Surgical Aspects of Infertility

Infertility is estimated to affect 12% of reproductive age women attempting to conceive in the United States.¹ Its consequences are manifested in many ways in our society, including the monetary costs of its investigation, diagnosis, and treatment as well as the psychosocial stresses^{2, 3} it imposes on this portion of the population.

Causes of infertility potentially amenable to surgical treatment include tuboperitoneal disease and abnormalities of the uterine cavity. These are estimated to account either alone or in part for the infertility of 40% of couples.⁴ Anatomic findings within this broad category include adhesions, implants of endometriosis, and partial or complete tubal obstruction. Their most common causes are prior surgery, endometriosis, and infection (either clinically recognized or unrecognized). With regard to abnormalities of the uterine cavity, some conditions such as Asherman's syndrome are surgically correctable causes of infertility. The contribution to the development of infertility of other surgically correctable pathologic findings, such as uterine septa or intracavitary myomas, is variable and is thought to depend on the specific characteristics of each lesion.

Over the past several decades, new techniques and instruments have been introduced that have restructured the performance of reproductive pelvic surgery. First was the recognition of a role of “microsurgery” in the performance of these procedures. The second was the application of endoscopic surgery as an alternative to laparotomy. Laparoscopy became the preferred approach in the surgical treatment of infertility, although there are no well designed randomized studies showing an outcome benefit of laparoscopy over laparotomy. However, laparoscopy adds the advantage of magnification, which is integral to microsurgery, early recovery, and potential cost saving. The recent introduction of robotics into endoscopic surgery made the performance of more complex procedures more feasible. Third was the choice for tissue incision or excision with the use of electrosurgery, harmonics, sharp dissection, or lasers. Laser application did not show any advantage over other forms of energy and its use has peaked and decreased over the past decade.

Although the role of tubal surgery has diminished with the advent of assisted reproductive technology (ART) and in vitro fertilization (IVF), this type of surgery is still widely performed in tertiary centers on selected patients and in units where IVF is not available.^{5, 6} Therefore, more than previously, the role of surgical aspects of infertility is undergoing re-examination. This review attempts to describe considerations in the decision-making process in regard to whether, when, and how to perform reproductive pelvic surgery. An introduction to each of these advances is given, followed by an evaluation of individual surgical operations. Direct comparisons of surgical series conducted under varying conditions are not entirely appropriate, but they do allow guidance for the aforementioned questions. The series included in the tables were selected from the literature; more complete listings are described in the references. Additionally, pregnancy outcome in many of the reports had to be modified based on the information provided to conform to the subset desired (e.g., conception rates versus term deliveries). Finally, it should be recognized that ART can also be considered an alternative to surgical repair in some patients with extensive tuboperitoneal disease. The appropriate applications of surgical treatment via laparotomy, laparoscopy, or ART vary among patients and are likely to be modified as advances are made in these areas. (A discussion of the role of ART in the treatment of tuboperitoneal-based infertility is beyond the context of this chapter; the reader is referred to the ART section of this series.)

Microsurgery

Use of microsurgery in the performance of reproductive pelvic procedures was initially described by Swolin in 1967.⁷ Microsurgery in this context includes, but is not limited to, the use of magnification. The advantage of magnification is that it allows close inspection so as to improve the ability to discriminate between normal and pathologic tissue and to allow the use of fine-caliber microsurgical instruments and sutures.⁸ Other tenets of gynecologic microsurgery include achieving meticulous hemostasis, minimizing tissue handling, preventing desiccation of tissue, avoiding introduction of foreign bodies (e.g., talc) into the operative field, using fine suture material of low tissue reactivity, and precisely reapproximating tissue planes.^{8, 9}

Although most of these tenets are probably correct, one that may be incorrect is the value of precise tissue approximation. In animal studies, suture closure of the ovarian cortex was associated with greater, not lesser, adhesion development.¹⁰ It remains unclear whether this represents a foreign body reaction to the sutures, a response to the increased tissue handling required to place the sutures, a consequence of tissue anoxia, or a combination of all these factors. Regardless of the actual mechanism, this report¹⁰ and others like it appeared to call into question the routine “reperitonealization” of all peritoneal and serosal defects, particularly those that do not protect vital structures. Finally, important corollaries exist to two of the tenets of microsurgery: achieving meticulous hemostasis and minimizing tissue handling. Although the presence of blood at the operative site increases postoperative adhesion development, it is important that hemostasis be achieved in a manner that devitalizes as little surrounding tissue as possible. When possible, the size of pedicles should be minimized and use of electrosurgery should be restricted to the actual bleeding site. With regard to tissue handling, manipulation of structures is required to achieve exposure and perform the procedure. However, tissue damage can often be minimized by the use of atraumatic graspers, moist (not dry) pads, and, when feasible, grasping of tissue structures to be excised.

A less recognized, yet extremely important, characteristic affecting the results of microsurgical gynecological procedures is the experience of the surgeon. Oelsner and coworkers¹¹ examined this issue directly, assessing microsurgical skill as a function of the number of isthmic reanastomoses performed on the rabbit fallopian tube. In evaluations of pregnancy rates, nidation index, and scanning electron microscopy sections of the anastomotic sites, they demonstrated a positive correlation between increasing experience and successful surgical outcome. Another report indirectly addressed this issue. DeCherney and associates¹² examined the failure rate of microsurgical tubal anastomoses, and noted that, compared with an initial success rate of approximately 50%, pregnancy outcome improved with experience, reaching a plateau at a rate of approximately 70%. In a third report on 387 cases of microsurgical tubal reversals, Kim and colleagues determined that the surgeon’s experience was among the factors that influenced success.¹³

Direct comparison of clinical outcome of surgical procedures performed by macrosurgery versus microsurgery is difficult because of differences in the procedures performed, in the extent of coexistent pelvic disease, and in chronology. Introduction of different centers and surgeons further complicates such comparisons. These observations notwithstanding, several investigators have conducted such examinations and have found trends favoring successful pregnancy with microsurgery.^{14, 15, 16, 17}

Many investigators have advocated adoption of the principles of gynecologic microsurgery;^{9, 18, 19, 20, 21, 22} its use and efficacy are described in subsequent sections of this chapter dealing with specific operative procedures. Use of these principles, however, should not be limited to the microsurgeon, but should be employed by all surgeons performing pelvic operations so as to minimize postsurgical development of pelvic adhesive disease.

Operative endoscopy

DeCherney, in 1985, predicted that “the obituary of laparotomy for pelvic reconstructive surgery has been written; it is only its publication that remains.”^{23, 24}More than 20 years later, this prophesy remains almost as controversial as when it was written. Clearly, a wide variety of the reproductive pelvic surgical procedures traditionally performed at laparotomy are becoming more commonly performed at laparoscopy or hysteroscopy. This list includes adhesiolysis, fimbrioplasty, neosalpingostomy, linear salpingostomy, salpingectomy, fulguration/vaporization of endometriosis, myomectomy, transection of uterosacral ligaments, drainage or excision of ovarian cysts, and removal of uterine septa. Later sections of this review describe the clinical efficacies of these procedures as performed by laparotomy or by laparoscopy/hysteroscopy. The reader is also referred to recent reviews for more in-depth considerations.

