Does levator ani injury affect cystocele type?




To determine the prevalence of levator ani injury in patients with different types of cystocele, as defined by translabial ultrasound, in order to shed light on potential pathophysiological mechanisms.


The datasets of 222 women who had undergone a physical examination, urodynamic testing and four-dimensional (4D) pelvic floor ultrasound were evaluated offline for prolapse, levator ani hiatal dimensions and levator ani trauma using tomographic ultrasound imaging (TUI), blinded against all clinical and urodynamic data. Cystoceles reaching below the symphysis pubis on ultrasound examination were classified based on bladder neck position, retrovesical angle (RVA) and urethral rotation as Green II (cystourethrocele) or Green III (cystocele with intact RVA).


Of 102 women who had a cystocele reaching below the symphysis pubis, 63 were classified as a Green type II cystocele and 39 as a Green type III cystocele. Women with Green type III cystoceles were older (59.4 vs. 48.7 years, P < 0.001), and had more severe prolapse (71 vs. 43%, P = 0.004) and objective voiding dysfunction (39 vs. 18%, P = 0.018). Women with Green III cystoceles also had larger hiatal dimensions and were more often diagnosed with an avulsion of the levator ani muscle (69 vs. 35%, P = 0.001).


A cystocele with an intact RVA is more likely to be associated with avulsion injury of the levator ani muscle and thus more likely to be caused by birth-related trauma. This contradicts the commonly held belief that such cystoceles are caused by central rather than by lateral fascial defects. Copyright © 2010 ISUOG. Published by John Wiley & Sons, Ltd.


Trauma to the levator ani muscle has recently been shown to be a common consequence of vaginal childbirth, affecting between 15 and 35% of vaginally parous primiparae1, 2. There is an established relationship between levator ani (puborectalis) muscle trauma and female pelvic organ prolapse3, 4, which at least partly explains the epidemiological link between vaginal childbirth and prolapse. This relationship seems to be strongest for cystocele and uterine prolapse4.

Cystoceles with a similar degree of anterior vaginal wall descent may in one patient be associated with stress urinary incontinence (SUI) and with voiding dysfunction in another. In the 1970s, this observation led to a radiological classification of cystocele by Green5, based on bladder neck position, posterior urethrovesical (PUV) or retrovesical angle (RVA) measurement (the angle between the proximal urethra and the trigonal surface of the bladder) and the degree of urethral rotation (proximal urethra) on Valsalva (Figure 1). Green type I describes a cystocele with an open RVA of 140° or more and urethral rotation below 45°. A Green type II cystocele, also named a cystourethrocele, has an open RVA of 140° or more and urethral rotation of 45–120°. A Green type III cystocele describes an isolated cystocele with an intact RVA below 140° and urethral rotation of 45° or more (Figures 1 and 2).

Figure 1.

Green5 classification of incontinence and cystocele type. (a) Normal configuration of the bladder neck. (b) Green type I: open retrovesical angle (RVA) ≥ 140°, urethral rotation < 45°. (c) Green type II: open RVA ≥ 140° and urethral rotation 45–120°. (d) Green type III: intact RVA < 140°.

Figure 2.

Translabial ultrasound appearance of cystourethrocele or Green type II (a) and cystocele with intact retrovesical angle or Green type III (b). Arrows represent the bladder neck.

Various authors have speculated that the main two types of cystocele, namely cystourethrocele (or Green type II) and cystocele with an intact RVA (or Green type III), may be caused by different anatomical situations (i.e. lateral vs. central fascial defects)5, 6. The anatomical distinction between the two types of cystocele seems to have clinical validity and, in the view of the authors, is worthy of further investigation even though the Green classification has fallen into disuse.

The aim of this study was to estimate the prevalence of levator ani injury in patients with different cystocele types, as defined by four-dimensional (4D) translabial ultrasound, in order to shed light on potential pathophysiological mechanisms.


Between January 2007 and May 2008, 222 women presenting with symptoms of lower urinary tract and pelvic floor dysfunction were assessed in a tertiary urodynamic service provided by a subspecialist urogynecologist. The routine assessment included an interview, clinical examination, independent flowmetry, multichannel urodynamic testing and pelvic floor imaging with three- and four-dimensional (3D/4D) translabial ultrasound, as previously described7. The interview, clinical examination, urodynamic testing and translabial ultrasound were performed by the same operator (H.P.D.) or under his direct supervision.

