaxin,¹² a protein that is known for its smooth muscle-relaxing and anti-fibrotic properties. Conversely, the protein fibulin-5 has decreased expression in the uterosacral ligaments of women with prolapse compared to those without. Fibulin-5 is a multifunctional extracellular matrix protein that regulates cell growth, motility, and adhesion. Presumably, diminished expression of this protein could lead to weakened connective tissue.¹³ While these findings are interesting, it is important to note that an association between a tissue-level factor and prolapse is not proof of causation.

These biochemical findings suggest that complex processes are at work in women with prolapse. At present, it is not known which changes—perhaps incited by genetics, trauma, pregnancy, or delivery—are a cause of prolapse and which are an effect of the abnormal forces placed on the ligaments because of the prolapse. Further work relating specific biomechanical and biochemical changes in the supportive tissues should help us to understand the role they play in the larger picture of pelvic organ support.

MRI Evaluation of Apical Ligament Geometry

MRI studies have revealed details regarding the orientation and length of the uterosacral and cardinal ligaments. Early cadaver studies on this topic were affected by the distortion of pelvic floor tissues due to loss of smooth muscle tone and alterations caused by high pressures during embalming. In contrast, MRI studies allow us to examine ligaments in living women. On MRI (Figure 3), the cardinal ligament appears as a web-like structure with an axis following the branches of the internal iliac vessels.⁶ The deep uterosacral ligament, however, has a more distinct band-like appearance.¹⁴ The superficial uterosacral ligament is not easily seen at present on standard MRI with 5mm slice thickness due to partial volume averaging.

Figure 3. MRI appearance of apical ligaments.

A. Note the location of the uterosacral ligament (red arrow) and cardinal ligament (black arrow) in this model generated from an axial MR scan. Inset: The bony pelvis, including the pubis (P), ilium (Il), ischium (Isch), and sacrum (S), is displayed. B. Positioning of the uterus (Ut), cardinal ligament (black arrow), and vagina (Vg) can be seen in this coronal image. C. The uterosacral ligament is marked by red arrows, and the cardinal ligament is marked by black arrows on the MRI used to make the model in A. D. The cardinal ligament is marked by arrows on the MRI that was used to make the model in B. Modified from Ramanah 2012.¹⁴

MRI studies have allowed for measurements of the axis and length of the ligaments at rest and strain in women both with and without prolapse. While many classic textbooks refer to the cardinal and uterosacral ligaments as “transverse” ligaments, MRI-based 3D reconstruction reveals a more vertical and dorsal orientation relative to the body axis. The cardinal and uterosacral ligaments each reach the cervix from a different angle, and each has a different 3D geometry within the pelvis. In 3D reconstruction from MRI, the cardinal ligaments lie at about an 18-degree angle from the cephalic-caudal body axis and merge with the uterosacral ligaments, which are oriented at about a 90-degree angle from the cephalic-caudal body axis—a 72-degree difference in direction (Figure 3).^14,15 The ligaments converge at the level of the cervix as previously described. Additionally, the cardinal ligament has been found to be longer than the uterosacral ligament by about 3 cm on average. The cardinal ligament also exhibits about double the curvature of the uterosacral ligament.¹⁵ Understanding the tension on the elements of a support structure depends on the load carried and the angle of the suspending element. Therefore, knowing the angles is necessary to understanding the mechanisms of failure.

The length and direction of the cardinal and uterosacral ligaments change in the setting of prolapse. These changes have been measured by Luo et al. with 3D stress MRI that documents the geometry both at rest and during maximal Valsalva. Broadly speaking, the cardinal ligaments lengthen, while the uterosacral ligaments rotate caudally during straining (Figure 4). At maximal Valsalva, the cardinal ligament elongates twice as much in those with anterior or posterior predominant prolapse than in normal controls. However, the change in length of the uterosacral ligament during maximal Valsalva is not significantly different between women with and without prolapse. The length of either ligament at rest does not differ significantly between cases and controls. The angle of the uterosacral ligament relative to vertical-body axis does change significantly more in cases (−32°±18°) than in controls (−23°±17°) during maximal Valsalva, whereas that of the cardinal ligament does not (Figure 4).¹⁶

Figure 4. Normal and prolapse ligament comparison under rest and maximum Valsalva.