Application of lasers

Several investigators have examined the application of lasers for the performance of reproductive pelvic surgical procedures, both at laparotomy and in endoscopic procedures. The most frequently used type to date is the carbon dioxide (CO₂) laser;²⁵others now in use include the argon, the potassium titanyl phosphate (KTP-532), and the neodymium:yttrium-aluminum-garnet (Nd:YAG) lasers. A comparison of characteristics of these different types of lasers is made in Table 1. Efficacy of their use is described for individual procedures. In general, there has been no evidence that use of a laser per se results in a greater reduction of adhesions or improvement in pregnancy outcome, compared with other surgical modalities.^{26, 27, 28, 29} However, individual surgeons, based on their own experience, equipment availability, and preference, may find use of a particular laser to be most advantageous for the performance of these procedures.^{30, 31, 32, 33}

Table 1. Properties of surgical lasers

(Adapted from Daniell JF: The role of lasers in infertility surgery. Fertil Steril 42: 815, 1984. Reproduced with permission of the publisher, the American Society for Reproductive Medicine.)

Adhesiolysis

Pelvic adhesions have been identified as a major contributing cause of infertility. In these patients, goals of reconstructive pelvic surgery include lysis of adhesions that interfere with physiologic processes of ovum release, pickup, and transport, and minimal subsequent development of postoperative adhesions.

Classification of the type or extent of pelvic adhesion has been described in several previous reports.^{34, 35, 36, 37, 38, 39, 40} Jessen³⁸ described an improved pregnancy rate in the presence of low-grade compared with high-grade adhesions (64% versus 22%, respectively). Subsequently, Hulka and colleagues³⁵classified the extent and type of adnexal adhesions, and later the same authors demonstrated their prognostic value.³⁷Pregnancy was more likely to occur when more than 50% of the ovary was visible and when adhesions were filmy and avascular, rather than dense and vascular. Similarly, Caspi and coworkers³⁶ demonstrated that pregnancy rates were higher in association with fine, avascular as opposed to fibrous, coarse adhesions. Mage and his group³⁹ reported on pregnancy in relation to extent of tubal damage and severity of adhesions determined preoperatively by hysterosalpingography and laparoscopy. The postsurgical pregnancy rate was 32% with mild adhesions, 26.6% with moderate adhesions, and 5.5% with severe adhesions.³⁹ In these reports, as well as that of Young and associates,⁴⁰ the significance of adhesions (as assessed by the pregnancy rate) was clouded by the classification systems, tubal patency, performance of fimbrioplasties or salpingostomies, and presence or absence of endometriosis as a possible contributing factor to the cause of adhesions.

The report that has most closely assessed the role of periadnexal adhesions alone on pregnancy outcome is that of Bronson and Wallach.³⁴They scored adhesions at varying locations (right and left ovaries, fallopian tubes, pelvic sidewalls, broad ligaments, cul-de-sac, and rectosigmoid) and observed that pregnancy rates were lower among women with high adhesion scores, although the difference was not significant. Importantly, however, in these studies looking at pregnancy outcome,^{34, 35, 36, 37, 38, 39, 40} the incidence, extent, and severity of adhesions at the time the women were trying to conceive is not known, nor were there extensive controls for other factors that may contribute to infertility.

In view of the need for a standard classification scheme for mechanical problems (such as adhesions) associated with infertility, the American Fertility Society (currently known as the Americam Society of Reproductive Medicine) published a classification system in 1988. In this classification both ovaries and tubes are evaluated. Different scores are given dependent on whether the adhesions are filmy or dense and whether they enclose less than one third, one third to two thirds, or more than two thirds of the ovary or the tube. Higher scores are given if the fimbriated end of the tube is completely enclosed. Depending on the sum of all the scores, adhesions are arbitrarily classified into minimal, mild, moderate or severe. The prognosis for conception is based on the score applied for the adnexa with the least amount of pathology.⁴¹

Adhesiolysis has traditionally been performed by sharp dissection or electrosurgery. It has also been performed using the CO₂,Nd:YAG, and argon lasers,^{42, 43} or linear cutters. More recently, harmonic energy has been employed in performing adhesiolysis. A comparison of pregnancy outcome after adhesiolysis at laparotomy and at laparoscopy with various techniques is shown in Table 2.^{16, 19, 36, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53} In general, the use of microsurgical techniques has led to a trend for improvement in pregnancy outcome after adhesiolysis. As shown, in women in whom there is no distal tubal obstruction, adhesiolysis can be expected to lead to pregnancy in 50–60% of cases. Use of the CO₂ laser, compared with other sources of energy or sharp dissection, has not resulted in improvements in pregnancy outcome,^{49, 50}nor has it resulted in consistent reductions in the adhesions observed at the time of second-look laparoscopy.⁵⁴ Consistent with these findings is the report of Pittaway and colleagues,⁵⁵ which described no reduction in postoperative intraperitoneal adhesion formation in rabbits and rats, respectively, with the CO₂ laser as compared to electrocautery.⁵⁶

Table 2. Pregnancy outcome following adhesiolysis 

With the advent of laparoscopic techniques, adhesiolysis as well as other reproductive pelvic surgical procedures, are commonly performed laparoscopically. Although relatively minor procedures can be performed by most laparoscopists, treatment of more extensive disease requires advanced training and experience. In experienced hands, laparoscopic treatment appears to yield results similar to those achieved with laparotomy.^{52, 53, 57} However, randomized controlled studies examining this issue are scarce. Mossa and colleagues randomized 224 women with distal tubal occlusion infertility to treatment with laparoscopy or laparotomy and microsurgery. After a 24-month follow-up period, there was no significant difference in the fertility rate of the two groups. The overall pregnancy rate was 43.7% vs. 41.6% with 33.3% vs. 29.1% term pregnancies in the laparotomy and microsurgery group and laparoscopy group, respectively.⁵⁸ A meta-analysis of six controlled studies comparing pregnancy outcomes of tuboplasty performed with laparoscopy or open microsurgery showed no significant difference in the intrauterine pregnancy rates between the two groups. When subgroup analysis was carried out according to the extent of tubal disease, there was no significant difference in the pregnancy rate in patients with mild tubal disease treated by laparoscopy or laparotomy and microsurgery. For patients with severe tubal disease, there was a significantly increased pregnancy rate in the laparotomy and microsurgery group.⁶

Adhesion reformation after reproductive pelvic surgery occurs frequently, even with the use of microsurgical techniques and surgical adjuvants.⁵⁹ Although adhesion reformation was defined in one report⁵⁹ as a redevelopment of adhesions at sites that were subjected to adhesiolysis at the time of the initial procedure, other authors have used the same term to refer to different situations, thereby creating considerable confusion in the literature.⁶⁰ To provide a common terminology for comparison of efficacy of different techniques and adjuvants in reducing postoperative adhesions, a classification system has been proposed that takes into consideration the presence or absence of adhesions initially, and whether a surgical procedure other than adhesiolysis was performed at that site (Table 3). Importantly, this classification also includes the concept of de novo adhesion formation (the development of adhesions, observed at the time of the second-look procedure, for which adhesiolysis was not performed at the time of the initial procedure). As shown in Table 4, de novo adhesion formation can be a frequent problem after microsurgery performed at laparotomy, occurring in 58% of available ovaries and at almost one third of all available sites in those affected.⁶¹

