Objective
The objective of the study was to compare the prevalence of incontinence disorders in relation with spontaneous vaginal delivery or cesarean section.

Study Design
Two hundred women with spontaneous vaginal deliveries only were compared with 195 women with cesarean deliveries only 10 years after first delivery.

Results
When compared with cesarean section, vaginal delivery was associated with an increased frequency of stress urinary incontinence (P = .006) and an increased use of protective pads (P = .008) as well as an increased frequency of fecal urgency (P = .048) and gas incontinence (P = .01). At multivariate regression analysis, mode of delivery showed no significant association with incontinence symptoms other than an increased risk for flatus incontinence in women with a history of obstetric anal sphincter injury (odds ratio 3.1; 95% confidence interval, 1.5 to 8.9).

Conclusion
Incontinence symptoms are more common following spontaneous vaginal delivery when compared with cesarean section 10 years after first delivery. However, cesarean section is not associated with a major reduction of anal and urinary incontinence.

Key words: cesarean section; incontinence; observational; vaginal delivery

Anal and urinary incontinence is often associated with significant lifestyle alterations and functional disability in women, with severity and prevalence being closely related to age. The vast number of women afflicted by incontinence disorders is rapidly becoming a global concern with major implications on health economy, for individuals and society alike.

The effects of childbirth on the pelvic floor have been the focus of growing attention in recent years, and accumulating evidence from various experimental and observational studies suggests that vaginal delivery has a detrimental effect on the pelvic floor. Vaginal delivery may traumatize pelvic floor supportive tissues by direct laceration, muscle distension, or injury to the distal branches of the pudendal nerves. The strain on pelvic floor tissues may be further aggravated by obstetrical interventions, such as vacuum or forceps extraction, episiotomy, and fundal pressure.

It is becoming exceedingly common for women to request elective cesarean section, often based on the assumption that cesarean section protects from postpartum pelvic floor dysfunction. Some support for this notion is provided by short-term observational studies, suggesting that vaginal delivery is associated with an increased risk of incontinence disorders when compared with cesarean section. It does, however, remain uncertain whether cesarean section protects from the subsequent risk for pelvic floor dysfunction at long term and to what extent subsequent obstetrical history affects the outcome. We performed a long-term comparative study to evaluate whether having had spontaneous vaginal deliveries only is associated with an increased risk for urinary and anal incontinence symptoms when compared with having had only cesarean sections 10 years after first delivery.

Materials and Methods

We recently described the natural incidence of incontinence symptoms in a single cohort of 309 consecutive primiparous women giving birth to their first child at our department, from April through June 1995. In this 10-year prospective, observational, single-cohort study, final analysis was restricted to women giving vaginal birth to their first child in 1995, with all subsequent deliveries performed vaginally. To compare the prevalence of urinary and anal incontinence 10 years after first delivery (index delivery) in relation to delivery mode, we identified a control group of 422 consecutive primiparous women giving birth to their first child by cesarean section at our department between September 1994 and December 1995.

All subjects in the cesarean and vaginal delivery groups received an identical self-administered questionnaire on lower urinary tract and anorectal symptoms as well as written information on the objective of the study by ordinary mail. The validated Cleveland Clinic Incontinence Score was used for anal incontinence symptoms. Questions on urinary incontinence symptoms were modeled from the Cleveland Clinic Incontinence Score, using an identical ordinal frequency scale including the questions: “Do you experience sudden urges to void urine that are difficult to hold back?”; “Do you experience sudden urges to void urine that are followed by involuntary loss of urine?”; “Do you experience involuntary loss of urine at physical activities?”; “Do you experience bladder emptying difficulties?”; and “Do you need to use protective pads because of urinary incontinence?” A response of “less than once/week” was considered a mild symptom, whereas response alternatives “more than once/week” and “daily” were considered moderate to severe symptoms for both anal and urinary incontinence symptoms.

Questions on general medical history, obstetrical history, and subsequent deliveries were added to the questionnaire. Return of the questionnaire was considered informed consent, whereas patients not responding following 2 reminding letters were considered nonresponders.