All clinical and urodynamic definitions used were those endorsed by the standardization committee of the International Continence Society (ICS)8. Prolapse symptoms were defined as a vaginal lump or a dragging sensation in the vagina. Symptoms of voiding dysfunction were defined as a departure from normal sensation or function, experienced by the woman during or following the act of micturition, such as hesitancy, poor stream or straining to void9.

Patients underwent a standardized (local, non-validated) interview assessing their symptoms, and were included in the analysis regardless of symptom severity but rather stratified based on clinical findings. The clinical examination was performed using the prolapse quantification system (POP-Q) of the ICS10. Cystocele was defined as descent of the anterior vaginal wall, most commonly because of bladder prolapse. A significant cystocele on physical examination was defined as cystocele Stage 2 or above. Patients with Stage I cystoceles were not included in the analysis.

Independent flowmetry and multichannel urodynamic testing was performed using fluid-filled catheters (Neomedix Uromac v. 3.6, Sydney, Australia). The urodynamic definitions used were those endorsed by the standardization committee of the ICS8. Voiding dysfunction was defined as a maximum flow rate centile of 5 or below and/or a residual of 50 mL or more on independent flowmetry; and sensory urgency as a strong desire to void at 300 mL or less.

Pelvic floor imaging was performed with 3D/4D translabial ultrasound, as previously described7, using a Voluson 730 Expert ultrasound system (GE Medical Systems, Zipf, Austria) with a 4–8-MHz volume transducer (acquisition angle 85°). Ultrasound imaging and volume data acquisition were performed after bladder emptying and with the patient supine in a modified lithotomy position. For the assessment of levator ani muscle integrity, volumes were obtained on maximal pelvic floor contraction, or at rest in those unable to contract (n = 9). The puborectalis portion of the levator ani muscle was observed as a V-shaped structure surrounding the anorectum, vagina and urethra, as previously described11. Volumes were obtained at rest, and upon maximal contraction and maximal Valsalva. The diagnosis of levator ani avulsion injury was made at the time of the urodynamic assessment, on the basis of tomographic ultrasound imaging (TUI)12, with eight slices obtained at 2.5-mm intervals, from 5 mm below to 12.5 mm above the plane of minimal hiatal dimensions, thus encompassing the entire puborectalis muscle (Figure 3). An avulsion was rated as positive if the plane of minimal dimensions and the two slices immediately cranial (2.5 and 5 mm cranial) all showed an abnormal insertion. For each level the presence of an avulsion gives a score of 1 for unilateral avulsion and 2 for bilateral avulsion. This gives a possible maximal score of 16 for complete bilateral avulsion. Good repeatability of the sonographic diagnosis of puborectalis muscle avulsion injury has been previously demonstrated11–13. For clarity and simplicity we did not subdivide the groups according to unilateral or bilateral avulsion injury.

Figure 3.

Rendered volume (a) and tomographic ultrasound image (b) of an avulsion injury of the puborectalis muscle, as seen on translabial volume ultrasound. The tomographic image shows a coronal reference plane (b, top left) and eight consecutive axial plane slices obtained at 2.5-mm intervals, from 5 mm below to 12.5 mm above the plane of minimal hiatal dimensions. The avulsion injury is indicated (*) in the rendered volume (a) and tomographic slices (b).

The post-processing analysis of ultrasound volume datasets for Green typing and levator ani biometry and injury assessment was performed offline (4D View 5.0, GE Medical systems), by a different operator (V.H.E.), who was blinded to all clinical and urodynamic data. All evaluation and statistical analyses were based on the post-processing analysis. Cystoceles reaching below the symphysis pubis on ultrasound were classified based on bladder neck position, RVA and urethral rotation as to their Green type5 (Figure 1). Green type I cystoceles were excluded from the analysis owing to their comparative rarity, and because they often seem to be iatrogenic, such as after anterior colporrhaphy.

This retrospective analysis is an approved extension of a project approved by the Institutional Human Research Ethics Committee (reference SWAHS HREC 05–029). Statistical analysis was performed with Minitab V.13 (Minitab, State College, PA, USA). All continuous parameters were tested for normality and found to be normally or near-normally distributed on Kolmogorov–Smirnov testing, except for residual urine. For this, we used a Mann–Whitney U-test. In other situations we employed t-tests and chi-square statistics. Prolapse stage analysis was based on ordinal stages. A P of < 0.05 was considered statistically significant.