Table 3. Classification of de novo adhesion formation and adhesion reformation 

 (Diamond MP, Neshat F: Letter to the Editor: Adhesions alter resection of ovarian endometriomas. Fertil Steril 59: 934, 1993)

Table 4. De novo adhesion formation* 

 *Frequency of adhesion identified at early second-look laparoscopy at sites throughout the pelvis after reproductive pelvic surgical procedures performed at laparotomy.
(Diamond MP: Adhesion prevention. In Gershenson DM, DeCherney AH, Curry SL (eds): Operative Gynecology, Chap 8. pp 147–158. Philadelphia, WB Saunders Company, 1993)

⁶² Polymeneas and co-workers compared the incidence of adhesions postlaparoscopic vs. open cholecystectomy. The incidence of adhesions postlaparoscopy was 44.3%, mostly filmy adhesions, while the open cholecystectomy group had 100% incidence of thick and extensive adhesions.⁶³ Nezhat and colleagues reported the findings at second look laparoscopy on 157 patients who had laparoscopic laser treatment of endometriosis and adhesiolysis. There was a significant reduction in adhesion scores compared to the first laparoscopy and there was no de novo adhesion formation.⁶⁴ In another study 105 patients with tubal ectopic pregnancy were randomized for laparoscopy vs. laparotomy for surgical treatment. There were 73 patients (31 laparoscopy and 42 laparotomy) who had a strong desire for pregnancy and had a second look laparoscopy 12 week after their surgery. Patients who had laparotomy developed significantly more adhesions at the operated side (76%) than those who had laparoscopy (42%) and more patients in the laparotomy group underwent adhesiolysis at second-look laparoscopy than did patients in the laparoscopy group.⁶⁵ However, in a multicenter study evaluating adhesion reformation at a second-look procedure after laparoscopic adhesiolysis, adhesion reformation was identified in 66 of 68 subjects (97%).⁶⁶ Laparoscopic adhesiolysis was able to reduce significantly the extent of pelvic adhesions, to approximately half of what was present initially. De novo adhesion formation occurred in only eight (12%) of these 68 women, and in 11 (23%) of 47 available sites in those affected. These studies and others^{67, 68} suggest that de novo adhesion formation may occur less frequently after laparoscopic surgery. Whether laparoscopy decreases adhesion reformation needs further studies, but does not appear to be likely based on most existing reports as summarized by Wiseman and colleagues.⁶⁹

Surgical adjuvants have been used frequently after adhesiolysis and other types of intra-abdominal reproductive pelvic surgical procedures. The rationale for use of adjuvants is to minimize development of postoperative adhesions and to maximize maintenance of tubal patency after the initial surgical procedure. As shown in Table 5, a variety of agents have been used for this purpose. Their use has been described in detail in several reviews.^{18, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80}

Table5. Classes of surgical adjuvants used in an attempt to minimize the occurrence of postoperative adhesions

(Adapted with modification from Diamond MP, DeCherney AH: Pathogenesis of adhesion formation/reformation: Application to reproductive pelvic surgery. Microsurgery 8: 103–107, 1987)

Despite application of microsurgical technique and use of surgical adjuvants by experienced surgeons, the development of postoperative adhesions is an all too frequent occurrence.^{81, 82, 83, 84, 85, 86, 87, 88, 89, 90} Currently, no serum marker or scanning technique is consistently able to identify adhesions, and a repeat operative procedure is required for evaluation. As shown in Table 6, pelvic adhesions were identified at the time of second-look laparoscopy in 55–100% of women who had undergone reproductive pelvic surgical procedures. Such adhesions represent both adhesion reformation and de novo adhesion formation.^{67, 91} Inasmuch as such adhesions might impair the ability to conceive, there appears to be room for improvement in preventing postoperative adhesion development. In a poll of the members of the Society of Reproductive Surgeons, Holtz⁹² reported widespread use of agents for this purpose throughout the United States, including perioperative antibiotics, corticosteroids, and antihistamines and intraoperative instillation of 32% dextran 70 (Hyskon). The literature is replete with conflicting results for most adjuvants, and their efficacy is not well established. A potential explanation is differences in the pathologic conditions being treated.

Table 6. Pelvic adhesions noted as second-look procedures

Three groups have demonstrated fundamental differences between adhesion formation and adhesion reformation in animal models. Holtz and associates^{93, 94} described reduction in adhesion formation with 32% dextran 70, but a similar inhibition of adhesion reformation could not be achieved with higher doses of dextran. Similarly, Elkins and coworkers^{95, 96} observed a greater extent of adhesion reformation than adhesion formation after dextran treatment. Finally, Diamond and associates^{97, 98} compared adhesion formation and reformation models and noted a greater extent of adhesions in the latter.

In the survey of the members of the Society of Reproductive Surgeons, Holtz⁹² reported that 45% performed second-look laparoscopy on occasion. There are several benefits of second-look procedures after reconstructive pelvic surgery. Such procedures allow for assessment of the adnexal disease, allowing patients to assess their potential for fertility and plan accordingly. Second, they provide an opportunity to lyse adhesions that have formed or reformed (or to perform additional procedures as needed, such as vaporization of endometriosis or deagglutination of fimbria). Third, based on the previously established principles that pelvic adhesions can cause infertility^{34, 36, 52} and that lysis of adhesions is associated with pregnancy establishment in many women,^{34, 36, 52} performance of second-look laparoscopy may improve the pregnancy rate after reproductive pelvic surgery, although this theoretical improvement has yet to be substantiated. Trimbos-Kemper and colleagues⁸⁶ and Jansen⁹⁹ reported that second-look laparoscopy was associated with a significant reduction in the subsequent observation of permanent pelvic adhesions, but the pregnancy rate was not improved. Trimbos-Kemper report⁸⁶ did, however, note a reduction in the occurrence of ectopic pregnancies after salpingostomies in women who underwent early second-look laparoscopy. On the other hand, in a nonrandomized study Ozaki and co-workers reported on 92 women who underwent pelvic surgery for infertility. A second look laparoscopy with lysis of adhesions was performed within 10 days from the pelvic surgery in 60 patients. The cumulative spontaneous pregnancy rate at 5 years was significantly higher in the second look laparoscopy group (39%) compared to control group (9.7%).¹⁰⁰ Finally, for the benefit of the surgeon, performance of second-look laparoscopy allows for feedback as to the surgical success of the procedures previously performed.