Statistical analysis was performed using Statistica software (Statistica; Statsoft Inc, Tulsa, OK). The final analyses were restricted to either women with the first and all subsequent deliveries performed vaginally without forceps or vacuum extractor or women with the first and all subsequent deliveries performed by cesarean section. Mann–Whitney U test was used to compare ordinal and continuous numerical data between the independent samples.

A multivariate logistic regression analysis was used to evaluate the association between symptoms of urinary and anal incontinence 10 years after index delivery and covariates: delivery mode, age at first delivery, number of deliveries, and anal sphincter injury (in the vaginal delivery group). Obstetrical perineal tears were classified according to the International Classification of Diseases (ICD)-9 and -10 nomenclatures. The multivariate analysis also included an interaction analysis between symptoms of anal and urinary incontinence. Odds ratios were estimated with 95% confidence intervals. A P value less than .05 was considered significant for all analyses.

The study was approved by the Research Ethics Committee at Karolinska Institutet Stockholm, Sweden (institutional review board approval number 04-754/1). All patients and control subjects gave their informed consent to participate in the study after receiving written information.

Results
In the vaginal delivery group, 246 of 309 subjects (81%) responded to the questionnaire. Of the participating subjects, 232 of 246 women (94%) reported only having had vaginal deliveries. Fourteen subjects were excluded from analysis because of subsequent cesarean deliveries following the birth of the first child in 1995, and 29 subjects were excluded because of instrumental deliveries, including forceps and vacuum extraction. An additional 3 subjects were excluded from analysis because of urinary incontinence surgery, leaving 200 women eligible for participation.

In the cesarean section group, 270 of 422 subjects (64%) responded to the questionnaire, and 195 of 270 women (72%) reported only having had cesarean section deliveries following the birth of the first child by cesarean section in 1995. No subjects reported incontinence surgery, leaving 195 women eligible for participation.

Subject characteristics
Mean age at the time of the study was 41.5 years (± 4.0 SD, range 21-46) in the cesarean section group, compared with 39.9 years (± 4.6 SD, range 19-45) in the vaginal delivery group (NS). Mean age in subjects not responding to the survey was 40.1 years (± 6.2 SD). Subsequent obstetrical history 10 years after index delivery showed that median parity was 2 (range 1-4) for both groups (NS). In the vaginal delivery group, a cumulative prevalence of obstetrical perineal tears ICD grade I-II was reported by 197 patients (86%), whereas perineal rupture ICD grade III-IV was reported by 31 patients (14%). All, except 1, grade III-IV obstetrical anal sphincter injury occurred at index delivery. Episiotomy was reported by 14 subjects (6%).

Conservative treatment of urinary incontinence (including physiotherapy and pelvic floor exercises) was reported by 4 subjects in both delivery groups, whereas surgical treatment of stress urinary incontinence was reported by 3 subjects in the vaginal delivery group (1%). None of the subjects in either delivery group had undergone pelvic organ prolapse surgery 10 years after index delivery, and none reported anticholinergical overactive bladder treatment.

Symptom prevalence
With regard to lower urinary tract symptoms, there was a significantly increased frequency of stress urinary incontinence episodes (P = .01) and an increased usage of protective pads (P = .04) in the vaginal delivery group when compared with the cesarean section group. There were no significant differences in symptoms of urinary urgency or urge incontinence. When comparing anal incontinence symptoms, there was a significantly increased frequency of fecal urgency (P = .048) but no significant differences in flatus or stool incontinence.

Risk analysis

In an adjusted multivariate analysis, vaginal delivery was associated with an increased risk for stress urinary incontinence. Mode of delivery showed no other interactions at statistically significant levels with symptoms of urinary or anal incontinence 10 years after index delivery. A history of obstetrical anal sphincter tear at vaginal delivery was independently associated with an increased risk of gas incontinence (odds ratio, 3.1; 95% confidence interval, 1.5-8.9). Stress urinary incontinence was a significant predictor for an increased risk of concurrent urge urinary incontinence as well as gas and loose stool incontinence. Other covariates tried in the multivariate model, such as maternal age at first delivery and number of deliveries, showed no significant association with the development of urinary or anal incontinence symptoms.

