Vaginal childbirth has a substantial effect on pelvic organ supports, which may be mediated by levator ani (LA) avulsion or hiatal overdistension. Although the impact of a first vaginal delivery on the hiatus has been investigated, little is known about the effect of subsequent births. This study was designed to evaluate the association between vaginal parity and hiatal dimension.
Retrospective observational study.
A tertiary urogynaecological unit in Australia.
A total of 780 archived data sets of women seen for symptoms of lower urinary tract and pelvic floor dysfunction.
Standardised in-house interview and assessment using the International Continence Society (ICS) pelvic organ prolapse quantification (POP-Q), and four-dimensional translabial ultrasound. Offline analysis for hiatal dimensions was undertaken blinded to history and clinical examination.
Main outcome measures
Hiatal area on maximum Valsalva.
Of 780 women, 64 were excluded because of missing ultrasound volumes, leaving 716 for analysis: 96% (n = 686) were parous, with a median parity of three (interquartile range, IQR 2–3), and 91.2% (n = 653) were vaginally parous. Levator avulsion was found in 21% (n = 148). The mean hiatal area on Valsalva was 29 cm2 (SD 9.4 cm2). On one-way anova, vaginal parity was significantly associated with hiatal area (P < 0.001). Most of the effect seems to occur with the first delivery. Subsequent deliveries do not seem to have any significant effect on hiatal dimensions. This remained true after controlling for potential confounding factors using multivariate regression analysis (P = 0.0123).
Vaginal parity was strongly associated with hiatal area on Valsalva. Most of this effect seems to be associated with the first vaginal delivery.
Symptomatic pelvic organ prolapse (POP) is a significant problem affecting up to 15% of women. The lifetime risk of undergoing surgery for POP was estimated to be 20% in a recent Australian study. In the USA, it has been estimated that approximately 200 000 prolapse procedures are performed annually, with a direct cost of over $1 billion.
Despite its high prevalence, the aetiology of POP remains incompletely understood. It has been attributed to multiple factors including childbirth, changes in collagen metabolism,[5, 6] obesity, ageing,[8, 9] and menopause. Evidence from epidemiological and observational cohort studies suggests that vaginal birth is the main aetiological factor.[7, 11-13]
During vaginal delivery the puborectalis muscle is exposed to a high degree of stretch, with an estimated stretch ratio of 1.5–3.5.[14-16] The degree of muscle lengthening seems to vary greatly, ranging from 25 to 250%. Muscle physiology studies have shown that substantial injury, both macro- and microscopic, may occur if skeletal muscle fibre is stretched to more than 1.5 times its original length. It is therefore not surprising that 10–35% of women suffer a levator avulsion, i.e. traumatic injury of the puborectalis muscle at its bony insertion.[18-21] This results in hiatal enlargement of 20–30%, and a more distensible and less contractile pelvic floor muscle.[23, 24] Apart from levator avulsion, vaginal birth may also result in another form of visible change in functional levator anatomy. In a study on peripartum change in hiatal dimensions, over 28% of primiparous women were diagnosed with irreversible hiatal overdistension or ‘levator microtrauma’ at 4 months postpartum, independent of avulsion, and without evidence of healing at the 2-year follow up. Levator avulsion and hiatal dimensions have both been shown to be strongly associated with symptoms and signs of POP,[21, 24, 27-29] and with recurrence after pelvic floor reconstructive surgery.[30-35] It is a reasonable assumption that levator trauma may be the missing link between childbirth and POP.
The effect of a first delivery on levator biometry has been investigated in prospective studies;[22, 25, 26, 36] however, the impact of subsequent births is not well defined. We therefore undertook a retrospective study to evaluate the relationship between vaginal parity and hiatal dimensions using a large data set of women presenting with symptoms and signs of pelvic floor dysfunction, mainly urinary incontinence and female pelvic organ prolapse, in a tertiary urogynaecological practice. Our hypothesis is that vaginal parity is associated with hiatal dimensions.