If second-look laparoscopy is to be performed, Swolin¹⁰¹ recommended that it be done early (6–8 weeks) to improve the possibility of lysis of postoperative adhesions. Subsequently, Raj and Hulka¹⁰² examined second-look laparoscopies performed up to 2 years after the initial procedure and demonstrated that bleeding was more common if the procedure was performed after 12 weeks or before 2 weeks of the initial procedure. In the former case, bleeding was attributed to increased density and vascularity of the adhesions; in the latter, bleeding was attributed to granulation tissue. Between 2 and 12 weeks, the adhesions were filmy and more amenable to lysis. Surrey and Friedman⁸⁴ observed that attempted adhesiolysis 6 months after the initial procedure was associated with more dense and vascular adhesions than that at 6–8 weeks. DeCherney and Mezer⁸² also compared early (4–16 weeks) versus late (approximately 18 months) second-look laparoscopy; in the latter group, 63% of patients had thicker neovascular adhesions. Finally, Daniell and Pittaway,⁸⁷ McLaughlin,⁸³and Diamond and coworkers⁸¹ have also observed that adhesions at “early” second-look laparoscopy are more likely to be filmy and avascular, making them more susceptible to easy lysis. However, in multiple studies involving an opportunity for early second-look laparoscopy after the initial surgical procedure, there appeared to be no difference in the type (filmy and avascular versus dense and vascular) of adhesions observed at second-look laparoscopy.⁵⁴

Distal tubal obstruction: complete

The surgical treatment of complete fimbrial occlusion is a neosalpingostomy: the creation of a new tubal ostium. Performance of neosalpingostomy is associated with a subsequent tubal patency rate of approximately 90–95%;¹⁰³however, pregnancy rates are only about 30% (Table 7).^{16, 44, 46, 47, 48, 49, 51, 53, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117} With prolonged follow-up of 5 years, pregnancy rates of 42% have been achieved.¹¹⁸ Comparison of series of macrosurgical and microsurgical neosalpingostomies fails to demonstrate an advantage of microsurgery.^{119, 120}Again, however, a study designed to directly compare the two techniques has not been performed. Both Mage and Bruhat⁴⁹ and Tulandi and Vilos¹⁰⁸ have compared microsurgical neosalpingostomies performed by laser and nonlaser techniques. Each group reported no difference in pregnancy outcome. In a Cochrane review in 2000 Watson and his colleagues concluded that utilizing laser in the treatment of distal tubal disease did not improve pregnancy rates.¹²⁰

Table 7. Pregnancy outcome after surgical treatment of complete distal tubal obstruction

Performance of neosalpingostomy at laparoscopy has been reported by several groups.^{42, 46, 49, 110, 111} Early reports showed lower pregnancy rates compared to laparotomy, however, later reports showed comparable results. The patient selection for neosalpingostomy at laparoscopy in the early reports includes women who refuse laparotomy, and therefore the patient populations may not be comparable. Furthermore, the follow up period in these early reports was relatively short. A meta-analysis of six studies comparing laparoscopy with laparotomy and microsurgery for treatment of distal tubal disease showed no significant difference in the overall intrauterine pregnancy (IUP) rate between the two groups. Subgroup analysis was carried out according to the extent of tubal damage showed no significant difference in the IUP rate with mild tubal disease, whereas the laparoscopy group had a significantly lower IUP rate with severe tubal disease, OR 0.34 (95% CI 0.14–0.86).⁶

The relatively poor pregnancy outcome after neosalpingostomy, compared with other types of reconstructive pelvic surgical procedures, and despite re-establishment of tubal patency in most patients, is thought to be caused in large part by intraluminal tubal pathology. The presence of coexistent pelvic adhesions, which restrict tubal motility and ovum capture, has also been implicated in reduction of successful pregnancy outcome.

Intraluminal tubal abnormalities primarily represent the pathologic sequelae of infection, although changes can also occur in response to endometriosis or exposure to diethylstilbestrol (DES)¹²¹ in utero. Chronic postinfectious histologic intraluminal changes observed include the presence of intraluminal adhesions, loss of villous folds, and ciliary destruction. It is unclear whether the latter is caused by the increased hydrostatic pressure resulting from tubal occlusion and hydrosalpinx or is an effect of the infectious process itself.

It has long been observed that the pregnancy outcome after neosalpingostomy depends on the extent of tubal disease. Several authors proposed classification systems that correlate the extent of tubal disease with pregnancy outcome. These classifications are based on intratubal and extratubal characteristics. Screening for these characteristics, to assess the likelihood of establishing a pregnancy, is performed by hysterosalpingography and laparoscopy, respectively. Rock and colleagues¹²² were among the first to develop a classification of distal tubal obstruction. Their classification is based on hydrosalpinx diameter, rugal pattern, extent of fimbrial damage, and extent of coexistent pelvic adhesions (Table 8); they were able to demonstrate a deterioration in outcome as the extent of disease progressed from mild to moderate to severe.

Table 8. Rock classification of the Eextent of tubal disease with distal tubal obstruction

 (Rock JA, Katayama KP, Martin EF: Factors influencing the success of salpingostomy techniques for distal fimbrial obstruction. Obstet Gynecol 52: 597, 1978)

The American Fertility Society proposed a classification of distal tubal occlusion that accounted for distal ampullary dilation, tubal wall thickness, mucosal folds at neostomy site, extent and type of adhesions.^41, Patients with severe forms of tubal disease have diminished pregnancy rates after neosalpingostomy and may benefit from direct referral for ART instead of surgery.

After surgical correction of complete distal tubal obstruction, the most likely time to conceive is more than 1 year after the surgical procedure.^{48, 104, 123} This time lag is thought to be required because the tubal mucosa needs to regenerate after the damage that occurred while a hydrosalpinx was present.

Salpingoscopy and fimbrial biopsy can help predict the likelihood of establishing a pregnancy after distal tubal obstruction treatment. In the former technique, a narrow-diameter scope or fiber is passed into the fimbrial ostium at the time of laparoscopy or laparotomy, or through the tubal cornua, which is visualized hysteroscopically. This allows direct examination of the intraluminal anatomy and identification of adhesions, areas of deciliation, and flattening of the mucosal folds. The findings with this technique correlated significantly with subsequent pregnancy outcomes and therefore may be helpful in referral of patients for tubal repair or ART.^{124, 125, 126, 127, 128, 129, 130, 131} This may particularly be the case as technologies for transcornual viewing are improved, making this an office procedure. Therefore, salpingoscopy may be beneficial in preoperative patient screening. Regarding fimbrial biopsy, Donnez and associates⁴⁸ showed that the ciliation index obtained on microbiopsy was impaired in the presence of tubal disease and that the index was predictive of subsequent pregnancy outcome.

Distal tubal obstruction: incomplete

Comparison of outcome after surgical treatment of incomplete distal tubal obstruction is difficult because of the diversity of pathology it may represent. The clinical findings range from slight fimbrial agglutination to peritubal bands to fimbrial phimosis causing near-total occlusion and can occur in association with varying degrees of pelvic adhesions. Therefore, analysis of efficacy of treatment of incomplete distal tubal obstruction must be carefully scrutinized. A summary of pregnancy outcomes after such procedures is shown in Table 9.^{48, 51, 53, 112, 132, 133, 134} These results tend to be better than those for complete distal obstruction, probably reflecting less intratubal damage. Although follow-up is limited, surgical treatments with the CO₂,¹³³ argon,⁴² KTP-532,¹³⁵ and Nd:YAG¹³⁶ lasers have been described; their use has not been associated with improved pregnancy rates. The effect of laparoscopic treatment of incomplete distal tubal obstruction on pregnancy outcome is comparable to that obtained with laparotomy, however, a study designed to compare the two surgical approaches was not performed. 

Table 9. Pregnancy outcome after surgical treatment of partial distal tubal obstruction

Endometriosis

The monthly fecundity rate in normal couples ranges between 15 and 20% and decreases with age, while in untreated women with endometriosis and infertility the monthly fecundity rate is 2–10%.¹³⁷ Endometriosis can contribute to infertility in more than one way. It could be a direct cause through formation of adnexal adhesions which form a mechanical barrier or otherwise impair tubal ovum pickup, and subsequently sperm–oocyte interaction. Endometriosis can also have effects on ovarian steroidogenesis, ovulation, tubal ovum pickup, tubal motility, tubal ciliary action, peritoneal macrophage activity, and implantation. In part, these effects are thought to be mediated by cytokines and prostaglandins.