We identified 222 ultrasound volume datasets of patients seen between January 2007 and May 2008. The mean age was 55 (range, 19–87) years and median parity was 3 (range, 0–10). Thirteen women (5.9%) were nulliparous, and six (2.7%) had delivered by Cesarean section only. Patients presented with SUI (78%), urge urinary incontinence (72%), urinary frequency (31%), nocturia (48%), voiding dysfunction (24%) and prolapse (39%) symptoms. Clinically, significant pelvic organ prolapse was detected in 58%. Urodynamic testing showed urodynamic stress incontinence (67%), detrusor overactivity (25%), voiding dysfunction (30%) and sensory urgency (16%). Overall, 86 women (38%) had levator avulsion defects and the mean for abnormal slices on TUI was 4.

Thirty-nine datasets were excluded from further analysis because of previous surgery affecting bladder neck anatomy, such as anterior vaginal wall repair, colposuspension, slings of any kind, or transobturator mesh placement. Five datasets could not be assessed because of missing or corrupted volumes. Of the remaining 178 datasets, 102 women had a cystocele below the symphysis pubis on ultrasound and were classified as a Green type II (n = 63) or a Green type III (n = 39) cystocele.

Women with Green type III cystoceles were older (59.4 vs. 48.7 years; P < 0.001), were more likely to have significant clinical prolapse (71 vs. 43%; P = 0.004), but were less likely to report SUI (64 vs. 92%; P < 0.001). They were non-significantly less likely to be diagnosed with urodynamic stress incontinence, but were more likely to have objective voiding dysfunction (39 vs. 18%; P = 0.018) and larger bladder capacity (482 vs. 443 mL; P = 0.019). A comparison of demographic, clinical and urodynamic data is given in Table 1.

Table 1. Comparison of demographic, clinical and urodynamic data in 63 women with Green type II cystoceles and in 39 women with Green type III cystoceles
CategoryMeasureGreen type II (n = 63)Green type III (n = 39)P
  1. Data are given as mean ± SD or %.

DemographicsAge (years)48.7 ± 13.459.4 ± 10.6< 0.001
 Instrumental delivery20.640.80.013
SymptomsStress urinary incontinence9264< 0.001
 Prolapse symptoms43710.004
UltrasoundCystocele descent relative to symphysis pubis (mm)− 13.7 ± 9.1− 25 ± 14.2< 0.001
Urodynamic dataUrodynamic stress incontinence77650.133
 Voiding dysfunction18390.018
 Bladder capacity (mL)443 ± 94482 ± 690.019

Women with Green type III cystoceles were more likely to suffer from levator ani muscle avulsion injury (69 vs. 35%; P = 0.001), and showed a higher number of abnormal slices on TUI (8.0 vs. 3.4; P = 0.001) than those with a Green II cystocele. This finding was independent of clinical cystocele stage, although, because of reduced power, not all associations reached significance on subgroup analysis. The location of abnormal slices had no bearing on the likelihood of cystocele type. The association between levator injury and Green type III cystocele was present in all eight standard levels assessed by TUI. Ultrasound data are given in Table 2.

Table 2. Comparison of hiatal dimensions and levator morphology in patients with Green types II and III cystoceles
MeasureGreen type II (n = 63)Green type III (n = 39)P*
  • Data are given as mean ± SD or %.

  • *

    Double-sided t-test, except for levator avulsion defect, calculated using the chi-square test.

  • TUI, tomographic ultrasound imaging.

At rest   
 Midsagittal hiatal diameter (cm)5.85 ± 0.865.41 ± 0.770.009
 Coronal hiatal diameter (cm)4.57 ± 0.694.96 ± 0.750.011
 Hiatal area (cm2)17.69 ± 4.2520.32 ± 5.530.014
On Valsalva   
 Midsagittal hiatal diameter (cm)6.84 ± 1.007.22 ± 0.930.053
 Coronal hiatal diameter (cm)5.76 ± 0.916.16 ± 0.930.036
 Hiatal area (cm2)30.61 ± 7.7034.19 ± 8.910.041
On pelvic floor muscle contraction   
 Midsagittal hiatal diameter (cm)4.61 ± 0.725.15 ± 0.820.002
 Coronal hiatal diameter (cm)4.37 ± 0.744.78 ± 0.970.034
 Hiatal area (cm2)14.24 ± 3.3017.33 ± 5.300.003
Levator avulsion defect35690.001
Abnormal slices on TUI (score)3.4 ± 5.78.0 ± 6.70.001


Vaginal childbirth has been linked to SUI and pelvic organ prolapse in epidemiological studies14, 15. Avulsion of the puborectalis muscle from the pelvic sidewall, also termed levator ani avulsion injury, has been recognized as a major etiological factor of pelvic organ prolapse, directly related to vaginal delivery1–3, and diagnosed both on magnetic resonance imaging3 and ultrasound4. Women with this type of levator ani trauma have altered pelvic floor muscle anatomy and function3, 16, 17.

