REVIEW: Reconstruction, destruction and IVF: dilemmas in the art of tubal surgery



The ‘IVF revolution’ in the last quarter of a century has been astounding for both patients and their doctors. Over a million babies have been born worldwide as a result of assisted reproduction technologies, and they now constitute 1–2% of all births in the United Kingdom. The demand for this treatment seems unlikely to diminish given the high rates of chlamydial infection causing tubal disease,1 and the increasing number of women who wish to delay childbearing until their fourth decade.

IVF was invented for tubal disease. But while tubal disease accounts for only 25–35% of infertility in women,2 IVF is now used to treat a wide variety of fertility problems including male factor, endometriosis, unexplained infertility and even recurrent miscarriage. The publication of league tables of assisted reproduction technologies clinic ‘success rates’3 is likely to increase public perception of the success of assisted reproduction technologies. Moreover, recent guidelines from the National Institute for Clinical Excellence (NICE) confirm that assisted reproduction technologies is indeed effective, and therefore should be more widely available in the NHS.4 Against this background, it is perhaps easy to dismiss alternatives to IVF, particularly when they predate IVF itself. But it is also notable that, although IVF was invented for tubal disease, IVF success rates are worst for that indication (particularly in the presence of hydrosalpinges), suggesting a more complex approach to tubal factor infertility might be necessary.

Practitioners of tubal surgery have certainly faced a substantial challenge to define their role, if any, in the management of tubal infertility. Data from the 1980s showed that, as a whole, the population with tubal infertility benefits little from surgery with intrauterine pregnancy rates of only 25%.5 This compares with rates of over 30% after one cycle of IVF and, in a study of 771 couples with tubal factor infertility, more than 70% had a live birth within four cycles of IVF.6 Unfortunately, no randomised controlled trial comparing IVF and tubal surgery has ever been done and there are few randomised trials even within the field of tubal surgery.7 In the modern era of evidence-based medicine, the continuing practice of tubal surgery may be judged by some to be ‘unscientific’. But that does not mean that it is not worth doing. On the contrary, there is a wealth of evidence that it can be highly effective. However, it does mean that the experience of the surgeon is critical. The ‘art’ of tubal surgery is accurate diagnosis, careful case selection and meticulous surgical technique. The RCOG has recognised the highly specialised nature of tubal surgery and recommended that it should only be done in a few tertiary centres.8

The RCOG guidelines state that tubal surgery is appropriate for mild distal disease and proximal tube obstruction. However, the choice of tubal surgery depends not only on the prognosis of surgical treatment but, crucially, on the accurate diagnosis of tubal disease. Unfortunately, there is no guidance whatsoever on how to make a definitive diagnosis, and moreover, many investigations (including laparoscopy) are performed before referral to a tertiary centre. This inevitably means that the difficult decision over whether to refer for surgery or not is almost always taken by gynaecologists who may not be trained to perform tubal microsurgery. They may not always appreciate the subtleties of proximal tubal disease, or the poor pregnancy rates achieved (11%) after treatment of disease affecting both proximal and distal tubal segments,9 or the very high rates (up to 85%) achieved for microsurgical reversal of sterilisation.10

This review aims to increase awareness of tubal disease diagnosis and surgery. It is not aimed to aggressively promote an alternative to IVF, but rather to show that tubal surgery and IVF should be complementary approaches to optimising a woman's reproductive potential.


Evidence for this review was gathered by searching PubMed with the following key words: tubal surgery, microsurgery, laparoscopic tubal surgery, reversal of sterilisation, proximal tubal disease, distal disease, hydrosalpinx, hysterosalpingogram, salpingography, tubal infertility. Only English language articles were included.

