Although increasing age and vaginal delivery are recognized as the major inciting factors,1–7 the etiology of pelvic organ prolapse (POP) remains poorly understood. Important individual variability in the predisposition to develop prolapse has been postulated.8,9 The role of lifestyle-related environmental factors is less well defined.10 Bump and Norton divided putative risk factors into three categories: predisposing (family history, race and ethnicity, variations of anatomy and collagen synthesis and structure), inciting (vaginal delivery through its effect on nerve damage, muscle damage, and tissue disruption), and promoting (other possible risk factors including lifestyle),10 but few studies address the relative contributions of these main types of risk factors.
The objective of this population-based, cross-sectional study was to quantify links between tentatively implicated nonobstetric factors and symptomatic prolapse. A detailed analysis of obstetric risk factors has been published elsewhere.6
MATERIALS AND METHODS
The study started in 2001 with a survey investigation among 8,000 female residents of Stockholm, 30 to 79 years of age, who were selected randomly from the computerized and continuously updated Swedish Population Register (Fig. 1). A validated 5-item questionnaire was mailed to all women selected.11 At the end of the data-collection period (October 2002), the response rate was 69% (n=5,489). Of these women, 454 gave self-reports that were consistent with POP, henceforth referred to as cases of symptomatic prolapse. Thus the prevalence of symptomatic prolapse was 8.3%.12 To validate the presumed POP status, we invited 206 randomly selected women from the 454 with symptomatic prolapse and 206 women randomly selected from the 5,035 without symptomatic prolapse to undergo gynecological examinations according to pelvic organ prolapse quantification (POP-Q). These examinations were carried out between February 2002 and March 2003 and were performed by gynecologists blinded to the questionnaire responses; 162 and 120 invitees, respectively, came for examination. In this nonconsulting female population sample, applying POP-Q as the gold standard, the questionnaire had a specificity of 94.2% and a sensitivity of 66.5%.11
To investigate possible risk factors for symptomatic prolapse, we mailed a 72-item questionnaire to all 454 patients classified in the initial survey investigation as having symptomatic prolapse (of whom 162 had undergone gynecological examination in the validation substudy) and to the 405 control participants. The latter group, which included the 206 women selected for the validation substudy, had been selected randomly among the 5,035 women whose responses in the initial survey investigation did not meet our criteria for symptomatic prolapse. With this sample size and an exposure prevalence of 10% and 20%, we would be able to detect relative risks of 1.77 and 1.57, respectively (significance level .05, power 0.80, two-sided test).
In addition to obstetric history, the 72-item questionnaire included questions about weight, height, hip and waist circumference, and previous weight changes. Questions about previous medical history covered gynecological operations, hypertension, asthma, chronic cough, and any conditions suggestive of deficient connective tissues (varicose veins, hernia, and hemorrhoids). Current drug history included the use of menopausal hormone therapy, bulking agents, and laxatives. We asked whether the interviewees considered themselves to have regular bowel habits (yes or no) and about frequency/week of bowel emptying. Constipation was defined as fewer than three bowel movements per week. Participants responded yes often, sometimes, occasionally, or never to the following questions: Bothersome hard/lumpy stool? Difficult evacuation? Pain at defecation? Gas incontinence? Fecal incontinence?
We asked about POP, pelvic organ surgery, and any conditions suggestive of deficient connective tissues in mother or sisters. Questions also were asked about the regularity of meals and about how often (daily, once to several times/week, or seldom) the participants consumed meat, fresh fruit, root vegetables, and whole-grain bread.
Moreover, we inquired about smoking status (never, former, and current), smoking dose (number of cigarettes per day), and duration of the habit (years). In addition, information was collected about the number of years of formal education (obligatory 9-year grade school or less, high school, or university) and current or previous occupations that involved heavy lifting (up to 10 kg daily, more than 10 kg) or exposure to vibrations.
We also inquired about physical activities such as walking or bicycling to work and regular, organized exercise (hours/week). The type of current exercise was divided into high-impact (running, aerobics, weightlifting, tennis, other ball sports) and low-impact (golf, swimming, horseback riding, bicycling, walking, light physical exercises). Finally, the participants were asked about the frequency of pelvic floor muscle training, if any, at time of interview as well as 10 and 20 years previously. The data collection about possible risk factors was concluded in August 2003.
We restricted our study to women with intact uteri and no prior surgeries for prolapse or incontinence. We postulated that symptoms and recall might differ among women who have undergone pelvic surgery.