Various systems have been used to classify endometriosis.^{138, 139, 140, 141} Most early reports used the Acosta classification,¹³⁸grouping cases into mild, moderate, and severe stages based on the location and extent of endometriosis and the degree of pelvic adhesions. Classifications by Kistner and associates¹³⁹ and by the American Fertility Society (now called the American Society for Reproductive Medicine) were subsequently established,¹⁴⁰ followed by a revised classification by the latter group in 1985,¹⁴¹ which is pictured in Figure 1. Use of classifications is advantageous because the scoring confers an extent of disease, allowing comparison of reports from different investigators and centers.

The relative efficacy of different methods of treating endometriosis is not well defined; in some reports, pregnancy outcome was no different with observant management than it was with medical or surgical treatment.^{142, 143} In a randomized control trail 341 infertile women with minimal or mild endometriosis were randomized to diagnostic laparoscopy with no treatment or operative laparoscopy and ablation or excision of endometriotic implants with lysis of adhesions. Patients were followed up for 36 weeks. The treatment group had a significantly higher pregnancy rate than the control group (30.7% vs. 17.7%, p=0.006).¹⁴⁴ In a multicenter trail, Parazzini and colleagues randomized¹⁴⁵ infertile women with minimal or mild endometriosis to laparoscopic treatment or observation. The follow up period was 1 year. There was no significant difference in the pregnancy rate between the treatment and the control groups (24% vs. 29%, respectively).¹⁴⁵ It should be noted, though, that in this multicenter trial the protocol allowed the use of medical treatment (gonadotropin-releasing hormone [GnRH] agonist) after surgery in both groups  according to the physician’s judgment. In a Cochrane review, a meta-analysis of the above two studies showed a significant increase in pregnancy rate after laparoscopic treatment of minimal or mild endometriosis in infertile women (OR 1.64, 95% CI 1.05–2.57).¹⁴⁶

The most widely used medical agents are oral contraceptive pills and progestins. Danazol^{147, 148}and GnRH analogues^{149, 150} are currently considered second line therapy for endometriosis.¹⁵¹ These agents are used alone for treatment of mild and moderate endometriosis, and in combination with surgery (both preoperative and postoperative) for treatment of these stages as well as severe endometriosis. Whether medical treatment of endometriosis provides a favorable outcome in the management of infertile woman with endometriosis is still uncertain.¹⁵² An in-depth discussion of studies describing the medical and medical-surgical treatment of endometriosis is beyond the scope of this chapter; the reader is referred to published reports.^{147, 148, 151, 153, 154, 155,}

Pregnancy outcome after treatment is the most important end point for infertile couples; an assessment of outcome is given in Table 10.^{133, 138, 144, 145, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171} Considering the use of laparoscopy for reproductive pelvic surgery, more series have been described for the treatment of endometriosis than for other procedures. From the numbers in these reports (which are limited in part by lack of comparable life table analyses of outcome), it appears that, in experienced hands treatment at laparoscopy and at laparotomy are equally efficacious, not only for mild and moderate endometriosis but for severe and extensive endometriosis as well. This observation is consistent with an extensive meta-analysis of equivalent pregnancy outcome after treatment at laparoscopy and at laparotomy.^170, Individual investigators^{172, 173} have treated large endometriomas laparoscopically with high pregnancy rates resulting. In many series of laparoscopic treatments for endometriosis, the CO₂ laser has been used.^{160, 161, 162, 163, 164, 165, 166, 174} However, similar pregnancy rates obtained with nonlaser laparoscopic techniques have also been described.¹⁶⁷ Other lasers that have been used for laparoscopic treatment of endometriosis are the argon,¹⁷⁵KPT-532,^{135, 171} and Nd:YAG^{176, 177}types. Nevertheless, pregnancy outcome as an end point has its limitations, in that many factors can impair fertility after adequate treatment of endometriosis; furthermore, pregnancy can occur after inadequate treatment of endometriosis and pregnancy can occur despite recurrence or persistence of endometriosis. Several investigators have tried to correct for other possible factors by looking at pregnancy outcome in women in whom endometriosis is the sole cause identified for infertility; slight improvements in pregnancy outcome have been identified.³⁷

Table 10. Pregnancy outcome after surgical treatment of endometrios

Lap, Laparotomy; Lsp, Laparoscopy.

 † Acosta.

 • American Fertility Society.

•• Revised American Fertility Society.

Endometriosis is noted to be present in 2–4% of women after surgical or medical treatment,¹⁷⁸ the actual likelihood, in part, depending on the initial severity of endometriosis and the treatment. However, it is difficult to differentiate between recurrence and persistence of endometriosis. This is clearly based on three observations suggesting persistence may be more likely than previously recognized. First, microscopic implants of endometriosis not visible to the naked eye (or the eye with magnification) are frequently present.^{179, 180, 181} Second, many implants do not have the gross purplish macroscopic appearance that has become clinically recognized as endometriosis but may be clear, yellow, red or black, or represented by converging vascular patterns or white scarring.¹⁸² Third, recognized implants of endometriosis may extend from the identified lesion beneath the peritoneum.¹⁸³ As a method to treat the latter, it has been suggested that visualized lesions be treated so as to include the immediately surrounding (normal appearing) peritoneum. The actual surface area that should be treated, and whether such treatment affects recurrence/persistence rates or pregnancy outcome has, however, not been established.

The extent of recurrence/persistence of endometriosis is best assessed at second look laparoscopy; however, this requires performance of an additional operative procedure. Clearly, a nonsurgical method of assessing recurrence/persistence would have tremendous advantages. CA 125 is a membrane antigen originally identified in many conditions including epithelial ovarian carcinomas.^{184, 185} This antigen is elevated in women with endometriosis, and initial studies suggest the levels correlate with disease and its clinical course.^{184, 185} CA 125 is a nonspecific antigen, its role in the diagnosis of endometriosis or its recurrence is still controversial. A meta-analysis of 23 studies of the value of CA 125 in the diagnosis of endometriosis showed its limited performance in the diagnosis of mild disease.¹⁸⁶ Although CA 125 was better in diagnosing moderate to severe disease, the sensitivity was still low.¹⁸⁶ Another potential serum marker that has been evaluated is an endometrial antibody assay.¹⁸⁷ In early studies, the antibody level did appear to correlate with observations at second-look laparoscopy. Follow up studies either failed to detect anti-endometrial antibodies in the serum of patients with endometriosis,¹⁸⁸ or showed no difference in antibody level between women with endometriosis and women without endometriosis, men or newborns.^{189, 190} Furthermore, levels of the antibody in women with endometriosis did not change after therapy.¹⁸⁹ For the detection of endometriomas, transvaginal ultrasonography is both highly sensitive and specific.¹⁹¹ Transvaginal ultrasonography is superior to CA 125 in diagnosing recurrence of endometriomas,¹⁹² however, it is of little value in diagnosis of nonovarian endometriosis.¹⁹¹ Additional studies are required to further examine the potential of these markers.