It is widely held that one of the main etiological factors in the pathogenesis of stress incontinence is loss of adequate support to the bladder base, vesical neck and proximal urethra with distortion of the urethrovesical anatomy5. Jeffcoate and Roberts18, 19 were among the first to associate the importance of the anatomical configuration of the urethrovesical junction and proximal urethra with the understanding of the continence mechanism in a series of radiological studies. They concluded that the presence of an intact PUV angle (an RVA) was essential for continence, regardless of the position of the bladder neck and urethra with respect to the symphysis pubis. Many patients with significant bladder descent and an intact RVA are continent, whereas others with severe stress incontinence have a relatively less mobile bladder base, with an obliterated RVA and resultant funnelling and descent of the bladder neck to the most dependent portion of the bladder.

The Green classification used in this study5 categorizes cystocele based on the configuration of the PUV angle or RVA and the mobility of the urethral axis (urethral rotation in degrees) as originally seen by urethrocystography. A Green type II has also been termed a cystourethrocele5 because the vaginal wall overlying the urethra and the bladder base forms a smooth surface. It has been hypothesized that Green type II represents a more profound weakening of the supports of the proximal urethra and the bladder neck5, 6. In Green type III (or isolated cystocele with an intact RVA), while the distal urethra retains a relatively normal urethral axis, the proximal urethra may rotate as much as in a cystourethrocele, resulting in urethral kinking. This may explain why these patients usually do not have stress incontinence, even though their cystocele may be quite large5.

Anterior vaginal wall prolapse was traditionally believed to be secondary to attenuation of the endopelvic fascia. White, in 1909, described a different etiology, namely detachment of the pubocervical fascia from its lateral attachment to the arcus tendineus fasciae pelvis, which is generally termed a ‘paravaginal defect’20. Richardson et al.21 described discrete breaks in the pubocervical fascia divided into midline, transverse and lateral (paravaginal defects). Richardson defined a paravaginal defect as detachment of the pubocervical fascia from the pelvic sidewall, and observed this in 67% of patients with anterior vaginal wall prolapse22. DeLancey agrees with the basic concept23, 24, although several attempts to validate this by imaging25 or intraoperatively26 have been unsuccessful. Other authors have stated that a cystourethrocele is more often related to paravaginal defects22, 27, than to midline stretching, in women with stress incontinence and urethral hypermobility, and this would naturally imply a traumatic etiology for these conditions.

Our study shows that a cystourethrocele (Green type II) is less likely to be associated with avulsion injury of the levator ani (puborectalis muscle), independent of the degree of prolapse, than a cystocele with an intact RVA (Green type III). This argues against the commonly held concept of a traumatic pathogenesis for cystourethrocele (Green type II), and argues for a traumatic etiology for a cystocele with an intact RVA (Green type III) because it is difficult to see how the paravaginal fascia could remain intact after major damage to the insertion of the entire puborectalis muscle. Therefore, if one wanted to postulate that paravaginal/lateral fascial defects cause cystocele, then this would be much more likely to apply in Green III cystocele than in Green II cystocele.

There are several limitations to this study which must be addressed. This was a retrospective analysis of routinely obtained data in a cohort of patients who attended a urodynamics clinic presenting with significant symptoms of pelvic floor dysfunction. Thus, our results may not be applicable to the general population. Also, most of our patients were Caucasians and our findings cannot be generalized to other ethnic groups. One may also criticise the fact that we did not try to assess the paravaginal fascia. However, to date, no-one has been able to present an imaging method that reproducibly demonstrates defects of this fascia, and even intraoperative diagnosis (whether laparoscopically or during laparotomy) is limited by the very real possibility of iatrogenic defects arising from the dissection process, because the paravaginal fascia is often a rather fragile structure.

A significant strength of our study is that offline ultrasound analysis was performed by an operator who was fully blinded to clinical and urodynamic data, and that the statistical analysis was based on this operator's assessment. This would serve to avoid any assessment bias.

In conclusion, we have shown that puborectalis avulsion injury is more prevalent in women with a Green type III cystocele (i.e. an intact RVA). These women more commonly suffer from voiding dysfunction than SUI. Women with a Green type II cystocele are more likely to suffer from stress incontinence, but are less likely to have a concomitant avulsion injury. This suggests that, in contrast to what has been commonly assumed, a cystourethrocele (Green type II) is less likely to be caused by paravaginal defects than a cystocele with an intact RVA (Green type III).