Possible links between the presence of symptomatic prolapse and the various factors first were studied in simple logistic regression models with and without adjustments for age and parity, using the prevalence odds ratio (OR) and 95% confidence intervals (CIs) as the measure of association. With a symptomatic prolapse prevalence of 8.3%, we considered the rare disease assumption to be reasonably valid; prevalence ORs for moderately strong associations therefore can be considered to be acceptable approximations of relative risks. A stepwise multivariable logistic regression analysis then was performed on the factors identified in trivariable analyses. At the end, the variables that were dropped in the stepwise approach were tested again in the resulting model. All the statistical tests performed were two-sided at the 5% significance level. No adjustments for multiple testing were made.
This study was approved by the ethics committee of Karolinska Institutet. Patients received written information before enrollment.
A total of 655 (76.2%) women returned completed questionnaires. Ninety-seven women were excluded owing to previous hysterectomy, incontinence surgery, or prolapse surgery. Thus, 558 women were included in this analysis—273 with symptomatic prolapse and 285 control participants. The mean age of the control participants was lower compared with the women with symptomatic prolapse (49.1, range 29–79, SD 13.5 compared with 53.3, range 29–79, SD 12.3) as was the mean parity (1.4, median 1, range 0–5 compared with 2.1, median 1, range 0–5). The cesarean delivery rate was low in both groups (8.6% and 13.0%). No difference in the use of hormone therapy was noted between those with symptomatic prolapse and control participants (28.0% and 29.4%).
Table 1 shows the prevalence ORs for symptomatic prolapse by factors under consideration, obtained in simple age and parity models. The prevalence of symptomatic prolapse was higher in women whose mothers or sisters had undergone surgery for prolapse compared with women without such family history (prevalence OR 3.1, 95% CI 1.7–5.5).
A history of conditions suggestive of deficient connective tissues (varicose veins, hernia, hemorrhoids) was associated significantly with symptomatic prolapse (prevalence OR 2.0, 95% CI 1.4–3.0).
A body mass index (BMI) of 25 or higher as well as a waist circumference of 88 cm or more also was found to be significantly associated with symptomatic prolapse, but asthma, chronic cough, and self-reported hypertension were not. Current or previous smoking status was not associated with symptomatic prolapse, nor was the number of cigarettes smoked. However, the odds of having symptomatic prolapse were found to fall with increasing duration of the smoking habit. Although level of education showed no relation to symptomatic prolapse, heavy lifting at work was found to be positively linked to this condition in the simple age- and parity-adjusted models. Participants currently engaged in physical activity with high impact on the pelvic floor (running, jumping, and ball sports) had a lower prevalence of symptomatic prolapse than those reporting low-impact exercise (walking, swimming, and light physical exercise). Age at menopause, weight gain, weight loss, the ratio of waist/hip circumference, and occupational exposure to vibrations were not significantly associated with symptomatic prolapse (data not shown).
All of the bowel symptoms considered—constipation, problems with hard/lumpy stool, difficult evacuation, pain at defecation, gas, and fecal incontinence—showed significant associations with symptomatic prolapse (Table 2), whereas this was not the case for any of the investigated aspects of dietary intake or meal habits (data not shown).
The final multivariable logistic regression model is shown in Table 3. We considered all the variables that were significantly linked to symptomatic prolapse in the simple age- and parity-adjusted models, but we did not consider incontinence of gas or feces because these symptoms are more likely to be part of pelvic floor dysfunction rather than a cause of prolapse. Instead, we considered level of education because this factor was a priori perceived as potentially linked to the inclination to report prolapse symptoms. Furthermore, we considered age at menopause, weight gain (less than 5 kg or 5 kg or more since age 20), weight loss (less than 5 kg or 5 kg or more), the ratio of waist/hip circumference, and asthma/chronic cough—factors for which there were strong a priori hypotheses.
Ultimately, eight variables turned out to be significant. Age, parity, and family history of prolapse were the dominating risk factors, although the CIs for prevalence ORs were wide. The prevalence OR point estimate for age had an apparent threshold of 50–59 years, whereas for parity the odds for symptomatic prolapse tended to increase with increasing parity. This tendency also was observed in relation to increased BMI, although the CI included 1 for underweight and obesity, the two categories with the lowest number of participants. The presence of any condition suggestive of deficient connective tissue was significantly associated with symptomatic prolapse, as was heavy lifting at work, although no clear dose-response was confirmed. Smoking duration remained significantly and inversely related to symptomatic prolapse (prevalence OR 0.98 per year of smoking, 95% CI 0.96–0.99). Presence of any abnormal bowel habits conferred twofold increased odds.