Left panels: Normal and prolapse ligament comparison under resting and maximal Valsalva. Left side view of 3D models of one healthy individual at rest (a) and at maximal Valsalva (b), with their relationship to the normalized arcus tendineus fascia pelvis (ATFP) [turquoise lines extending from the public symphysis (P) to the ischial spines (yellow squares)], or the pubic symphysis to the ischial spine line (P-IS) for spatial reference. Left-side view of models of one individual with prolapse at rest (c) and at maximal Valsalva (d). PS and sacrum (S) are shown in the midsagittal plane. Ut: uterus, V: vagina, CL: cardinal ligament, USLd: deep uterosacral ligament. Right panel: Straight length and angle of ligaments at rest and maximal Valsalva. Average straight length and angle of the cardinal ligament (CL) and deep uterosacral ligament (USLd) are shown based on one average midsagittal image from a healthy woman, with outline for uterus, vagina, pubic symphysis, and sacrum. Reproduced from Luo 2013.¹⁶ ©DeLancey

MRI data can also be used to create computational models of the entire pelvic floor support system, including ligaments. The material properties of each tissue type are incorporated into the models to enable simulations of prolapse. These subject-specific 3D finite element models can simulate prolapse formation with increasing intra-abdominal pressure. The models can then be used to study how support structure (cardinal and uterosacral ligaments) impairment affects prolapse size, which will be discussed in more detail later in the paper.

In Vivo Ligament Mechanical Properties

The next step in understanding the apex is to examine the ligament properties in vivo. Historically, ligament laxity was thought to be responsible for prolapse, but more recent studies have in fact shown that ligament length seems to be a more significant factor. Ligament stiffness has been calculated in vivo using a device called a servo-actuator, which was connected to a tenaculum on the cervix. The servo-actuator measured the displacement of the cervix via a computer-controlled linear motor connected to a load cell that measured traction force (Figure 5). The force-displacement information was then used to calculate ligament stiffness. Results of these measurements indicate that only 19% of the variation in POP-Q point C (cervix) location can be explained by ligament stiffness (Figure 6).¹⁷ In these experiments, the greatest predictor of the variation in point C is the location of the cervix under minimal traction force (1N), which may represent straightening of the ligaments to their full length.¹⁸

Figure 5. Schematic of servo-actuator test setup.

Zero indicates the location of the hymen, with −/+ meaning above or below the hymen. S: sacrum; U: uterus; P: pubic bone; Cx: cervix. A, B, and C denote the servo actuator, force transducer, and surgical tenaculum, respectively; D the tripod; E the motor controller; and F the microprocessor. Vertical dashed-dotted line represents the initial location of the hymenal ring. This defines the origin (“0”) for the pelvic organ prolapse measurement system used to assess uterine position. The dashed long and short lines represent the CL and USL, respectively, under load. Note the 1 m long vertical suture suspending the weight of C. ©Luo, Ashton-Miller, and DeLancey

Figure 6. Cervix location force-displacement graph.

Force-displacement graph shows hyperelastic ramp curves (solid lines) for the 17 individual subjects. The X-axis shows cervix location in millimeters based around the hymen. The Y-axis shows force in Newtons. The dotted line indicates Pelvic Organ Prolapse Quantification system (POP-Q) point C for each subject. Modified from Smith 2013.¹⁷

Experiments with the servo-actuator device in women with prolapse have also confirmed that the apical ligaments demonstrate visco-hyperelastic behavior.¹⁹ Viscoelastic materials have the properties of both elastic materials (e.g., spring), in which the amount of deformation is directly proportional to the amount of force applied, and viscous materials, where how rapidly a force is applied affects the degree of deformation (e.g., silly putty).²⁰ A constant weight applied to a spring always results in the same elongation, and when the weight is removed, the spring always shortens back to its original length immediately. But in the case of a viscoelastic material, not only will it exhibit increasing elongation over time when lengthened by a force, but when that force is removed, it takes time to return to its original length. If the material also exhibits viscoplastic behavior, then it will never return completely to its original length when unloaded. This behavior explains how repeated stresses on the apical support tissues could lead to permanent lengthening of the ligaments over time. The aforementioned findings have been synthesized into the “slack cord” paradigm, which posits that the cardinal ligaments naturally rest in a curved state, but straighten over time as the uterus position descends due to various pelvic floor changes. The major shortfall of this paradigm, however, is that it does not account for histologic changes that occur in the ligaments. We have therefore expanded on the “slack cord” paradigm to develop a more comprehensive conceptual model that will be discussed later in this review.