The fallopian tube: function

On the face of it, the 7–14 cm long fallopian tubes simply establish a connection between the ovaries and the uterine cavity. Hence, ‘blocked’ tubes are the most obvious features to diagnose tubal factor infertility. But the tubes are certainly not just passive conduits for sperm, eggs and embryos. Transvaginal hydrolaparoscopy has visualised an actual human ovum pickup by the fimbria.11 On the side of ovulation, fimbriae were noted to be congested and had pulsatile movements which swept the ovulated ovum into ciliated folds. Tubal secretions (which form the basis of artificial IVF culture media) provide the essential environment for fertilisation,12 and the cilia and muscular layers provide active transport for sperm and embryo.13 These critical factors cannot be easily assessed, but the gross appearances of tubal mucosa and fimbriae are perhaps the most important in determining the likely outcome of tubal surgery.14

The surgical approach

There is no place for open macrosurgery of the fallopian tube. The development of tubal surgery took a major step forward with the adoption of microsurgical principles.15,16 These include:

  • minimising tissue trauma

  • use of fine, non-absorbable sutures

  • use of intra-operative magnification

  • repairing all peritoneal surfaces

  • exquisite haemostasis

  • avoiding tissue dessication

Meta-analyses of non-randomised studies have shown microsurgery to be superior to macrosurgery for salpingostomy, adhesiolysis and reversal of sterilisation.7 Cornual reanastamosis was itself developed as a microsurgical technique,17 although superiority compared with macrosurgical treatment of proximal tubal disease has not been proven.

Laparoscopy has clearly presented a new challenge for tubal surgery. With its well-documented benefits of less post-operative pain and faster recovery, it also shares the basic principles of open microsurgery. However, while some procedures have been readily adopted (salpingectomy, adhesiolysis, salpingostomy), others (cornual reanastamosis) have proven to be extremely difficult.7,18,19 No randomised controlled trials have been conducted and, although there are some very successful reports,20,21 it will take time for those skills to be disseminated outside a few specialised centres.

Proximal disease: pathology and diagnosis

Proximal disease is found in 10–25% of cases of tubal infertility (Table 1). There may be mild obstruction associated with mucus plugs, cornual synechiae or polyps, and tubal endometriosis (endosalpingiosis) is found in 7–14% of cases.2 However, the most common histopathological finding is salpingitis isthmica nodosa. First described in 1887, it consists of firm nodular enlargements with a smooth intact serosal covering. The hallmark of salpingitis isthmica nodosa is the presence of diverticula or outpouchings of the tubal epithelium which are surrounded by hypertrophied smooth muscle. While the diagnosis of salpingitis isthmica nodosa can only be confirmed by histology, it can be suspected by hysterosalpingography if there is proximal obstruction or by a stippled appearance indicating contrast medium in the diverticular projections. At laparoscopy there is a nodular or shouldered appearance often associated with abnormal vasculature over the cornual region.

Table 1.  Diagnostic features of proximal and distal disease. Distal disease may be non-obstructive, or obstructive leading to hydrosalpinx.
 Proximal disease (salpingitis isthmica nodosa, endometriosis, mucus plugs, synechiae, polyps)Distal disease (congenital, inflammatory/ infective)
High perfusion pressureFlattened mucosa
‘Stippled’ appearance 
LaparoscopyObstructionHydrosalpinx (infective or congenital)
High perfusion pressureWall thickness
‘Nodular’ or ‘shouldered’ appearanceFlattened mucosa
Abnormal vasculatureFimbrial clubbing
Selective salpingography and catheterisationObstruction? hydrosalpinx
High perfusion pressure 
Intraluminal disease (mucus plugs, synechiae, polyps) 
SalpingoscopyFlattened mucosa
 Fimbrial clubbing
Hysterosalpingo-contrast-sonographyObstruction? hydrosalpinx

The cause of salpingitis isthmica nodosa is not known.22 Infection cannot be excluded although associated inflammatory lesions such as scarring and hydrosalpinx are not generally present. Salpingitis isthmica nodosa is found in 0.6–11% of healthy fertile women, up to 50% of women with ectopic pregnancy, more frequently in multiparous (over 70%) and black women, and is almost always bilateral.22 It is recommended that progression of salpingitis isthmica nodosa cannot be excluded with a hysterosalpingography done more than 12 months previously.23