When we excluded the variable abnormal bowel habits from the model, assuming that it might be a consequence rather than a risk factor, it did not materially change the results concerning the other covariates (data not shown), nor did a restriction to women without a family history of prolapse (data not shown).
We repeated our analyses, restricting them to parous women only (n=365). Our results essentially were unaltered by this restriction (data not shown).
The association of symptomatic prolapse with various aspects of pelvic floor muscle training were analyzed, having been restricted to women who had a history of childbirth before the respective exposure windows (at time of interview, 10 years before interview, and 20 years before interview) because we assumed that advice about pelvic muscle training had been given only to parous women. Although the point estimates of prevalence OR were below 1 among the women with the highest self-reported frequency of pelvic floor muscle training, none of these variables attained statistical significance and there were no clear dose-response patterns (Table 4).
Although previous studies have demonstrated convincingly the significance of vaginal delivery as the dominant risk factor for POP,13 the results from the present study indicate that BMI, conditions suggestive of deficient connective tissues (varicose veins, hernia, and hemorrhoids), and heavy lifting at work also are associated independently with the presence of symptomatic prolapse. In line with previously published data, a family history of prolapse also seemed to be a risk factor.8,9,14,15 These observations support the hypothesis that certain women are genetically more predisposed to the condition. Whereas the observed familial concordance of prolapse may have been inflated owing to information bias, the association between symptomatic prolapse and a history indicative of deficient connective tissue is less likely to be explained by such a bias, particularly because most of the affected women had never seen a doctor for their symptoms and had never been given a prolapse diagnosis.
The association with high BMI receives ample support in the previous literature,2,7,16–18 but published data do still conflict.3,5,8,9,15,19
Heavy lifting at work commonly is cited as a risk factor for the development of POP, and this association was confirmed in our study. We tried to specify heavy lifting in terms of regularly lifted weight over a time period exceeding 6 months, but we did not ask for specific job descriptions. The same hypothetical role of repetitive heavy lifting was suggested in studies in which women in certain occupations were found to have a higher risk compared with women in professional employment.8,20,21 None of these studies, however, including our own, can exclude credibly the possibility that heavy lifting does not cause anatomic prolapses but only makes them symptomatic enough to become a clinical problem. Moreover, because aggravation by heavy lifting was one of the criteria used for our case definition, our observed association might well be an artifact.
A long history of smoking was linked to a decreased prevalence of symptomatic POP in our study and is also noted by others.22 It has been proposed that this inverse association may be mediated by decreased BMI,23 but the apparent protection remained after adjustment for BMI, which suggests an independent effect of smoking. The mechanisms of this counterintuitive association are unknown but might well be indirect and related to the effects of smoking-related lifestyle factors.
Although essentially unconfirmed in previous literature, asthma/chronic cough has been proposed to increase the risk for POP9 based on the intuitive link to increased intraabdominal pressure. In our simple adjusted analyses, asthma/chronic cough was linked to a moderate but statistically nonsignificant increase in the odds of having a symptomatic prolapse. This tentative association was further attenuated in our multivariable model. Hence, our data provide little support for the significant role of asthma/chronic cough in the etiology of symptomatic prolapse.
Clinical studies investigating the associations of POP with physical activity are scarce.24 Women engaged in high-impact physical training in our cross-sectional study had, paradoxically, lower prevalence compared with women engaged in low-impact training, whereas women with no organized physical training also tended to demonstrate lower prevalence than the latter group. The direction of any causality (high-impact training protecting against prolapse or absence of prolapse permitting high-impact training) cannot be determined in our data.
Concerning pelvic floor muscle training, available evidence has been deemed insufficient to guide practice.25 The participants in our study were asked to state how often they exercised rather than to record the actual time. We chose to analyze pelvic floor muscle training only among parous women because such training is recommended routinely at maternity clinics but is not promoted among nonparous women, and we believed that the motivation among the latter group was low. Although there is some risk of information bias and reversed causation, our data suggest that pelvic floor muscle training does not reduce the risk of symptomatic prolapse importantly. Further evidence from larger controlled trials is, however, necessary.