Tubal anastomosis/implantation for tubal obstruction

An overview of pregnancy outcome following tubal anastomosis and tubal implantation procedures is shown in Table 11.^{9, 12, 17, 44, 47, 51, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204} Compared to other types of reproductive pelvic surgical procedures, pregnancy outcome is best following tubal anastomosis, approaching 70% in many series. Many factors have been identified that affect the outcome of these surgical procedures designed to reestablish tubal continuity. These factors include the cause of the obstruction; its location; the extent of coexistent pelvic disease; the length of the resulting tube after restoring patency; and the length of time since tubal blockage. 

Table 11. Pregnancy outcome after tubal anastomosis/tubal implantation 

Among women desiring reversal of previous tubal ligations, the method of tubal occlusion is a major prognostic factor in subsequent pregnancy outcome. The best results are obtained with Hulka clips or fallope rings; the worst outcome follows unipolar cautery. Intermediate pregnancy rates are achieved following segmental resection or bipolar cautery. These differences may be partially explained by the resultant length of tube after anastomosis; increasing tubal length correlates directly with improved pregnancy outcome. An additional factor may be the extent of damage to the tubal blood vessel arcade; both human and animal studies have suggested a possible role of this network in ovulation and pregnancy outcome. In tubocornual anastomosis, an additional advantageous prognostic factor is preservation of as much of the intramural segment of the fallopian tube as possible.¹⁹⁷

The length of time between tubal ligation and reversal has been shown to be a prognostic factor. Time inversely correlates with subsequent pregnancy outcome. DeCherney and coworkers¹² examined the causes of failure to conceive in 35 women who underwent midsegment tubal anastomosis. A surgically related factor was noted in only 20% of the couples (bilateral tubal reocclusion). The remaining couples were noted to have ovulatory dysfunction, a husband who had not previously fathered a child, or had unexplained infertility.

Location of tubal occlusion is another major prognostic factor. Those women with tubal blockage that allows reanastomosis to be performed generally have improved conception rate as compared to those requiring tubal implantation. In their in-depth review of surgical management of utero-tubal obstruction, Musich and Behrman²¹² concluded that microsurgical anastomosis has replaced implantation as the primary method for re-establishing tubal patency under most circumstances. Exceptions were surgeon preference and complete intramural occlusion. Additionally, in a review of all series in the literature from 1965 to 1982 describing pregnancy outcome after uterotubal implantation and tubocornual anastomosis, Haseltine and Von Attas²¹³ demonstrated improved pregnancy rates following anastomosis. Additionally, anastomosis obviates the decision of whether a conception following a tubal implantation procedure should always be delivered by cesarean section, and reduces the risk of uterine rupture at the site of implantation, a complication that has been described with the implantation technique.²¹⁴

There are multiple reports of tubal anastomosis performed through laparoscopy with a term pregnancy rate comparable to that achieved with laparotomy and microsurgery.^{205, 206, 207, 208, 209, 210, 211} The reported laparoscopic techniques differed from a two layers four suture anastomosis (similar to laparotomy and microsurgery),²⁰⁷ to a single suture anastomosis (suture placed at the antimesentaric side of the tube),²⁰⁵ to sutureless anastomosis using fibrin glue.²¹⁰ Few reports compared laparoscopic anastomosis to laparotomy and microsurgery.^{209, 210} Wiegerinck and colleagues²¹⁰ compared 41 laparoscopic tubal sterilization reversal to 41 age matched cases performed by laparotomy and microsurgery. The reported ongoing pregnancy rate in laparoscopy group (37%) was lower than that at laparotomy (59%), however, the period of follow up was shorter for the laparoscopy group. Kaplan−Mayer curves showed no statistically significant difference in intrauterine pregnancy rate at 3 years (45% and 52% for laparoscopy and laparotomy groups, respectively). Laparoscopic anastomosis was associated with shorter hospital stay and quicker recovery.

At the turn of the century, robotic assistance for laparoscopic tubal anastomosis was first reported by Falcone and co-workers.²¹⁵ Following that there were several small case series published.^{211, 216, 217, 218} The results were similar to laparoscopic or open microsurgical anastomosis, however, the robotic assisted procedures required significantly longer operative time.

Women who have undergone fimbriectomy for sterilization represent a special category because the surgical correction required is performance of a neosalpingostomy. In two small series of such women, pregnancy occurred in four of seven (57%)¹⁶ and four of nine (44%) patients.²¹⁹ Over a period of 10 years, Tourgeman and his colleagues performed neosalpingostomy for reversal of fimberiectomy sterilization on 41 women with six achieving pregnancy (15%).²²⁰ Women who conceived had significantly longer fallopian tubes at the end of the procedure than those who did not conceive.

Use of lasers in the performance of tubal anastomosis has seen limited application. Kelly and Roberts¹⁰⁹ described excising the blocked tubal stumps using the CO₂ laser. It is unclear whether this has any advantage over nonlaser techniques. Other investigators have attempted to anastomose the fallopian tube by “welding” using CO₂ laser, but have noted this technique to be of little value.²²¹

The incidence of ectopic pregnancy has increased almost five fold over the past three decades with current estimates at 2% of all pregnancies in the United States.²²² Ectopic pregnancy remains the leading cause of maternal mortality in the first trimester and accounts for 9% of all pregnancy-related deaths.²²² The apparent increase in the incidence of ectopic pregnancy could be a result of improved diagnostic aids, alteration in sexual activity with increased prevalence of chlamydia and gonorrhea infection or perhaps due to the advances that have occurred in the performance of reproductive pelvic surgery. With regard to the latter, perhaps the current ability to establish and maintain tubal patency in damaged tubes allows conception to occur in women who previously would have been sterile, but the extent of the tubal damage precludes (or reduces) the ability of the fertilized ovum to traverse the tube and reach the uterine cavity. While only a theory, such a possibility is pertinent because of the issue of whether gamete intrafallopian transfer (GIFT) procedures should be performed on women with damaged tubes. There was no observed increase in the incidence of ectopic pregnancy when GIFT was compared with IVF²²³ or intrauterine insemination (IUI)²²⁴ in infertile couples, however, whether women with surgically corrected tubal factor infertility undergoing GIFT have a higher incidence of ectopic pregnancy is unknown. 

In the past, ectopic pregnancy was a diagnosis done at the time of surgery. With the advent of transvaginal ultrasonography, a noninvasive diagnosis of ectopic pregnancy can be made with a sensitivity more than 90% and specificity approaching 100%.²²⁵ With the current availability of high resolution transvaginal ultrasound and highly sensitive human chorionic gonadotropin (β-hCG) assay, ectopic pregnancy can be diagnosed at an earlier stage. Once tubal ectopic pregnancy is diagnosed, the options for treatment include expectant management, medical treatment with methotrexate or surgical intervention. Expectant management is an option when the patient exhibits no symptoms, is reliable and willing to have close follow up. The reported success rates in the literature for expectant management of ectopic pregnancy varied between 48 and 100%.²²⁶ A low starting β-hCG level, a decreasing trend in β-hCG levels and a longer period of amenorrhea are associated with higher success rate.²²⁶ A follow up study of 30 women with successful expectant management of ectopic pregnancy reported patency of the affected tube in 93% of cases and an 88% pregnancy rate.²²⁷ Medical treatment is another option that has gained popularity in the past 2 decades. Medical treatment consists of administration of a single dose of methotrexate (commonly 50 mg/m²) systematically or locally. Less commonly, methotrexate is administered in a multiple dose regimen (1 mg/kg on days 1, 3, and 5). Medical treatment is a viable option when the patient is asymptomatic and is willing to follow up with β-hCG levels post-treatment. Table 12 summarizes reported contraindications for medical treatment of ectopic pregnancy. The reported success rate of the single dose methotrexate regimen varies between 65 and 95% in well selected cases with 3–27% of cases requiring an additional dose.²²⁸