This was a retrospective study involving 780 women attending a tertiary urogynaecological unit with lower urinary tract symptoms and/or pelvic floor dysfunction, between July 2009 and February 2012. All patients had undergone a standardised interview to assess symptoms and medical history, a clinical examination, which included the International Continence Society (ICS) pelvic organ prolapse quantification (POP-Q), and translabial three- or four- dimensional (3D/4D) pelvic floor ultrasound using a GE Kretz 730 Expert system (GE Medical Systems, Zipf, Austria), after bladder emptying, in the supine position at rest and upon maximum Valsalva, as previously described. The volume acquisition angle was set to the system maximum of 85°. Levator co-activation upon Valsalva manoeuvre was avoided with meticulous observation and patient education by visual biofeedback. At least three volume cine loops upon Valsalva were acquired. The volume upon maximum Valsalva manoeuvre resulting in the greatest degree of pelvic organ descent was used for analysis. ‘Significant prolapse’ on clinical examination was defined as ICS POP-Q stage 2 or higher.
Off-line analysis of ultrasound data sets for dimensions of the levator hiatus was undertaken at a later date, blinded against all clinical data, using the proprietary software 4d view 7.0 (GE Medical Systems). Hiatal area was measured as previously described. In brief, the plane of minimal hiatal dimensions was identified in the midsagittal orthogonal plane, where the distance between the hyperechogenic posterior aspect of the symphysis pubis and hyperechogenic anterior border of the levator ani muscle, just posterior to the anorectal muscularis, is shortest. Hiatal area was measured in rendered volumes of 1–2 cm in thickness containing the plane of minimal hiatal dimensions (Figure 1).
This retrospective study was approved by the Institutional Ethics Committee (NBMLHD HREC ref. 12-07). Statistical analysis was carried out using spss 16 (SPSS, Chicago, IL, USA) and minitab 13 (Minitab, State College, PA, USA). Normality was assessed visually and using the Kolmogorov–Smirnov method. A linear model was fitted to hiatal area, with vaginal parity fitted as a categorical covariate. This model was extended in the multivariate analysis by including possible confounding factors such as age, body mass index, previous hysterectomy, previous incontinence/prolapse surgery, vaginal operative delivery, levator avulsion, and significant prolapse as additional covariates. P < 0.05 was regarded as statistically significant.
A total of 780 women were seen during the study period. Of these, 64 were excluded for missing ultrasound volume data because of equipment unavailability in 59 and operator error in five, leaving 716 women for analysis. All subsequent analyses pertain to these 716 women.
The population age ranged from 18 to 85 years, with a mean of 57 years (SD 13 years). Body mass index ranged from 17 to 56 kg/m2, with a mean of 29 kg/m2 (SD 6 kg/m2). A total of 686 (96%) women were parous with a parity ranging from 0 to 10, and with median parity of three (interquartile range, IQR 2–3). Mean age at first delivery was 24 years (SD 5 years). The records indicated that 91% (n = 653) were vaginally parous and 12% (n = 86) had undergone instrumental deliveries with either forceps or vacuum. Thirty-three (4.6%) women had delivered by caesarean section only (Table 1). The main presenting symptoms were stress incontinence (n = 535, 75%), urge incontinence (n = 536, 75%), and symptoms of prolapse (n = 370, 52%). Other symptoms included nocturia (n = 327, 46%), urinary frequency (n = 229, 32%), and symptoms of voiding dysfunction, such as urinary hesitancy, poor or interrupted flow, or straining to void (n = 218, 30%). There was a history of hysterectomy, incontinence, and/or prolapse surgery in 30% (n = 212) and 22% (n = 156) of the women, respectively. Clinically, 74% (n = 526) had significant POP (ICS POP-Q stage 2 or higher), involving the anterior (n = 434, 61%) and posterior (n = 380, 53%) compartments in the majority. Significant uterine prolapse and enterocoele was noted in 8% (n = 60) and 5% (n = 36) of women, respectively. Levator avulsion was diagnosed in 21% (n = 148), and it was bilateral in 43% of these women (n = 64/148), i.e. 8.9% of the total population (n = 64/716). The hiatal area on Valsalva ranged from 9.1 to 68.7 cm2, with a mean of 29 cm2 (SD 9.4 cm2): 21.5 cm2 (SD 9.1 cm2) for the vaginally nulliparous (P0); 29.1 cm2 (SD 9.3 cm2) for P1; 29.2 cm2 (SD 8.6 cm2) for P2; 30.7 cm2 (SD 9.2) for P3; 30.9 cm2 (SD 10.3 cm2) for P4; and 29.3 cm2 (SD 9.1 cm2) for P5.