Hysterosalpingography is still the most commonly used investigation for proximal tubal disease. Nevertheless, important considerations are that, on the one hand, proximal obstruction is often overdiagnosed (as a result of tubal spasm or temporary mucus obstruction), and on the other hand, that the tube can be diseased without necessarily being obstructed. Thus, any hysterosalpingography abnormality should be investigated further by laparoscopy to reassess both tubal patency and the possibility of non-obstructive disease. Furthermore, a ‘normal’ hysterosalpingography report should not necessarily exclude further investigation by laparoscopy. It has been suggested that excessive tubal filling pressures (measured during selective salpingography) are probably indicative of pathology24 and warrant further investigation. But filling pressures are not routinely measured at hysterosalpingography. Moreover, there is considerable variation in both hysterosalpingography performance and interpretation.25 Hence, in spite of recent NICE guidelines, a ‘normal’ hysterosalpingography should not contraindicate a laparoscopy to assess, among other factors (e.g. endometriosis), the possibility of proximal tubal disease. The complete evaluation of female infertility necessitates a combined approach of both intrauterine assessment by hysterosalpingography and/or hysteroscopy and pelvic assessment by laparoscopy.

More recently, newer techniques for assessing tubal status have stimulated some considerable interest, but their ability to diagnose proximal disease is unproven. Hysterosalpingo-contrast-sonography (HyCoSy) is an outpatient ultrasound technique involving the injection into the uterine cavity of a solution of galactose microparticles in 20% galactose (Echovist, Schering, Berlin).26 The accompanying air bubbles can be followed through the uterus and fallopian tubes enabling an assessment of intrauterine abnormalities (e.g. polyps) and tubal blockage or distal disease (e.g. hydrosalpinx). However, it is unlikely to detect non-occlusive proximal disease, and can fail to diagnose large hydrosalpinges. HyCoSy also cannot be used to assess ampullary folds, the presence of which are major prognostic factors for effective tubal function.

‘Fertiloscopy’ is a combination of transvaginal hydrolaparoscopy, salpingoscopy and hysteroscopy, and has been proposed as a new alternative to diagnostic laparoscopy.27 However, it is not widely available and there is concern over the risk of large bowel perforation during insertion of the fertiliscope into the pouch of Douglas.28 Indeed, the RCOG recommendation of bowel preparation beforehand makes it arguably more unpleasant than hysterosalpingography. As with HyCoSy, there is no evidence that fertiloscopy can reliably diagnose proximal disease.

Proximal tubal patency may also be assessed by selective salpingography, with tubal catheterisation to treat obstruction where necessary.29 This technique uses radiologically controlled coaxial wire-guide systems and is most effective for diagnosis and treatment of tubal spasm, synechiae and mucus plugs with some authors reporting pregnancy rates of 20–40%.30 However, the value for tubal wall disease such as salpingitis isthmica nodosa is uncertain.

Proximal disease: treatment

Proximal tubal surgery is rarely performed outside very few specialised centres (Table 2). Perhaps the difficulty with diagnosis and the allure of the immediacy of IVF have led to its decline as a major treatment but this, ironically, ignores the consistently high success rates reported. Reconstructive proximal surgery was first reported for tubal occlusion in 1896, and by 1971 those macrosurgical techniques yielded pregnancy rates of 34%.2 A major advance was the development of microsurgical principles.15,16 Proposed advantages of this approach are a decreased risk of adhesions and less disruption of cornual blood supply, and hence a greater likelihood that tubal function is preserved. Post-operative term pregnancy rates of up to 56% with ectopic pregnancy rates of less than 10%15,31 have been reported. In a more recent study, 120 women with proximal occlusion of one or both tubes treated by microsurgery had a cumulative intrauterine pregnancy rate of 68% at 24 months.32 Fertility outcome was better in women aged less than 37 years and when anastamosis was carried out bilaterally. The overall ectopic pregnancy rate was 10.8%. The success of this surgery is further enhanced by the possibility of subsequent pregnancies. In a study of 312 women who had tubal microsurgery, over half of the women who wanted a second child succeeded over a two-year period.33

