been associated with increased OASIS at the time of birth and at 3 months after birth in other studies. 5,6 Kalis et al. 1 also state that we extrapolate evidence from studies investigating angle shrinkage of mediolateral episiotomies, and use this to support our findings concerning healed lateral episiotomy. We are reluctant to classify the episiotomies in our study as lateral. The tradition in our two units is mediolateral episiotomy. Episiotomy incision lateral to the midline has only recently been advocated. Also, Tincello et al. 7 have shown that one-third of UK obstetricians and midwives start the episiotomy lateral to the midline. Therefore, we included the point of episiotomy incision, d, in the figure, believing that this would mirror episiotomy techniques more accurately. Consistent with earlier reports, we found that the episiotomies actually performed varied greatly from theoretical recommendations, with 'mediolateral episiotomies' being in practice anything from lateral to medial. 5,6 This discrepancy between technique and theory was one of the main reasons for our research.Kalis et al. 1 also suggest that data on the length from the end of the episiotomy to the midline, b, are of limited clinical value. We disagree. In clinical practice the posterior fourchette and the anal opening are the two points of reference, and d, length (e) and angle (a) are independent variables. As b is a function of theses variables, and can be approximately estimated visually from the two reference points and the independent variables, b might be of great clinical value if future studies indicate an association between b and OASIS. With an angle of 45°one will need an episiotomy with a longer distance e when performing a mediolateral episiotomy, compared with a lateral episiotomy, to achieve the same distance b. Thus, b might be an important guiding parameter to perform an episiotomy with optimal unloading effect.Lastly, Kalis et al. 1 imply that a major limitation in our discussion is the use of data from studies investigating mediolateral episiotomy. We do agree with Kalis et al. 1 that different episiotomy techniques have inherent differences in outcome, but disagree with the idea that previous research is not relevant to the discussion of our results. Because of the variation in how mediolateral episiotomy is performed, it is currently unrealistic to limit the data from episiotomy studies to specific types of episiotomy. [5][6][7] In response to Knobel et al., 2 we do not promote a liberal use of episiotomy. On the contrary, we strongly support the restricted but correct use of episiotomy. We argue that there are situations where episiotomy is clearly indicated, for example with instrumental delivery.

Please note that letters and emails to the editor should be no more than 500 words with a maximum of five references.The management of urogynaecological problems in pregnancy and the early postpartum period Dear Sir Given that pregnancy alters both physiological and anatomical parameters, it is pleasing to read an article on urogynaecological problems specifically relating to pregnancy and the puerperium. 1 We were concerned that the section on success rates following repeat colposuspension after failed primary surgery lacks clarity. The figure of 81% after the first procedure refers not to the primary colposuspension but after the first repeat procedure and the figure of 25% after the second repeat (third colposuspension) procedure.2 Robinson and Cardozo summarise nine studies on the objective outcome of redo colposuspension 3 suggesting cure rates of between 65 and 86% following repeat colposuspension. Similar success rates have also been reported following repeat midurethral sling procedures. Rezapour and Ulmsten 4 report an 82% cure rate following the retropubic tension-free vaginal tape procedure in the treatment of recurrent urinary stress incontinence, with success also being reported in women with more than one prior procedure. Whilst retropubic tape procedures demonstrate similar continence rates after failed procedures, transobturator tapes appear less successful. Stav et al.5 report a statistically greater subjective cure rate in women who underwent the retropubic sling as a repeat procedure compared with the transobturator sling, irrespective of previous procedure (71% vs 48%, P = 0.04). Recent evidence suggests that subjective cure rates of 85% can be achieved when open colposuspension is used after an unsuccessful primary tape procedure although the development of de novo detrusor overactivity, known to be increased in patients undergoing repeat surgery, can be as high as 30%. Mike DiversNobles Hospital, Isle of Man Chrissy RadonNobles Hospital, Isle of Man Authors' reply Dear SirWe would like to thank our colleague for clarification on the efficacy of repeat colposuspension after a primary, secondary or tertiary procedure. A recently published Cochrane review of Burch colposuspension found insufficient data to comment on redo surgery. 1 However, the study published by Robinson et al., quoted by our colleague, states success rates of 65-86% for repeat colposuspension. 2 We would agree that similar results are obtained with mid urethral slings for redo surgery. A recent review of the management of recurrent stress incontinence specifically after a failed mid urethral sling suggests a cure rate of up to 70% for a secondary sling. Furthermore, an evaluation of risk factors for failed surgery suggests that multiple vaginal deliveries is the main risk factor for repeat surgery, but that synthetic slings at index surgery may be associated with lower risk of failure.