Ligament stiffness does not have a large effect on uterine position until the ligaments have straightened.¹⁷ Interestingly, measures of ligament stiffness vary widely depending on how it is tested. Most published measurements of ligament stiffness are based on testing small pieces of excised cadaveric tissue using traditional engineering techniques (e.g., Rivaux et al. 2013²¹). This type of testing, however, has not represented measures that are in the physiological range normally found in vivo. We find that they turn out to be 650 times greater than that measured in the operating room (OR) in living women.²² Understanding which properties are most physiologically relevant is important to truly understanding the role of apical supports.

Clinical Assessment of Apical Support

Although normal diagnostic cutoff values for problems like gestational diabetes and hypertension, as well as standard laboratory values, are well known among clinicians, normal values for uterine support are not often discussed or widely known. In vivo measurements of apical support vary depending on the testing situation. This is confusing when normative values obtained during pelvic examination are compared with measurements made in the OR with traction on the cervix. In the OR, for women with normal support, the location of the lateral margin (“3 o’clock”) of the cervix with traction is 0.8 cm (standard deviation, 1.4 cm) above the hymen and ranges from 4.8 cm above to 2 cm below. By contrast, the mean location of the POP-Q point C (cervix) in clinic is 6 cm above the hymen with a standard deviation of 1.95 and with 99% occurring between the range of −11 and −2.²³ It is important to recognize normal values in order to correctly identify those that lie outside of the normal range, as in the case of prolapse.

In women with prolapse, when point C is measured in the clinic with Valsalva and subsequently in the OR with traction, point C is on average 3.5 cm lower in the OR.¹ More recent in vivo studies conducted with the servo-actuator and load cell device that measures movement in response to specific applied force have attempted to answer the question of how much force it takes to move the uterus a physiologic distance—or the distance observed on MRI when a patient exerts maximum Valsalva. The answer is surprisingly small—only 90 g of force (the weight of a chicken egg).²² This finding implies that the common practice of applying traction to the cervix preoperatively in the OR to evaluate apical support is likely to greatly overestimate the extent of uterine descent compared to physiologic forces. It also shows that factors other than the ligament properties affect the location of the cervix.

Apical Relation to Cystocele and Rectocele

Both cystocele and rectocele are associated with loss of apical support. More than half of the size of anterior compartment prolapse is explained by the degree of apical descent and vice versa.^1,2 Accordingly, there is a strong linear relationship between point C (cervix) and point Ba (most protruding point of anterior vaginal wall) when measured with POP-Q.²⁴ Mechanistically, the anterior vaginal wall can be visualized as a trapezoid—with the apex as the broad side of the shape, held upward by the suspending action of the cardinal and uterosacral ligaments. When apical support fails, the broad side of the trapezoid swings downward like a trap door, allowing the bladder to descend with it and thereby creating a cystocele (Figure 7).²

Figure 7. Conceptual diagram showing the mechanical effect of apical descent on bladder location.

A. The trapezoidal plane of the pubocervical fascia on which the bladder rests. B. Loss of apical support allows pubocervical fascia to descend and bladder to move downward. Reproduced from DeLancey 2002.³⁵ ©DeLancey

When there is a 90% impairment of apical support in biomechanical models, anterior vaginal wall prolapse increases by more than five-fold. While not as dramatic, this effect is still present with increases in levator ani defects.²⁵ The bilinear relationship between exposed vaginal wall length and extent of anterior and apical prolapse has been elucidated as well.²⁶ When combined with anterior vaginal wall length, apical descent explains up to 77% of variation in anterior wall prolapse size.²⁷ Furthermore, when the triad of apical location, paravaginal location, and hiatus size are considered together, they explain 83% of variation in cystocele size.²⁸ Of note, apical impairment and paravaginal descent are highly correlated (R=0.84), which suggests that these are essentially different measurements of a single phenomenon—confirming the findings based on computational modeling study.²⁹ Although association does not imply causation, the theory that cystocele may cause apical descent by creating a downward traction force on the uterus is supported by operative findings. In women who undergo anterior repair alone with the uterus in situ, point C is higher postoperatively than preoperatively in the majority of cases.³⁰

Loss of apical support contributes to rectocele size as well. Rectocele modeling using a subject-specific 3D finite element model (Figure 8) examined simulated rectocele size (at POP-Q point Bp, or most protruding point of posterior vaginal wall) with increases in abdominal pressure. This study showed that rectocele size increased more dramatically in the setting of 60% apical impairment when compared to the no apical impairment condition.³ Clinically, the correlation between apical and posterior wall support is evident in POP-Q measurements. When Bp is recorded as stage 3 or 4, it has a strong positive correlation with point C.³¹ However, the association between apical support and rectocele is not as strong as that seen with the anterior vaginal wall. Women with posterior-predominant prolapse have a POP-Q point C (representing apical support) that is about 3.5 cm higher than women with anterior-predominant prolapse.³²

Figure 8. 3D finite element modeling.