Table 2.  Tubal surgery summary.
 ProcedureReported pregnancy rates
Proximal diseaseOpen macrosurgery34%2
Selective salpingography/ catheterisation9–40%35
Open microsurgery cornual reanastamosis 
 occlusive disease56–68%31,32
 non-occlusive disease46%37
Distal disease (hydrosalpinx)Open microsurgery salpingostomy22–31%43,44
Laparoscopic salpingostomy14–33%43,46
Laparoscopic salpingectomy + IVF/IVF alone38–46%/10–25%43
Laparoscopic tube occlusion + IVF/IVF alone36%/8%66
Transvaginal aspiration + IVF/IVF alone31%/5%57
SterilisationOpen microsurgery reversal55–85%10,70
Laparoscopic reversal56–80%20,21,73

The restoration of proximal tubal patency can also sometimes be achieved by a less invasive transcervical recanalisation procedure using selective salpingography followed, if necessary, by tubal catheterisation with a soft Teflon catheter and finally, if needed, by wire-guide cannulation.32 In a study of 66 women, patency was achieved in 34.5% by selective salpingography alone, in 46% by tubal catheterisation and in 8.9% by wire-guide, with 10.6% of tubes remaining occluded.34 Pregnancy occurred in 36% of those women. Other studies report even lower pregnancy rates of 12.8%,35 and the technique is not without risk with a rate of tubal perforation of 2%.36 While catheterisation is not yet widely available, there is certainly diagnostic and therapeutic value for spasm, synechiae and mucus plugs, but there is less consensus for benefit for occlusive tubal wall disease (salpingitis isthmica nodosa, fibrosis, endometriosis).

Proximal tubal obstruction is most commonly due to salpingitis isthmica nodosa, and disease is usually limited to the proximal tube (unlike distal disease which is more often pan-tubal). Whether or not salpingitis isthmica nodosa has an infectious aetiology, it is hypothesised that the occlusion may impede ascent of infection and so spare the distal tube. This phenomenon would explain the relatively high success rates for proximal tubal surgery compared with distal surgery for hydrosalpinx where inflammation has travelled along the entire tube.37 However, in a prospective study of 26 women with non-occlusive salpingitis isthmica nodosa (confirmed histologically) who underwent microsurgical resection and reanastamosis, similarly high pregnancy rates were reported (46% within 10.5 months).37 Three women had ectopic pregnancies after surgery, which was less than the ectopic pregnancy rate before surgery. Thus, resection of diseased proximal tube appears to be beneficial, not just restoration of patency. It is likely that local inflammation of the tubal wall affects tubal function and hence gamete and embryo transport.

Distal disease: pathology and diagnosis

Disease of the distal tube can be secondary to any pelvic inflammatory condition including infection, endometriosis, appendicitis and abdomino-pelvic surgery (Table 1). The common endpoint of distal disease is occlusion which results in the collection of watery fluid and formation of a hydrosalpinx. The first attempts to grade distal tube disease were based on the size of hydrosalpinx, thickness of the tubal wall, and the presence of peritubal adhesions.38,39 More recently, studies using salpingoscopy performed at the same time as laparoscopy have emphasised the importance of an assessment of the tubal mucosa.14 Thus, hydrosalpinx can be classified, with increasing degrees of severity, into hydrosalpinx simplex (thin-walled tube with flattened mucosal folds and without mucosal adhesions), hydrosalpinx follicularis (thin-walled tube with adhesions) and thick-walled hydrosalpinx.

