Introduction 2 4 5 6 9 1 10 11 12 15 16 Materials and methods One hundred and seventeen women who attended the outpatient clinics of the University Hospital Groningen and the Martini Hospital Groningen enrolled for the study, with a mean age of 30 years (range 17–41). All women were of Caucasian origin, except three who were of Mediterranean origin. All women were nulliparous and had no history of incontinence, pelvic operations or neurological disease. Written informed consent was obtained from all participating women. The study was approved by the medical ethical committees of both hospitals. For this study, the women were investigated at 28–32 and at 36–38 weeks of pregnancy and 6 weeks and 6 months post-partum. 17 11 Statistical analysis Pad test results have a non-parametric distribution. For continuous variables, non-parametric tests are used. Numeric data are analysed by cross tabulation, chi-squared test and risk analysis. Pearson correlation test was used to identify significant relationships between variables. Data are presented as median or numbers. Results Pregnancy, pad test 1 Fig. 1 Distribution of 24-h pad test results at 28 and 38 weeks of pregnancy  r p r p Pregnancy, questionnaire r p r p Pregnancy, pad test and questionnaire p 2 p 2 Fig. 2 a b  Puerperium, pad test 3 r p r p Fig. 3 Distribution of 24-h pad test results at 6 weeks and 6 months post-partum  Puerperium, questionnaire r p r p  Puerperium, pad test and questionnaire p p p 4 p 4 Fig. 4 a b  Discussion 1 18 16 19 20 It is remarkable that the group of controls as referred to had similar pad test results but did not report incontinence, whereas in our pregnant group, 30 (28 weeks of pregnancy) and 35% (38 weeks of pregnancy) of the women did. With the same weight gain in pregnancy, women report more incontinence than in the sample of non-pregnant women. It seems therefore that not the amount of weight gain in the pad test but the pregnancy state itself is more discriminating for the chance that a woman qualifies herself as incontinent. 21 22 The calculation of the diagnostic strength of our 24-h pad test was made with the self-reporting of symptoms of urinary incontinence as the gold standard. At 28 weeks, the pad test failed to capture eight subjects who stated they were wet; at 38 weeks, this was higher. Such results are possibly related to the high threshold for definition of incontinence in the women with some leaking less than 9 g describing some leakage. The rationale for using a high cutoff is established in both men and women, but subjects themselves may perceive this a severe incontinence. It is possible that pregnant and post-delivery women perceive leakage differently than their non-pregnant counterparts. When adding severity of symptoms to the gold standard, as reported by VAS, the diagnostic strength of the pad test did not improve. 23 From our study, we conclude that pad testing measures fluid loss over a certain period but does not quantify self-reported symptoms of urinary incontinence. Both measurements are of interest but cannot replace each other. Stressing of the pelvic floor by pregnancy and childbirth modulates the sensation of urinary leakage in such a way that women in this state do report symptoms of urinary incontinence more frequently than nulliparous pre-menopausal women do.