Table 12. Reported contraindications for medical treatment of ectopic pregnancy 

Ectopic pregnancy patients need surgical intervention when they have failed medical or expectant treatment, are symptomatic, hemodynamically unstable, or non-reliable to comply with follow up. Laparoscopy is frequently utilized for the surgical treatment of ectopic pregnancy. With the wide spread use of laparoscopy the conversion rate to laparotomy ranges between 12.2 and 18.2%.^{229, 230} Presence of pelvic adhesions was the strongest predictor for conversion with an odds ratio of 607.²³⁰ There is no contraindication for laparoscopy in ectopic pregnancy treatment. Several studies have shown that laparoscopy is a valid approach with ruptured ectopic pregnancy and presence of large hemoperitonium when conducted by an experienced surgeon with proper operating room set up and prepared staff.^{231, 232} Laparoscopy is safe and feasible in obese^{233, 234} and morbidly obese²³⁵ women with ectopic pregnancy. Laparotomy still remains the primary approach when the surgeon is inexperienced in operative laparoscopy, the necessary laparoscopic equipment is not available or there is no adequate pelvic access. Regardless of the mode of intervention, the options for surgical treatment are whether to perform a conservative intervention and preserve the affected fallopian tube or a radical intervention and remove the tube. This decision will be based on the location of the ectopic pregnancy, the extent of tubal damage, the patient’s desire to preserve fertility and willingness to follow up with β-hCG levels to rule out persistent ectopic pregnancy. Isthmic, ampullary, and infundibular ectopic pregnancies are more likely to be amenable to conservative management. Isthmic ectopic gestations are usually treated by segmental resection, with possible reanastomosis of the two segments at the same or usually a later procedure. Tubal ampullary and infundibular gestations are able to be treated by linear salpingostomy along the antimesenteric border. Although an early series described no persistent ectopic pregnancies in 79 sequentially treated ampullary ectopic gestations,²³⁶ the reported persistence rate in literature varies between 10 and 11.5%.^{237, 238} The only two factors able to be identified that were associated with persistent ectopic pregnancies were smaller ectopic size and younger gestational age.²³⁷ A summary of pregnancy outcome following treatment of ectopic pregnancies among selected series is shown in Table 13.^{239, 240, 241, 242, 243, 244, 245}

Table 13. Pregnancy outcome after treatment of tubal ectopic pregnancy

Women who have had an ectopic gestation have an 8–13% risk of repeat ectopic gestation.^{240, 246, 247, 248} This risk is independent of whether the initial ectopic pregnancy was treated conservatively or by salpingectomy. In the groups treated conservatively, it is equally likely to occur in the tube in which the initial ectopic occurred as in the contralateral tube.²⁴⁹ This suggests that the processes that contributed to the initial ectopic (whether pathologic disease or innate variants in tubal motility) are likely to affect each fallopian tube.

Among women with two prior ectopic pregnancies, the risk of a third ectopic gestation increases to 20%.²⁵⁰ Furthermore, among women with an ectopic pregnancy in the sole remaining fallopian tube treated conservatively, the subsequent pregnancy rate ranges in most series from 32% to 60%, with repeat ectopic pregnancy rates of 12–40%.²⁴³ Additionally, women who have undergone a segmental resection for treatment of an ectopic pregnancy and have a patent contralateral tube are at risk for development of repeat ectopic pregnancy in the blind-ending distal segment of the partially resected tube.²⁵¹

Prior to the widespread availability of agents for induction of ovulation in women with polycystic ovary syndrome (PCOS), ovarian wedge resection was a common method by which to attempt to achieve ovulation. In theory, such a procedure reduced ovarian steroid production, thus, temporarily allowing follicular development to resume. In one series, after wedge resection, resumption of ovulation occurred in 91% of women; however, subsequent conception occurred in only 48% of women.²⁵²

A potential cause for the discrepancy between ovulation and conception rates is the postoperative development of pelvic adhesions. Kistner²⁵³ noted peritubal and periovarian adhesions at culdoscopy in 16 women after wedge resection. Weinstein and Polishuk²⁵⁴ reported that 14% of women undergoing wedge resection and subsequent second-look developed pelvic adhesion. This number undoubtedly represents an underestimation, as only a select nonpregnant group underwent the second-look procedure. Buttram and Vaquero²⁵⁵ and Toaff and co-workers²⁵⁶ reported adhesion observance in 59 of 59 and seven of seven women, respectively.

Open ovarian wedge resection is frequently replaced by a laparoscopic procedure that aims at reducing the androgen levels in the ovaries by draining ovarian cysts (or follicles) using electrocautery or other techniques. This procedure is known as “laparoscopic ovarian drilling”. Eight to 12 puncture holes are made with a laparoscopic electrocautery needle in each ovary. It is not known whether the drainage of follicular fluid or the tissue destruction causes a drop in the ovarian hyperandrogenic milieu and resumption of ovulation. Using this technique, ovulation occurred in 57 of 62 (92%) women.²⁵⁷ Of 35 women with PCOS as sole cause of infertility, a conception occurred in 24 (69%) after electrocautery treatment. In two large series, Gjonnaess²⁵⁸ and Naether and colleagues²⁵⁹ reported presumptive spontaneous ovulation rates of 82% and 92%, and pregnancy rates of 69% and 70%, respectively. This procedure has also been performed using both CO₂ laser¹³⁵ and argon laser.²⁶⁰ There are a few studies that evaluated the development of postoperative adnexal adhesions with conflicting results. The incidence ranged from as low as 27% to as high as 100%.^{261, 262} Studies with a high incidence of adhesions still reported a high pregnancy rate.²⁶² The effects of laparoscopic ovarian drilling seem to be long lasting. Gjonnaess followed up 51 anovulatory PCOS women up to 20 years after laparoscopic ovarian drilling and reported at least two thirds of the patients were still ovulating and the majority of the patients had persistence of decreased androgen and luteinizing hormone levels.²⁶³ Recently, ultrasound guided transvaginal ovarian needle drilling was described as a substitute to laparoscopic electrosurgery. A total of 82 clomiphene-resistant PCOS patients randomized to in-office ultrasound guided transvaginal ovarian needle drilling were compared to 81 clomiphene-resistant PCOS patients randomized to laparoscopic drilling and followed up for 6 months. There was no significant difference in the ovulation rate (49% vs. 56%) or pregnancy rate (22% vs. 25%, respectively).²⁶⁴

Two operations which women of reproductive age have frequently undergone are appendectomy and cesarean section. Several reports have examined the effects of these procedures on subsequent infertility. Trimbos-Kemper and associates²⁶⁵ noted a 42% incidence of tubal abnormalities in women who had undergone appendectomy; however, this was not increased from the 37% incidence in the control group. Mueller and co-workers²⁶⁶ observed that women who underwent appendectomy with findings of an unruptured appendix were not at increased risk for infertility. However, those women with a ruptured appendix at the time of appendectomy were significantly more likely to be infertile. In a cohort study from the Swedish general population, Andersson and coworkers compared the fertility outcome in 9840 women who had appendectomy in childhood with 47,590 age matched controls. Women with a perforated appendix had similar fertility outcome to controls, while those whose appendix was found to be normal at appendectomy had a better fertility outcome than controls. These findings made the authors conclude that appendectomy did not have a negative effect on fertility.²⁶⁷