Comment

Over the last decade, many Western countries have experienced an increase in the rate of elective cesarean sections on maternal request. The increase is mostly patient driven and often based on the notion that cesarean section protects women from postpartum sexual dysfunction and incontinence disorders. Contrary to widespread belief, the evidence supporting this assumption is limited, and controlled clinical studies assessing the impact of various delivery modes are few and mostly short-term.

We found an increased overall prevalence of symptoms associated with stress urinary incontinence in the vaginal delivery group when compared with the cesarean section group. However, differences in symptom frequency mostly consisted of symptoms in the range of mild to moderate bother. In a similar fashion, differences between the 2 groups were rather limited when considering symptoms associated with anal incontinence. At 10 years’ follow-up, only fecal urgency was significantly more prevalent in women with vaginal deliveries only. Frank fecal incontinence was uncommon, regardless of delivery mode, and in agreement with Bek and Laurberg, the predominant problem associated with anal incontinence in both groups was flatus incontinence. Again, the difference in fecal urgency mostly consisted of frequencies in the range of mild to moderate bother. Not surprisingly, women with an obstetrical history of anal sphincter injury carried a substantially increased risk for gas incontinence. This association has repeatedly been demonstrated by our and other research teams.

In the present study, approximately 10% of patients in both delivery groups had a urinary incontinence frequency score that we classified as severe. Among these subjects, a majority also experienced flatus incontinence, regardless of delivery mode, an association confirmed by the adjusted multivariate analysis. This finding suggests that urinary and anal incontinence may share a common etiology, not limited to trauma at vaginal delivery, and it seems that a full-term pregnancy by itself is associated with adverse effects on continence status, although the causal mechanism remains uncertain. Thus, double incontinence should be kept in mind when patients present with severe urinary incontinence, regardless of obstetrical history.

When considering possible predictors of outcomes in a multivariate analysis, cesarean section was associated with a decreased risk of wearing protective pads because of stress urinary incontinence 10 years after index delivery. Delivery mode showed no other significant associations with anal or urinary incontinence symptoms. The number of subjects having undergone urinary incontinence surgery was few in both groups and did not suffice for a valid statistical comparison. We were, however, able to conclude that women only having had vaginal deliveries following the birth of their first child did not show a dramatic increase in incontinence surgery rates when compared with women only delivered by cesarean section.

Strengths of our study included the consecutive population samples, a well-defined source population, the homogeneity of the groups, and the long-term follow-up from first exposure to assessment. To increase the internal validity of our observations, we excluded women with a history of both delivery modes, as well as instrumental delivery, from the final analysis. Internal validity was prioritized, even though it may have come at the cost of loss of some statistical power and also limited our ability to generalize our findings to women with both delivery modes and instrumental deliveries.

When assessing our main results, one should also consider the limitations associated with the present study design. The population sampling method used in the present study was highly efficient, but the comparison was limited to indicating an association or the lack thereof. Furthermore, the temporal ordering of associations cannot be determined in a cross-sectional design. The response rate in the present study was acceptable, yet the concern for selection bias should be addressed. We found no indication for selection bias by age because mean age in responders and nonresponders was nearly identical. It seems unlikely that nonresponders would share a common personality trait, making them less likely to respond, which at the same time acts as a confounder for delivery mode and incontinence disorders, although this possibility cannot be excluded.

It has been suggested that increasing maternal age at first delivery, and increasing number of deliveries, may cause both severity and prevalence of incontinence symptoms to increase. Although we were unable to confirm that these factors were associated with increased symptoms in either delivery group, the influence of maternal age may need to be considered in several ways. The follow-up was performed 10 years after first delivery, and all subjects in our study were assumed to be premenopausal at follow-up. Considering that the prevalence of anal and urinary incontinence is at its highest in postmenopausal women, it is possible that degenerative changes of the pelvic floor at menopause may cause symptoms to worsen. Thus, the rather minor differences in incontinence symptoms observed in the present study may change considerably following menopausal estrogen depletion, possibly for the worse in women whose pelvic floor supportive structures and nerves have been traumatized by vaginal delivery. The postmenopausal impact of delivery mode on the pelvic floor and the delayed onset of incontinence disorders are in need of further evaluation as recently highlighted.