Model development: (A) and (B) MR-based 3D pelvic floor reconstruction model with and without bone in a left three-quarter view in the standing posture. PB: pubic bone; U: uterus; V: vagina; CL: cardinal ligament; USL: uterosacral ligament; PeB: perineal body; PM: perineal membrane; LA: levator ani; AVW: anterior vaginal wall; PVW: posterior vaginal wall; APS: anterior paravaginal support; PArcus: posterior arcus tendineus fascia pelvis; PPS: posterior paravaginal support; AS: anal sphincter; ATLA: arcus tendineus levator ani. Rectocele characteristics shown in left three-quarter view for the finite element model at rest (C) and under load (D). Downward displacement, forward protruding, and perineal descent behaviors are seen for the rectocele. Modified from Luo, Chen 2015.³

Synthesis

The role of apical descent must be incorporated into an overall conceptual model that also includes the many different types of data that currently exist. In addition, such a model must relate to observations about levator damage and hiatal closure that are well-established as important factors in causing prolapse. Such a model can also be useful in identifying knowledge gaps and directing future research.

The aforementioned observations inform an overall disease model of prolapse. In addition to apical support, the genital hiatus has a very important role in maintaining normal pelvic support. Increasing genital hiatus size is highly associated with loss of apical support.³³ In women with normal support, the hiatus is closed and the pressures on the apical support structures are balanced. When the hiatus widens, increasing amounts of the vaginal wall become exposed to atmospheric pressure, which is significantly lower than abdominal pressure. The resultant pressure differential leads to prolapse³⁴ and might lead to subsequent tension on the apical ligaments (Figure 9).

Figure 9. Diagrammatic representation of interactions between levator ani muscle, anterior vaginal wall prolapse, and cardinal/uterosacral ligament suspension.

With normal levator function, (a) the vaginal walls are in apposition, and anterior and posterior pressures are balanced. Levator connective tissue damage (b) results in hiatal opening, and the vagina becomes exposed to a pressure differential between abdominal and atmospheric pressures. This pressure differential (c) creates a traction force on the cardinal ligament (CL) and uterosacral ligament (USL). Modified from DeLancey 2012.³⁶ ©DeLancey

These observations suggest a conceptual framework for the apical ligaments that we call the “3-Phase Apical Support Theory” (Figure 10). The first phase, straightening, occurs when ligaments go from a curved resting state to straightened, but without tension. The second phase is tensioning, during which ligaments lengthen beyond their maximal straightened length and are then subjected to tensioning forces, but still within their elastic limits, before recovering to their original length when there is no longer tension placed on the ligaments. These first two phases are physiologic and reversible. Straightening and tensioning happen, for instance, during a dilation and curettage procedure, when the uterus is pulled down to the hymen before promptly returning to its original position when the tension is released. If the cervix descends below the physiologic limits of extension, however, then some tissue change must have taken place—which brings us to the third and final phase. This phase is structural change, which we hypothesize involves permanent tissue adaptation and remodeling resulting from repetitive loading over time; this manifests as increased ligament resting length. The third phase is pathologic and represents a failure of the normal support of the apex.

Figure 10. New 3-phase apical support theory.

Illustration of conceptual framework for sequential straightening, tensioning, and remodeling of cardinal (CL) and uterosacral (USL) ligaments, leading to progression of prolapse. ©DeLancey

In summary, it is widely known that the apical ligaments play a key role in the support of the uterus and upper vagina. Over the last several decades, our knowledge about the specific biochemical and biomechanical properties of the apical ligaments has significantly evolved. We now have a much more nuanced understanding of some of the ways in which uterovaginal prolapse relates to changes in the protein expression and load bearing of the ligaments within a complex pelvic support system. We hope that future research will investigate the concept of tissue adaptation and remodeling in the ligaments and will elucidate the question of which biochemical changes in the ligaments are a cause—versus an effect—of prolapse, as well as ways in which we can use our biomechanical understanding of the apical ligaments to improve surgical techniques for prolapse repair.