We noted statistically significant associations of symptomatic prolapse with all the investigated bowel symptoms. However, the establishment of the direction of causality, if any, is not straightforward and is particularly difficult in our cross-sectional study. In previous research, chronic constipation13,26,27 and straining at defecation27,28 were both found to be associated with the presence of POP, but only weakly with respect to stage and severity.27,29
The principal strength of this study was its basis in a population-based sample obtained via the continuously updated computerized population register. Most previous studies used patient populations with acknowledged prolapse, but we selected participants regardless of whether they had consulted health care services or not. In addition, a large proportion of the women in our study were examined according to POP-Q. The estimated prevalence of symptomatic prolapse was consistent with the findings from other countries.16,30
Among the limitations, misclassification of prolapse status is the most problematic. Given the sensitivity of the initial questionnaire vis-à-vis an anatomical prolapse of 66%, a specificity of 94%, and a prevalence of POP of approximately 10%, the positive predictive value of the questionnaire was around 60% and the negative predictive value was 97%. This implies it is likely that approximately 40% of the women classified as having symptomatic prolapse did not have an anatomic prolapse. On the other hand among the women classified as being without symptomatic prolapse, no more than approximately 3% were, in fact, expected to have an anatomic prolapse. The effect of this misclassification would be an underestimation of the associations with risk factors; observed associations could not be explained conceivably by this misclassification, but a lack of associations must be interpreted cautiously. Furthermore, the cross-sectional character of this study makes it difficult to infer the causal direction of observed associations.
Moreover, it cannot be predicted whether or not the pelvic support defects in this population will advance to more severe prolapse, and, for this reason, the results of the study cannot be generalized readily to apply to a prolapse of such severity that it necessitates surgical repair. Risk factors may differ depending on the definition of prolapse.19 The restriction to women without hysterectomy and no prior POP surgery may further challenge the generalizability, particularly with respect to populations with a high prevalence of hysterectomy. However, the exclusion of women who had undergone pelvic surgery was deemed to make associations clearer and more interpretable from an etiological perspective.
In conclusion, with reservations for the possible attenuation of associations due to misclassifications of prolapse status and the problems of establishing the direction of causality in cross-sectional investigations, this study supports the existence of some genetic predisposition but also the influence of modifiable risk factors such as BMI and possibly heavy lifting and disturbed bowel habits. Prospective cohort studies may shed further light on the sequences of occurrence, but large-scale randomized intervention trials remain the only way to establish causality.
1. Olsen AL, Smith VJ, Bergstrom JO, Colling JC, Clark AL. Epidemiology of surgically managed pelvic organ prolapse and urinary incontinence. Obstet Gynecol 1997;89:501–6.
2. Mant J, Painter R, Vessey M. Epidemiology of genital prolapse: observations from the Oxford Family Planning Association Study. Br J Obstet Gynaecol 1997;104:579–85.
3. Samuelsson EC, Victor FT, Tibblin G, Svärdsudd KF. Signs of genital prolapse in a Swedish population of women 20 to 59 years of age and possible related factors. Am J Obstet Gynecol 1999;180:299–305.
4. Swift SE. The distribution of pelvic organ support in a population of female subjects seen for routine gynecologic health care. Am J Obstet Gynecol 2000;183:277–85.
5. Swift SE, Pound T, Dias JK. Case-control study of etiologic factors in the development of severe pelvic organ prolapse. Int Urogynecol J Pelvic Floor Dysfunct 2001;12:187–92.
6. Tegerstedt G, Miedel A, Maehle-Schmidt M, Nyrén O, Hammarström M. Obstetric risk factors for symptomatic prolapse: a population-based approach. Am J Obstet Gynecol 2006;194:75–81.
7. Jelovsek JE, Maher C, Barber MD. Pelvic organ prolapse. Lancet 2007;369:1027–38.
8. Chiaffarino F, Chatenoud L, Dindelli M, Meschia M, Buonaguidi A, Amicarelli F, et al. Reproductive factors, family history, occupation and risk of urogenital prolapse. Eur J Obstet Gynecol Reprod Biol 1999;82:63–7.
9. Rinne KM, Kirkinen PP. What predisposes young women to genital prolapse? Eur J Obstet Gynecol Reprod Biol 1999;84:23–5.
10. Bump RC, Norton PA. Epidemiology and natural history of pelvic floor dysfunction. Obstet Gynecol Clin North Am 1998;25:723–46.
11. Tegerstedt G, Miedel A, Maehle-Schmidt M, Nyren O, Hammarström M. A short-form questionnaire identified genital organ prolapse. J Clin Epidemiol 2005;58:41–6.