Table 1. Demographic and delivery characteristics of study population
Age (mean, SD)
57 (13) years
Age at first delivery (mean, SD)
24 (5) years
Median parity (IQR)
Body mass index (mean, SD)
29 (6) kg/m2
Delivery characteristics (%)
History of instrumental delivery
Delivered exclusively by caesarean section
As there was no significant difference in hiatal dimensions (P = 0.21) between nulliparous women (n = 30) and parous women exclusively delivered by caesarean sections (n = 33), we merged those two groups as ‘vaginal parity = 0’. On one-way analysis of variance (anova), mean hiatal area on Valsalva was found to be significantly associated with the number of vaginal births (P < 0.001). The difference in mean hiatal area on Valsalva was most marked between women with no vaginal birth, i.e. nulliparous women and women who had delivered by caesarean sections only, and women with one vaginal birth. The difference for each additional vaginal birth was less obvious. It appears that most of the effect of vaginal childbirth on hiatal dimension was associated with the first delivery. Subsequent deliveries did not seem to have any significant effect on hiatal area on Valsalva (Figure 2).
On multivariate analysis, a similar pattern was observed (Figure 3) after controlling for potential confounding factors such as age, body mass index, previous hysterectomy, previous incontinence/prolapse surgery, vaginal operative delivery, and levator avulsion (P < 0.001; Figure 3; Table 2). Controlling for prolapse stage made the association between hiatal area and vaginal parity non-significant (P = 0.08), but upon controlling for significant prolapse (ICS POP-Q stage 2 or higher), the significant association remains (P = 0.01).
Table 2. Univariate analysis of the association between potential confounding factors and hiatal area on Valsalva
Potential confounding factors
Despite being non-significant predictors of hiatal area, these factors were forced into the multivariate analysis as they are regarded as important factors in the literature.
Prospective studies on primiparous women evaluating peripartum change in hiatal area and delivery mode have shown an association between vaginal delivery and hiatal dimensions.[22, 36] In the latter study, Toozs-Hobson et al. reported a significant increase in hiatal area at rest and upon maximum Valsalva 6 weeks and 6 months postpartum in the vaginal delivery group. In a more recent study using transperineal ultrasound, a significant increase in hiatal area on Valsalva, of 28 and 6% in women with and without levator avulsion, respectively, was demonstrated. In this current retrospective study using data obtained in a tertiary urogynaecological practice we have also found a significant effect of a first vaginal delivery on hiatal area, concurring with reports in the literature. It also conforms to findings in epidemiological studies showing that the prevalence of symptomatic POP doubled after vaginal delivery, compared with caesarean section, two decades after one birth.
