Editorials

Urinary stress incontinence

Benefits of using tension-free vaginal tape remain unproved

New minimal access surgical sling procedures such as the tension-free vaginal tape procedure are now being used to treat urinary stress incontinence in women. These procedures use minimal tensionurethral support is perhaps achieved from a tissue reaction to the tape, which produces a collagen scar along the length of the tape and increases support of the bladder when the rectus muscle contracts. The tension-free vaginal tape procedure is often carried out under regional or local anaesthesia. Many women would undoubtedly welcome the choice of a less invasive procedure than open retropubic colposuspensionas long as the minimal procedure cures urinary incontinence and does not result in major complications.

What does the currently available evidence say? Two recent systematic reviews concluded that, although the minimal access surgical sling procedures (and particularly tension-free vaginal tape) may be promising, the quality of the evidence available so far is not conclusive. ^{1 2} So far almost all the evidence has come from case series. So the six month results for the first multicentre randomised trial of tension-free vaginal tape have been eagerly awaited,³ since only preliminary results could be included in these systematic reviews.

Ward et al concluded that in their trial tension-free vaginal tape was as effective as colposuspension in treating stress incontinence. Several aspects of the trial, as well as results from other studies of tension-free vaginal tape, indicate that this claim may still be somewhat premature. In the trial, the authors prespecified that a difference of 10% in cure rates between tension-free vaginal tape and colposuspension would be clinically important, but unfortunately they were unable to recruit the required number of 394 patients overall. Ultimately 344 women were randomised, and only 287 completed the urodynamic investigations at six months156/175 (89%) after tension-free vaginal tape and 131/169 (78%) after colposuspension. Ward et al analysed most of their data on an intention to treat basis, making the assumption that all missing patients were treatment failures. This assumption is questionable, however, and it would have been better also to reanalyse assuming missing patients were treatment successes and then attempt to explain any differences between the sets of results.⁴ Unfortunately, a more correct interpretation (taking into consideration the underpowered study and missing results) is that tension-free vaginal tape may be better, worse, or the same as colposuspension in this study.

Of greater concern, however, may be the biases inherent in this study that seem to favour tension-free vaginal tape. A large number of women who agreed to join the trial seem to have withdrawn when placed in the colposuspension group. Given that the study was performed in the NHS and that tension-free vaginal tape may have been readily available for trials only, it may well be that patients were willing to continue in the trial only if they were allotted to the group treated with the less invasive approach. The report mentions that the women who withdrew from the colposuspension group before surgery had less severe incontinence. This reinforces the importance of closer analysis of these women in interpreting the results obtained, ideally in a true intention to treat analysis, which requires testing of several assumptions.

Irrespective of these statistical and methodological concerns, women treated with vaginal tape had shorter operating times and reduced hospital stays than women treated with open colposuspension, as would be expected from a less invasive technique. They were also able to return to work and normal activities more rapidly. What remains to be addressed, even if the operations are of equal short term benefit for incontinence, is the issue of long term cure rates. This will require further follow up of patents in both groups. Women will also need to base their decisions on the possibility of complications. The trial showed more intraoperative complications, such as perforation of the bladder, in the vaginal tape group. Postoperative complications, such as infections, were more prevalent in the colposuspension group.

Although this study begins to add evidence from randomised controlled trials, at best we can conclude only that we need further studies with adequate power and long term follow up, whose outcomes are analysed more stringently. The study by Ward et al also shows that, even when resources can be found to start a randomised controlled trial, many factors coincide to make surgical research problematic. It is often difficult to recruit sufficient patients to produce a robust result. Especially in surgery, authors of underpowered trials tend to make unsupported recommendations for changes in practice.⁵ The often unknown motivations of participants who withdraw from trials once they are aware of their treatment allocation and the multiple reasons for becoming lost to later follow up make intention to treat analyses more complex. For surgical randomised controlled trials, issues related to the learning curve and differences in performance between surgeons complicate analysis of results even further. ^{6 7}

We must, however, find ways to optimise the ability of randomised controlled trials to answer the questions that they were designed to answer. Funders need to be prepared to fund randomised controlled trials on the basis of realistic recruitment rates, or they may need to provide "top up" or contingency funding so that trials do not remain underpowered. Trialists should also have access to specialised statisticians and methodologists who have the skills to model scenarios, particularly for inputing missing data, such as losses to follow up. The alternative is that we revert to reliance on observations of current practices, where variation in practice might be seen as a large but poorly controlled experiment,⁸ to attempt to make informed choices between treatments. The difficulties experienced with the tension-free vaginal tape trial mean that women needing to choose between minimal access and conventional surgery still do not have enough evidence to make this decision, even though the difficulties were probably surmountable.

G J Maddern, RP Jepson professor of surgery. 
P F Middleton, senior research officer. 

ASERNIP-S, PO Box 688, North Adelaide SA 5006, Australia

A M Grant, director. 

Health Services Research Unit, University of Aberdeen, Aberdeen AB25 2ZD, Scotland

Footnotes

   Competing interests: AMG is a member of a group commissioned by the National Institute for Clinical Excellence to conduct a technology assessment review of tension-free vaginal tape.

1.	Bezerra CA, Bruschini H. Suburethral sling operations for urinary incontinence in women. Cochrane Database Syst Rev 2001;(3):CD001754.
2.	Merlin T, Arnold E, Petros P, MacTaggart A, Faulkner K, Maddern G. A systematic review of tension-free urethropexy for stress urinary incontinence: intravaginal slingplasty and the tension-free vaginal tape procedures. ASERNIP-S Report No.11. Adelaide, South Australia: ASERNIP-S, February 2001. www.surgeons.org/open/asernip-s.htm (accessed 10 Sep 2002).
3.	Ward K, Hilton P, on behalf of the United Kingdom and Ireland Tension-free Vaginal Tape Trial Group. A prospective multi-centre randomised trial of tension-free vaginal tape and colposuspension as primary treatment for stress incontinence. BMJ 2002; 325: 67[Abstract/Free Full Text].
4.	Deeks JJ, Altman DG, Bradburn MJ. Statistical methods for examining heterogeneity and combining results from several studies in meta-analyis. In: Egger M, Davey Smith G, Altman DG, eds. Systematic reviews in health care: meta-analysis in context. London: BMJ Books, 2001:285-312.
5.	Orseck M, Johnson J, Orr R. Type II error in randomized controlled trials with negative resultsare methods improving? Current Surgery 2001; 58: 561.
6.	Varma R, Neale E. Tension-free vaginal tapeis there a learning curve? [electronic response to Ward et al. Prospective multicentre randomised trial of tension-free vaginal tape and colposuspension as primary treatment for stress incontinence]. BMJ 2002. bmj.com/cgi/eletters/325/7355/67#23951 (accessed 9 Sep 2002).
7.	Hilton P. Re: Tension-free vaginal tapeis there a learning curve? [electronic response to Ward et al. Prospective multicentre randomised trial of tension-free vaginal tape and colposuspension as primary treatment for stress incontinence]. BMJ 2002. bmj.com/cgi/eletters/325/7355/67#24182, 25 Jul 2002 (accessed 9 Sep 2002).
8.	Alderson P, Roberts I. Should journals publish systematic reviews that find no evidence to guide practice? Examples from injury research. BMJ 2000; 320: 376-377[Free Full Text].