12. Tegerstedt G, Maehle-Schmidt M, Nyrén O, Hammarström M. Prevalence of symptomatic pelvic organ prolapse in a Swedish population. Int Urogynecol J Pelvic Floor Dysfunct 2005;16:497–503.
13. Rortveit G, Brown JS, Thom DH, Van Den Eeden SK, Creasman JM, Subak LL. Symptomatic pelvic organ prolapse: prevalence and risk factors in a population-based, racially diverse cohort. Obstet Gynecol 2007;109:1396–403.
14. Jack GS, Nikolova G, Vilain E, Raz S, Rodriguez LV. Familial transmission of genitovaginal prolapse. Int Urogynecol J Pelvic Floor Dysfunct 2006;17:498–501.
15. Buchsbaum GM, Duecy EE, Kerr LA, Huang LS, Perevich M, Guzick DS. Pelvic organ prolapse in nulliparous women and their parous sisters. Obstet Gynecol 2006;108:1388–93.
16. Risk factors for genital prolapse in non-hysterectomized women around menopause. Results from a large cross-sectional study in menopausal clinics in Italy. Progetto Menopausa Italia Study Group. Eur J Obstet Gynecol Reprod Biol 2000;93:135–40.
17. Hendrix SL, Clark A, Nygaard I, Aragaki A, Barnabei V, McTiernan A. Pelvic organ prolapse in the Women’s Health Initiative: gravity and gravidity. Am J Obstet Gynecol 2002;186:1160–6.
18. Moalli PA, Jones Ivy S, Meyn LA, Zyczynski HM. Risk factors associated with pelvic floor disorders in women undergoing surgical repair. Obstet Gynecol 2003;101:869–74.
19. Nygaard I, Bradley C, Brandt D, Women’s Health Initiative. Pelvic organ prolapse in older women: prevalence and risk factors. Obstet Gynecol 2004;104:489–97.
20. Jørgensen S, Hein HO, Gyntelberg F. Heavy lifting at work and risk of genital prolapse and herniated lumbar disc in assistant nurses. Occup Med (Lond) 1994;44:47–9.
21. Woodman PJ, Swift SE, O’Boyle AL, Valley MT, Bland DR, Kahn MA, et al. Prevalence of severe pelvic organ prolapse in relation to job description and socioeconomic status: a multicenter cross-sectional study. Int Urogynecol J Pelvic Floor Dysfunct 2006;17:340–5.
22. Bradley CS, Zimmerman MB, Qi Y, Nygaard IE. Natural history of pelvic organ prolapse in postmenopausal women. Obstet Gynecol 2007;109:848–54.
23. Molarius A, Seidell JC, Kuulasmaa K, Dobson AJ, Sans S. Smoking and relative body weight: an international perspective from the WHO MONICA Project. J Epidemiol Community Health 1997;51:252–60.
24. Larsen WI, Yavorek T. Pelvic prolapse and urinary incontinence in nulliparous college women in relation to paratrooper training. Int Urogynecol J Pelvic Floor Dysfunct 2007;18:769–71.
25. Hagen S, Stark D, Maher C, Adams E. Conservative management of pelvic organ prolapse in women. The Cochrane Database of Systematic Reviews 2006, Issue 4. Art. No.: CD003882. DOI: 10.1002/14651858.CD003882.pub3.
26. Arya LA, Novi JM, Shaunik A, Morgan MA, Bradley CS. Pelvic organ prolapse, constipation, and dietary fiber intake in women: a case-control study. Am J Obstet Gynecol 2005;192:1687–91.
27. Spence-Jones C, Kamm MA, Henry MM, Hudson CN. Bowel dysfunction: a pathogenic factor in uterovaginal prolapse and urinary stress incontinence. Br J Obstet Gynaecol 1994;101:147–52.
28. Kahn MA, Breitkopf CR, Valley MT, Woodman PJ, O’Boyle AL, Bland DI, et al. Pelvic Organ Support Study (POSST) and bowel symptoms: straining at stool is associated with perineal and anterior vaginal descent in a general gynecologic population. Am J Obstet Gynecol 2005;192:1516–22.
29. Weber AM, Walters MD, Ballard LA, Booher DL, Piedmonte MR. Posterior vaginal prolapse and bowel function. Am J Obstet Gynecol 1998;179:1446–9.
30. Nygaard I, Barber MD, Burgio KL, Kenton K, Meikle S, Schaffer J, et al. Prevalence of symptomatic pelvic floor disorders in US women. JAMA 2008;300:1311–6.
Figure. No caption available.