To estimate the risk for stress urinary incontinence and pelvic organ prolapse surgery related to vaginal birth or cesarean delivery.
A cohort study of all women having their first and all subsequent deliveries by cesarean (n = 33,167), and an age-matched sample of women only having vaginal deliveries (n = 63,229) between 1973 and 1983. Hazard ratios were calculated using Cox regression models with 95% confidence intervals.
Women only having vaginal deliveries had increased overall risks of incontinence (hazard ratio, 2.9; 95% confidence interval, 2.4–3.6) and prolapse surgery (hazard ratio, 9.2; 95% confidence interval, 7.0–12.1) compared with women only having cesarean deliveries.
Having only vaginal childbirths was associated with a significantly increased risk of stress urinary incontinence and pelvic organ prolapse surgery later in life compared with only having cesarean deliveries.
Almost one-fifth of US women experience symptoms of urinary incontinence or pelvic organ prolapse (POP). Surgery remains the mainstay of treatment for both disorders, and regional US studies estimate that the life time risk of having surgery for stress urinary incontinence (SUI) or POP is 11%. Childbirth is widely considered an established risk factor for both SUI and POP and women rarely enter pregnancy with preexisting symptoms of either of these diseases. In 2006, one third of all US childbirths were performed by cesarean deliveries, and an increasing proportion of cesarean deliveries are performed on maternal request ( www.hcup-us.ahrq.gov/reports/statbriefs/sb71.pdf ). The reason for the rising incidence of elective cesarean deliveries performed on maternal request is multifactorial, but may to some extent be driven by women’s apprehension for pelvic floor sequela after vaginal delivery.
Vaginal childbirth may induce functional disorders of the lower urinary tract and pelvic floor as a consequence of delivery trauma to pelvic organ neuromuscular function and morphology. Although little is known about the long-term effects of cesarean delivery on POP, several epidemiologic studies suggest that cesarean delivery significantly decreases the risk for postpartum urinary incontinence. However, prevention of disorders later in life by cesarean delivery is controversial. Cesarean delivery involves major surgery and studies suggest that the protective effects of cesarean delivery on urinary incontinence diminishes over time and even disappears after multiple deliveries.
Whether cesarean delivery provides a long-lasting protection against SUI or POP surgery must be considered an unresolved issue. The aim of this population-based cohort study was to compare long-term effects of vaginal vs cesarean delivery with respect to risks of urinary incontinence and POP surgery.
Materials and Methods
The nationwide Swedish Medical Birth Register, kept by the National Board of Health and Welfare, contains prenatal, obstetric, and neonatal data from almost 99% of all women giving birth in Swedish hospitals from 1973 and thereafter ( www.socialstyrelsen.se/register/halsodataregister/medicinskafodelseregistret/inenglish ). Within the Medical Birth Register, we initially identified all women who gave birth to their first child by cesarean section and thereafter (if multiparous) only gave birth by cesarean section from January 1973 through December 1982. To each woman in this cesarean delivery cohort, we randomly selected 2 control women, individually matched by year of birth, who gave birth to their first child by vaginal delivery at the same calendar year with all subsequent deliveries performed by vaginal birth only (if multiparous).
Using individually unique national registration numbers, assigned to all Swedish residents at birth or immigration, we linked both cohorts to the nationwide Inpatient and Cause of Death Registers. The Swedish Inpatient Register contains data for individual hospital discharges. The register was established in 1964 and covered in 1973 more than 60% of the population, in 1983 coverage was 85%, and from 1987 complete national coverage was achieved. Every inpatient discharge record contains: (1) dates of hospital admission and discharge; (2) up to 8 discharge diagnoses, coded according to the International Classification of Diseases (ICD-8 from 1973 to 1986, ICD-9 from 1987 to 1996, and ICD-10 thereafter) ; and (3) up to 12 operation codes from the Swedish Classification of Operations and Major Procedures. Correct coding for surgical procedures is achieved in 98% of records in the register, with a less than 1% yearly loss to registration.
From the Inpatient Register, we identified all women in the vaginal delivery and cesarean cohorts having had SUI or POP surgery. Women contributed person-time to the study from Jan. 1 the calendar year after their first delivery until the first occurrence of SUI surgery, POP surgery, death, or the end of the observation period (Dec. 31, 2004).