In practice, the value of routine salpingoscopy is not yet established, and no one grading system has been adopted widely. It is clear though that an assessment of the tubal mucosa is essential, particularly as there is a poor correlation between mucosal and serosal adhesions.40 This is most easily done by a combination of hysterosalpingography and laparoscopy. Clinicians may sometimes be unaware that hysterosalpingography is not solely used to assess tubal patency, but it can also diagnose both non-occlusive proximal disease, and distal tube mucosal flattening, adhesions and hydrosalpinx.

Overall, the factors that determine the likely outcome of tubal corrective surgery are the condition of the endosalpinx, the amount of tubal dilation, the condition of the tubal wall and the presence of mucosal and serosal adhesions. In practice, without any consensus grading system, management decisions are based on firstly whether or not there is both proximal and distal (bipolar) disease,41 and secondly whether the distal tube assessment constitutes ‘severe’ or ‘mild’ disease.5 These judgements inevitably rely on the experience of the clinician, and this is a major reason for focussing tubal surgery expertise in a few specialised centres.

Distal disease: treatment

A blocked distal tube leading to the hydrosalpinx is the most common and intuitively obvious cause of tubal infertility.42 What is not so obvious is the variety, and often diametrically opposite approaches to managing the problem. Traditional tubal surgery has focussed on treating the ‘blockage’per se, whether by salpingostomy (opening the distal tube), neosalpingostomy (fashioning a new stoma, including everting the end of the tube to expose tubal mucosa to the peritoneal cavity), or fimbrioplasty (removing fibrous tissue and freeing the fimbriae). The majority of women with hydrosalpinges treated by reconstructive surgery have relatively lower clinical pregnancy rates of 14.5–33%.43,44 Fimbrioplasty is slightly more successful with rates of up to 51.4% reported.45 These data are based on small numbers of women (trials involving 35 to 211 cases), operated on by a few highly specialised surgeons using different techniques (open microsurgery or laparoscopic surgery), and reported over a 20-year time scale. As IVF success rates have become more comparable, it is perhaps hardly surprising that enthusiasm for distal tubal surgery has waned (Table 2).

It is likely that better results from reconstructive distal tubal surgery can be obtained from careful case selection, but no one means of assessment has been universally applied. Current RCOG guidelines state that IVF rather than tubal surgery should be considered as first line treatment for ‘moderate to severe distal tubal disease’, although no guidance is given in defining such a diagnosis.8 Vasquez et al.40 identified a good prognosis group, comprising only about a third of all patients, with thin-walled hydrosalpinx and reasonably healthy mucosa. Women generally deemed unsuitable for reconstructive surgery are those with dilated tubes over 3 cm in diameter, thickened tubal walls, mucosal adhesions, dense or extensive serosal adhesions,5 and those with both proximal and distal (bipolar) disease.41,43 And it was recently shown that women who have unilateral salpingostomies have better spontaneous pregnancy rates (up to 44%) than those who require bilateral salpingostomies.42

A further important factor when contemplating distal surgery is the relatively high rate of ectopic pregnancy, as was confirmed recently in139 women treated by laparoscopic salpingostomy in whom the intrauterine and ectopic pregnancy rates were 24.5% and 16.5%, respectively.46 Those authors went on to show by logistic regression analysis that careful patient selection (excluding women with previous ectopic pregnancy, use of intrauterine coils, pelvic inflammatory disease, the presence of perihepatic adhesions and bilateral hydrosalpinges) could potentially reduce the ectopic risk and so improve outcome.

