Gyhagen et al. reported on the prevalence of urinary incontinence (UI) after 20 years of vaginal or caesarean delivery in BJOG (January 2013), which has major implications in the management of pregnant women.
In the title it would be better to state ‘a national cohort’ rather than ‘a national cohort study’ to avoid confusion regarding the study design.
Prevalence data of UI in relation to maternal age at delivery and current body mass index (BMI) have been reported in the article. The current age may also be looked into, as increasing age has a strong relationship with UI. The current BMI was the most important BMI-determinant for UI in this study; the scientific justification for the causal relationship between UI developing as a result of vaginal delivery and the current BMI may be provided. The data on maternal weight at delivery was retrieved from the Swedish Medical Birth Register, but the method of obtaining the current weight and height of the participants was not described. If it is by the questionnaire filled in by the respondents, then the standardisation procedure to ensure the reliability of these measurements may be described. This is integral, as the most important finding was the association between current BMI and UI.
The severity of UI was assessed using the Sandvik score, but this has not been presented in the results, which would provide useful information on the severity of UI associated with vaginal delivery versus caesarean section.
Four variables were found to be statistically significant in the logistic regression analysis, as presented in table 4 (mode of delivery, BMI current, BMI at delivery, maternal age at delivery). The comparison of these variables and their crude and adjusted odds ratios in the vaginal and caesarean groups is presented in table 5, but in table 6 the comparison of these variables is shown within the vaginal group and within the caesarean group only. The confounding variables should have been compared with the independent variable (vaginal and caesarean groups) to understand the net effect of the independent variable on UI.
From table 4, the birthweight and head circumference of the infant were not statistically significant. The reasons for including these two factors in further analysis (tables 5 and 6) is not clear. Furthermore, the infant birthweight was first categorised into four groups, and significant association was found with UI in all the groups except <3000 g. The infant birthweight was further re-categorised into two groups <4500 g and >4500 g, which did not show any association with UI; the reasons for this categorisation and re-categorisation would provide more clarity to the readers.
The recommendation from the study is to perform eight or nine caesarean sections to avoid one case of UI. This study finding alone may not substantiate this claim, as both groups showed a prevalence of UI, with a greater prevalence in the group with vaginal delivery. Secondly, the risks associated with caesarean section need to be considered. Finally, the influence of current age on UI may also be a risk factor involved.
Disclosure of interest
We declare that we have no conflicts of interest.