Cesarean section for the prevention of anal incontinence
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
In this systematic review, evidence of efficacy will be been obtained from studies that compare the prevalence of anal incontinence in women who had vaginal delivery vs. Cesarean section.
Background
There is currently a lively and voluminous discussion concerning the performance of elective primary Cesarean delivery (Abramowitz 2000, Amu 1998, Faridi 2002, Farrell 2002, Guise 2004, Lockwood 2004, Minkoff 2003, Nygaard 2003, Patterson‐Brown 1998, Zetterstrom 1999). The indications for urgent Cesarean section are governed by the health of the infant. However elective primary C. section is on the other hand undertaken with primarily maternal factors in mind. Among the most commonly mentioned factors in the above discussions are preservation of maternal anal and urinary continence by avoiding the direct trauma to the anal sphincters and distal pelvic nerves that is brought about by vaginal delivery. Though it is intuitive that C. section should be effective in this regard, an intervention such as elective C. section should require rigorous evidence of efficacy before anal continence preservation can be accepted as a principal reason for its choice.
If elective primary Cesarean section doesn't prevent anal incontinence, it doesn't have much to recommend it. There are many potential risks (Klein 2004), which include: the probability that all subsequent pregnancies will be delivered by C. section. Only 12 per cent of women with primary section deliver their next pregnancy per vaginum (Greene 2004). For these there is an increased, if small, risk of uterine rupture, endometritis, and neonatal morbidity (Landon 2004). With each subsequent C. section, the risk of operative morbidity increases (Makoha 2004). Readmission to the hospital of the mother is more likely after C. section than vaginal delivery (Liu 2005). Thromboembolism is more likely after C. section (Jacobsen 2004), women with primary elective C. section have fewer children than women who deliver vaginally (Mollison 2005), and C. section costs more than vaginal delivery (Henderson 2001).
Objectives
In this systematic review, evidence of efficacy will be been obtained from studies that compare the prevalence of anal incontinence in women who had vaginal delivery vs. Cesarean section.
Methods
Criteria for considering studies for this review
Types of studies
Studies will be sought in which women are either randomised to C. section vs vaginal delivery, or alternatively studies which compare in cohorts or in cross sectional surveys women having either C. section or vaginal delivery, in all cases for the outcome of continence after pregnancy for faeces and flatus. When possible, emergency section will be separated from elective section in the analyses, primarous pregnancies separated from women having prior pregnancy and, in multiparous women, C. section after prior vaginal delivery separated from women only having prior C. section.
Types of participants
Women with a history of pregnancy and delivery of a live infant, including breech presentations and twin pregnancies.
Types of interventions
Vaginal delivery and Cesarean section, done either electively or as an emergency.
Types of outcome measures
Incontinence of flatus and faeces
Search methods for identification of studies
Search terms will include: "Cesarean section vaginal delivery incontinence and randomised". Single institution and single practice publications will be included along with population based surveys Medline‐PubMed will be the search engine and the date range from 1966 through August, 2005. Reference lists and authors of relevant publications will also be screened for potential studies.
Data collection and analysis
Crude incidence data of anal incontinence will be presented as well as odds ratios and 95 per cent confidence intervals for individual studies. Randomized studies and non‐randomised studies will be analysed separately. In the randomised studies, if the cross over after randomisation exceeds 20% (i.e. for instance women randomised to vaginal delivery instead are treated by Cesarean section), then the data from that study will also be presented "as treated", i.e., as a non‐randomised study. In the non‐randomised studies, when odds ratios and confidence intervals adjusted for potential confounders, notably maternal age and parity, are available in published reports, these will be extracted. When these adjusted analyses are not available in the original publication, study authors will be contacted to obtain adjusted analyses or individual patient data requested to adjust their data. Meta‐analysis will be used to calculate a summary relative risk among studies using HEpiMA software (Costa‐Bouzas 2001) and graphic representation of the results of each individual study will be shown using the Revman Analysis software. Three criteria of study quality were developed by review of the included studies. Maternal age was found in most studies to be the most significant predictor of post partum anal incontinence, so adjustment for maternal age was one criterion. Longitudinal assessment of continence in the post partum period demonstrated in several studies that premature assessment of anal incontinence had no correlation with long term dysfunction. Lastly some women having had Cesarean section may have had previous vaginal delivery, and so be mis classified as C. section in the analysis, when their anal dysfunction might have been due to the original vaginal delivery. Therefore sensitivity analyses will be conducted excluding studies:
1 Without age adjustment,
2 In which continence was assessed 4 months or less post partum,
3 In which women having had Cesarean section had a prior vaginal delivery, i.e., including in this case either primipara mothers only or, in the C. section group, women whose prior pregnancies had only been Cesarean deliveries.
Statistical heterogeneity will be calculated using chi square and the Q statistic with the HEpiMA software. The number needed to treat will also be presented, that is, the number of C. sections which must be done to prevent a single case of incontinence of either gas or stool. The random effects model will be used in the meta‐analysis if heterogeneity is encountered.