However, it is not just the ‘blockage’ that has a negative effect on pregnancy rates. Although IVF was initially developed primarily for couples with severe tubal disease, IVF success rates are consistently lower in the presence of hydrosalpinges.47–52 This phenomenon is most likely caused by a combination of a mechanical effect of fluid sweeping embryos away from the implantation site,53 and an effect on endometrial receptivity as observed by the impairment of endometrial αvβ3 integrin expression.54 Although human hydrosalpinx fluid is toxic to animal embryos,55 there is little evidence to suggest a toxic effect on human embryo quality in vitro or in vivo.56

Thus, drainage of hydrosalpinx may in itself improve the likelihood of pregnancy, and this may be one mechanism for the success of salpingostomy.42 The problem with simple drainage (by laparoscopy or transvaginal ultrasound guidance) is the high chance of reaccumulation of the hydrosalpinx fluid, although it has been shown to be of benefit in the short term. In a recent study of 34 women with hydrosalpinges diagnosed by ultrasound scan and undergoing IVF treatment, 16 had their hydrosalpinges aspirated at the time of oocyte retrieval.57 Although a small study, aspiration was associated with a significantly higher pregnancy rate (31%vs 5%). The risk of infection associated with the puncture of a hydrosalpinx has not been properly evaluated.58

An alternative to drainage, and the subject of intense recent debate, has focussed on preventing hydrosalpinx fluid from reaching the endometrial cavity by either salpingectomy or proximal tubal occlusion. Thus, we have the irony of tubal surgery to specifically destroy or even remove the fallopian tube for the purpose of improving fertility. A meta-analysis including 14 studies and 1004 women with hydrosalpinx concluded that IVF pregnancy rates were significantly lower in the presence of hydrosalpinx.51 These observations led to reports that, in cases where hydrosalpinx is detectable by ultrasound scan, salpingectomy significantly increases IVF pregnancy rates.59 This has been confirmed in a recent Cochrane review of three randomised controlled trials60 although is not currently recommended by the RCOG guidelines.8

These studies have sparked a fierce debate because of their potential to lead to widespread practice of unnecessary salpingectomy, particularly in cases which may be amenable to reconstructive surgery by reproductive surgery specialists. In some cases, salpingectomy may be performed before an unexpectedly poor-response IVF cycle leaving little left to offer. There is also some evidence that salpingectomy may decrease ovarian blood supply and hence ovarian responsiveness to IVF. In a study of 29 women who had undergone unilateral salpingectomy for ectopic pregnancy (mostly by laparotomy), fewer follicles developed on the operated side although there were no differences in total numbers or pregnancy rates.61 And similarly, in 32 women who had salpingectomy for ectopic pregnancy (18 by laparoscopy), the operated side had significantly reduced antral follicle counts and 3D power Doppler indices.62 In contrast, in a similar sized study, Tal et al.63 found that unilateral salpingectomy (by laparoscopy or laparotomy) for ectopic pregnancy did not affect ovarian response even after repeated IVF cycles one to nine years after surgery.

It has also been argued that salpingectomy for ectopic pregnancy is different from a prophylactic salpingectomy because the pregnant tube may be haemorrhagic and oedematous, rendering surgery more imprecise and difficult.64 This is supported by studies showing that prophylactic salpingectomy does not alter ovarian response to IVF treatment.64,65 Nevertheless, these concerns, as well as the sometimes difficult surgery associated with pelvic adhesions, have led to a simpler alternative: proximal tubal occlusion by cauterisation. Although there are no randomised data, comparative studies have shown laparoscopic tubal occlusion to be as effective as salpingectomy in improving IVF pregnancy rates.66,67 There are no data yet describing any decreased ovarian reserve or responsiveness after tubal occlusion as a treatment for hydrosalpinx. However, in a five-year follow up study of women sterilised by laparoscopic bipolar coagulation, there was no effect on ovarian follicular reserve or ovarian hormone production.68

It has also recently been reported that laparoscopic unilateral salpingectomy or occlusion not only improve IVF outcome, but also the potential for spontaneous pregnancy.69 In a retrospective study of 25 women with an average of three years infertility, 88% achieved spontaneous intrauterine pregnancies in an average of 5.6 months after treatment. However, unilateral hydrosalpinx is unusual, particularly when the cause is ascending infection. Thus, those exceptionally high pregnancy rates may reflect a different tubal pathology or at least very mild disease.