During the observational period, 33,167 women were delivered by cesarean section at their first and all subsequent births and were individually matched to 63,229 women with vaginal childbirth at their first and all subsequent deliveries. We excluded 274 women with diagnosis of both instrumental vaginal delivery and cesarean section at the same delivery and 2060 women with multiple births. We further excluded 6 women where SUI or POP surgery preceded or were performed in the same year as first birth and 36 women who died before Dec. 31 of the year of first birth, as well as, 3018 women who lacked control subjects after exclusions were performed. After exclusions, 30,880 women in the cesarean cohort and 60,122 women in the vaginal delivery cohort were eligible for analyses.
Classification of exposures and outcomes
We used codes from the Swedish Classification of Operations and Major Procedures for 1973-1996 and 1997-2005 to identify SUI and POP surgery in the Swedish Inpatient Register. Surgical procedures for SUI included: Kelly sutures, Kennedy sutures, abdominal and laparoscopic Burch colposuspensions, Stamey procedures, Ingelman-Sundberg plasty, intravaginal slingplasty, suburethral slings, and tension-free vaginal tapes (operation codes 6355, 6356, 6358, 7470, 7471 for 1973-1996, and LEG00, LEG10, LEG20, LEG96 and KDG10-40 for 1997-2004). POP surgery was categorized as: anterior repair, posterior repair, Manchester procedure, abdominal sacrocolpopexy, vaginal sacrospinous fixation, abdominal and vaginal enterocele obliteration, colpocleisis, and hysterectomy (operation codes 4840, 4841, 4844, 4935, 7120, 7121, 7460-7464, 7466, 7469 and 7541 for 1973-1996 and LEF00, LEF03, LEF10-50, LEF53, LEF96, LFE20, JHB 40, JHB96, JFJ00, JGC00, JGC01 for 1997-2004).
We retrieved information on potential confounders and effect modifiers including: age at first and last birth, parity, pregestational and gestational diabetes, instrumental delivery, gestational age, infant’s weight, and head circumference at birth, from the Medical Birth Register. From the Swedish Inpatient Register, we derived time to surgery, and from the Cause of Death Register, we obtained information on deaths occurring before the end of the observational period.
Crude and parity-specific incidence rates for SUI and POP surgery were calculated as the number of events per 10,000 person years with 95% confidence intervals (CIs) based on the Poisson distribution. The risk of SUI and POP surgery related to delivery mode and parity was estimated using Cox proportional-hazards models, adjusted for maternal age and obstetric covariates. Only first surgery for either prolapse or incontinence was included in the analysis after which patients were censored if secondary procedures were performed. Parity was categorized as 1, 2, or 3 and more, childbirths and maternal age at first delivery was categorized in 5-year bands. Possible confounders and effect modifiers were dichotomized (yes/no) in, ever given birth to a child with: a birthweight ≥4 kg, head circumference ≥38 cm, gestational week ≥40 pregnancy, or ever having a recording of pregestational and gestational diabetes. Effect modification of variables was evaluated by type 3-test before being entered in the Cox proportional hazard model. The results of the different Cox’s regression hazard models are presented as hazard ratios (HRs) with 95% CIs.
To address potential confounding of instrumental vaginal delivery, subanalyses were performed with risk estimates for women who have ever delivered by vacuum extraction or forceps, presented separately. Parity-specific risks were estimated through stratification. Effects of delivery mode in relation to time from first delivery until pelvic floor surgery are presented as incidence rates. As a secondary outcome measure, we calculated the number needed to harm (NNH). NNH is an epidemiologic measure that estimates how many patients need to be exposed to a specific risk-factor during a period to cause harm in 1 patient that would not otherwise have been harmed during that period.
Statistical analyses were undertaken with SAS software (version 9.1; SAS Institute, Cary, NC) and performed by the principal investigator in cooperation with a biostatistician. The study was approved by the Research Ethics Committee at Karolinska Institutet, Stockholm, Sweden, and conforms to the STROBE guidelines for reporting observational studies.