Our results suggest that most of the effect of vaginal parity on hiatal dimensions seems to result from the first vaginal birth, and that subsequent deliveries do not seem to have any substantial influence on hiatal biometry. This is in agreement with a much smaller prospective study on the effect of a second delivery on the pelvic floor, where a second vaginal birth was not found to have a significant impact on hiatal dimension. The findings are largely in agreement with epidemiological studies that demonstrate a rather minor effect of subsequent births. In the Women's Health Initiative study, the odds ratio for uterine prolapse was 2.13 for the first birth and 1.10 for each subsequent birth, and 1.91 for cystocele and 2.22 for rectocoele, with each subsequent birth conferring a risk of 1.21 for both cystocoele and rectocoele. However, in a study describing the epidemiology of hospital admission for POP a much stronger association with increasing parity was noted, with four-fold and eight-fold increases in risk in mothers of one and two children, respectively. The discrepancy may be explained by the fact that the pathogenesis of POP is likely to be multifactorial. Apart from levator integrity and hiatal dimensions, other factors, for instance, connective tissue or fascial defects, may also play a role, and such potential aetiological factors were not addressed by our current study design. This has to be acknowledged as a weakness.
Strengths and limitations
The main strength of this study is the large data set involving more than 700 patients; however, its retrospective study design is a weakness. In order to study the research question, a long-term prospective study comparing antenatal and postpartum hiatal area after each vaginal delivery would be preferable. Such a study, however, is difficult to accomplish in view of the need for a large sample size and extended follow-up in women who are likely to be highly mobile at this time in their lives. Another criticism may be that one should assess not just hiatal area but also other anatomical factors associated with POP, such as organ mobility and the integrity of the rectovaginal septum. Furthermore, it may be argued that women with a history of prolapse surgery should be excluded from the study; however, on multivariate analysis controlling for a number of potential confounding factors, including ‘history of prolapse surgery’, the association between vaginal parity and hiatal area on Valsalva remained significant.
Another potential confounding factor is the presence of prolapse, which was common in this population. Controlling for stages of prolapse made the association between vaginal parity and hiatal dimensions non-significant, which is not surprising given the strong cross-correlations between vaginal parity, prolapse, and hiatal dimension; however, upon controlling for significant prolapse (ICS POP-Q stage 2 or higher), the association remained significant. In this context, it is important to point out that our study design is unsuitable for proving a direct causative link between vaginal childbirth and hiatal dimensions. We do not know whether enlarged hiatal dimensions are directly caused by vaginal childbirth (e.g. as a result of levator avulsion) or the result of prolapse caused by childbirth (e.g. as a result of fascial disruption). Either or both may be the case in any given individual. All this study is able to show is an association between vaginal childbirth and hiatal dimensions, not causation. Finally, our study population comprised women who were symptomatic of pelvic floor dysfunction only, and who were mostly white, and therefore the results may not apply to other ethnic groups and to the general population.
Whereas the aetiology of POP is likely to be complex, to date there seems little doubt that altered function and integrity of the levator ani muscle is important in the development of POP.[15, 29, 41] The findings of the current study suggest that it is mainly the first vaginal delivery that is associated with such alterations. Any efforts in preventing levator trauma and the subsequent development of POP should probably focus on the first delivery.
In this retrospective observational series of 716 symptomatic women attending a tertiary urogynaecological centre, vaginal parity was strongly associated with the area of the levator hiatus upon Valsalva. Most of this effect seems to be associated with the first vaginal delivery. Subsequent vaginal birth does not seem to have any major additional effect on hiatal dimensions.
Disclosure of interests
HPD and KLS have received an unrestricted educational grant from GE Medical. IKA and BG have no conflicts of interest to declare.
Contribution to authorship
IK: data acquisition; analysis and interpretation; drafting the article and final approval of the version to be published. BG: data acquisition; analysis and interpretation; drafting the article and final approval of the version to be published. KS: assessed repeatability series; analysis and interpretation of data; drafting the article and revising it critically and final approval of version to be published. HD: conception and design; analysis and interpretation of data; drafting the article and revising it critically for important intellectual content and final approval of the version to be published.
Details of ethics approval
This retrospective study was approved by the Institutional Ethics Committee (NBMLHD HREC ref. 12-07).
This study was not funded.
The authors would like to thank Andrew Martin, PhD, Senior Biostatistician at Sydney Medical School Nepean, University of Sydney, for his help with statistical analysis.