We were unable to confirm a positive association between repeat deliveries and prevalence of urinary incontinence symptoms, regardless of delivery mode. This finding is consistent with several observational studies, suggesting that the first vaginal delivery is associated with the highest risk of loss of urethral support and urinary incontinence. When considering the association between repeat childbirth and anal incontinence, previous results are more divergent. Whereas some describe that the risk is not affected by parity, others suggest that multiple childbirth is associated with aggravated symptoms. In the present study, we found no relationship between increasing parity and prevalence of either urinary or anal incontinence symptoms.

In summation, our main findings are in concurrence with the findings of Goldberg et al, suggesting that cesarean delivery is not associated with a major reduction in long-term anal and urinary incontinence. Only having had spontaneous vaginal deliveries was associated with an increased prevalence of incontinence symptoms when compared with only having had cesarean sections 10 years after first delivery. However, severity distribution was similar between delivery groups, and only subjects with an obstetrical sphincter injury at vaginal delivery were at increased risk of flatus and fecal incontinence.

References

_{P. Miner, Economic and personal impact of fecal and urinary incontinence, Gastroenterol 126 (Suppl 1) (2004), pp. S8–S13.}

J.L. Melville, M.Y. Fan, K. Newton and D. Fenner, Fecal incontinence in U.S. women: a population-based study, Am J Obstet Gynecol 193 (2005), pp. 2071–2076.

K. Kenton and E.R. Mueller, The global burden of female pelvic floor disorders, BJU Int 98 (Suppl 1) (2006), pp. 1–7.

S.J. Snooks, M. Swash, S.E. Mathers and M.M. Henry, Effect of vaginal delivery on the pelvic floor: a 5-year follow-up, Br J Surg 77 (1990), pp. 1358–1360.

A.H. Sultan, M.A. Kamm, C.N. Hudson, J.M. Thomas and C.I. Bartram, Anal-sphincter disruption during vaginal delivery, N Engl J Med 329 (1993), pp. 1905–1911.

A.H. Sultan, M.A. Kamm and C.N. Hudson, Pudendal nerve damage during labour: prospective study before and after childbirth, Br J Obstet Gynaecol 101 (1994), pp. 22–28.

A.M. Ryhammer, S. Laurberg and A.P. Hermann, Long-term effect of vaginal deliveries on anorectal function in normal perimenopausal women, Dis Colon Rectum 39 (1996), pp. 852–859.

H.P. Dietz and A.B. Steensma, The role of childbirth in the aetiology of rectocele, BJOG 113 (2006), pp. 264–267.

K.C. Lien, B. Mooney, J.O. DeLancey and J.A. Ashton-Miller, Levator ani muscle stretch induced by simulated vaginal birth, Obstet Gynecol 103 (2004), pp. 31–40.

A.R. Smith, G.L. Hosker and D.W. Warrell, The role of pudendal nerve damage in the aetiology of genuine stress incontinence in women, Br J Obstet Gynaecol 96 (1989), pp. 29–32.

R.E. Liebling, R. Swingler, R.R. Patel, L. Verity, P.W. Soothill and D.J. Murphy, Pelvic floor morbidity up to one year after difficult instrumental delivery and cesarean section in the second stage of labor: a cohort study, Am J Obstet Gynecol 191 (2004), pp. 4–10.

Daniel Altman MD, PhD, Asa Ekstrom MD, Catharina Forsgren MD, Johan Nordenstam MD and Jan Zetterstrom MD, PhD

Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden
Division of Obstetrics and Gynecology, Department of Clinical Sciences, Danderyd Hospital, Karolinska Institutet, Stockholm, Sweden
Division of Colon and Rectal Surgery, University of Minnesota Medical School, Minneapolis, MN.
Received 7 November 2006;  revised 23 January 2007;  accepted 27 March 2007.  Available online 31 October 2007.