Reversal of sterilisation

‘Reversal of sterilisation’ procedures are rarely funded by the NHS but, nevertheless, are among the most successful infertility treatments available. The operation is usually performed by laparotomy, and the use of an operating microscope has been shown to be beneficial.7 Pregnancy rates of 55–85% have been reported,10,70 with higher rates in women aged 35 years or less. The pregnancy rate is not significantly correlated with the method and duration of sterilisation, but the interval from operation to pregnancy decreases as the post-operative tubal length increases.70 The majority of women conceive within a year of surgery.71

The first intrauterine pregnancy after laparoscopic reversal of sterilisation was reported in 1989.72 Although obviously advantageous in terms of post-operative pain and recovery, it has been a significant technical challenge, and initial studies found poor success rates compared with open microsurgery.19 Nevertheless, as technology and skills have improved, so has the potential for laparoscopic tubal reanastamosis to one day replace open microsurgery. In a series of 102 women in whom reanastamosis was achieved with a single stitch, an intrauterine pregnancy rate of 70% (and 7.2% ectopic) was reported.20 And, although even more demanding, a two-layer reanastamosis equivalent to open microsurgery produced a success rate of 80.5% in a series of 202 women.21 However, more recently, a study using ‘standard’ laparoscopic instruments with a single-layer reanastamosis produced a success rate of only 56.5%, after an operating time of up to two and a half hours.73 There is clearly still a significant ‘learning curve’ to be overcome.


Present evidence suggests that surgical adhesiolysis (using sharp dissection, laser or diathermy) is beneficial, although this is based on just one non-randomised controlled trial comparing open adhesiolysis and no treatment.7 In that study, 69 infertile women with periadnexal adhesions were treated by laparotomy and salpingo-ovariolysis and 78 were not treated. After two years, the cumulative pregnancy rates were 45% and 16%, respectively.74 Since then numerous small studies have suggested laparoscopic adhesiolysis to be at least as beneficial as open adhesiolysis with cumulative pregnancy rates of 50–60% after two to three years5,75 and a randomised trial has been recommended.7 However, this is unlikely to be done given the fact that laparoscopic surgery is believed to result in fewer de novo postsurgical adhesions than open surgery,76 and the increasing tendency to manage women with a single ‘see and treat’ laparoscopic procedure.


Tubal microsurgery (whether open or laparoscopic) is often a demanding undertaking which, given the relatively small numbers of women who need it, should be restricted to specialised centres. What is of more general importance, though, is the need for all Gynaecologists and their patients to be aware that it may be an option for women with tubal disease.

The option of tubal surgery depends on there being an accurate diagnosis of tubal disease (including severity). And in the absence of a universally accepted grading system, early referral to a specialist centre, ideally with any imaging or photographic evidence, should be encouraged.

So who should be considered for tubal surgery? Any coexistent male factor abnormality is the most important contraindication. Generally, the more severe the tubal disease the less likely surgery will be effective, and those women should be treated by IVF immediately. On the other hand, women who have a predicted or have had a poor response to IVF may still become pregnant spontaneously after tubal surgery. Management decisions are also often influenced by ethical and financial factors.

In the presence of hydrosalpinges, it may be unwise to perform salpingectomy or tubal coagulation prior to at least one IVF attempt. A poor response or cancelled IVF cycle would then leave no option of any salvageable surgery. But salpingectomy or tubal coagulation may be an effective treatment option if IVF was unsuccessful.

Finally, it is important to emphasise that tubal surgery (reconstructive or destructive) and IVF have the same aims to maximise a woman's reproductive potential. Once embarked on a course of IVF treatment, there is no reason not to reconsider tubal surgery if it does fail. And conversely, tubal surgery should be at least discussed prior to starting IVF.