Among American women who underwent cesarean section for their first childbirth, Hemminki and colleagues²⁶⁸ described a reduction in subsequent fertility as compared to controls. This reduction was attributed to physical difficulties related to the operation, and potentially may have been caused by the cesarean section. Hemminki²⁶⁹ subsequently confirmed these findings in a report based on the Swedish Birth Registry, again demonstrating that women who undergo cesarean section have fewer subsequent children. These studies, however, did not address whether the reasons for the reduction in fecundity was voluntary or related to factors that led to the cesarean section. In the international randomized Term Breech Trial, the only trial where patients with term breech singleton pregnancies were randomized to have elective cesarean section or vaginal delivery, there was no difference in the fecundity of women randomly assigned to the cesarean section group and those in the vaginal delivery group up to 2 years after their delivery of the breech.²⁷⁰

A variety of congenital and acquired abnormalities of the uterus have been suggested as factors possibly contributing to the establishment of an infertile state. Several of these conditions and advances in their surgical management are described. For a summary of uterine distention media used for hysteroscopy, the reader is referred to a recent review.²⁷¹

Asherman’s syndrome

Asherman’s syndrome is the presence of adhesions within the uterine cavity. The adhesions partially or completely obliterate the uterine cavity, and are diagnosed by saline infusion sonogram (SIS), hysterosalpingogram (HSG), hysteroscopy, or a combination of either SIS or HSG and hysteroscopy. Obliteration of the uterine cavity by adhesions causes amenorrhea, in its most severe forms, and is a contributing factor for infertility. The most common cause of intrauterine adhesions is sharp curetting of a gravid or recently gravid uterus. Adhesion formation is noted at hysteroscopy in 8–19% of women who had suction, dilation and curettage for the treatment of spontaneous abortion.^{272, 273} Uterine infection is another common cause of intrauterine adhesion formation and when pelvic tuberculosis is present, the treatment outcome is poor.²⁷⁴ If these women with intrauterine cavity adhesions are able to conceive, only a small percentage are able to carry the pregnancy to term; this percentage increases markedly after lysis of intrauterine adhesions.^{275, 276, 277} Treatment of these adhesions currently is at the time of hysteroscopy. While lysis can be performed bluntly with the hysteroscope, adhesions in the noncentral sections of the uterine cavity can be missed, alternative therapy is by dilation and curettage, mechanical dissection using hysteroscopic forceps or scissors, electrosurgery, or lasers. The American Fertility Society classified intrauterine adhesions into three different stages, mild, moderate, and severe, depending on the extent of cavity involved (<1/3, 1/3–2/3, >2/3), type of adhesions (filmy or dense), and menstrual pattern (normal, hypomenorrhea or amenorrhea). The postadhesiolysis pregnancy rate decreases with the severity of adhesions.²⁷⁸ Pregnancy outcome following surgical treatment of Asherman’s syndrome is shown in Table 14.^{275, 277, 279, 280, 281, 282, 283, 284, 285, 286} The relative efficacy of the different methods of lysing the intrauterine adhesions has not been examined. Two studies reported increased abnormal placentation in pregnancies achieved after hysteroscopic adhesiolysis with increased incidence of placenta accreta (5–10%) requiring cesarean hysterectomy or hypogastric artery ligation.^{284, 285}

Table 14. Pregnancy outcome after treatment of Asherman’s syndrome in women attempting pregnancy

    * Term or ongoing.

  ** All patients had severe intrauterine adhesions.
 

At the completion of these operative procedures, particularly with more extensive adhesions, either an intrauterine device or a pediatric foley is placed in the uterine cavity. This reduces the apposition of the anterior and posterior uterine walls, thereby reducing redevelopment of the adhesions. Most reports also describe administration of estrogen to these women. However, there is not enough evidence to support the use of these preventive measures.²⁸⁷ Recently, the off label intrauterine use of adhesion prevention barriers like Seprafilm was reported to be effective in preventing recurrence of intrauterine adhesions.²⁸⁸

Uterine septum

Removal of a uterine septum has traditionally been performed at laparotomy using either the Jones or Tompkins technique. The rationale for performance of this operative procedure is that the septum might contribute to infertility or early miscarriage if the embryo were to “implant” on the septum; poor survival is attributed to inadequate blood supply to the uterine septum.²⁸⁹

While such elimination of the septum reduces the likelihood that a pregnancy will miscarry, it is likely that adhesions that develop at laparotomy and hysterotomy will reduce the actual conception rate. Thus, prior to performing a metroplasty at laparotomy, women have been given the opportunity to conceive and carry pregnancies. Additionally, since infertility is rarely caused by uterine anomalies alone,²⁹⁰ these couples should undergo the same rigorous evaluation as infertile couples without uterine anomalies.

Currently, uterine septum resection is more commonly performed through hysteroscopy. The procedure involves incision of the septum using hysteroscopic scissors, resectoscope, or Nd:YAG laser. The procedure is invariably performed concomitantly with laparoscopy to confirm the presence of septate uterus as opposed to bicornuate uterus. Although difficulties identifying both cavities may occur, complications in experienced hands have been minimal. The septum has been thought to contain fibrous tissue with no blood vessels, though, a recent report by Kraemer and co-workers²⁹¹ on Doppler flow in 66 septate uteri showed evidence of blood flow into the septum in 41% of the studied uteri. The classic teaching for hysteroscopic resection of a septum is that the surgeon starts cutting the septum from its lower edge up towards the fundus at its thinnest level and stops when bleeding is encountered which is a sign that the myometrium has been reached at the fundus. The surgeon could also use the laparoscopic light as a guide for the limit of excision. When the brightness of the laparoscopic light is clearly seen across the whole fundus excision should stop. Pregnancy rates using these techniques or abdominal metroplasty are listed in Table 15^{289, 290, 292, 293, 294, 295, 296, 297, 298, 299, 300, 301, 302, 303, 304, 305, 306} and appear to be similar to success rates to those of conventional metroplasty. Additionally, the hysteroscopic procedures are done as an outpatient procedure, without the morbidity associated with laparotomy. In a report comparing outcome in one center, Fayez²⁹⁶ compared abdominal and hysteroscopic metroplasties and observed no significant difference in the postoperative hysterosalpingogram or subsequent pregnancy outcome. However, hysteroscopic procedures were associated with decreased operating time, less blood loss, and shorter hospitalization. Thus, many have concluded that it would be reasonable to perform prophylactic hysteroscopic “metroplasties” in women undergoing infertility evaluation, prior to assessing whether conception occurs and pregnancy outcome following conception.
 

Table 15. Pregnancy outcome after metroplasty for uterine septum 

Fibroids

Uterine fibroids are a frequent finding, and occur with increased frequency in older women prior to menopause. While uterine fibroids are a recognized cause of pregnancy loss,^{307, 308} as a general rule, they are considered to be a rare cause of infertility.³⁰⁸ In certain circumstances fibroids may contribute to infertility by causing tubal occlusion at the cornua, filling the uterine cavity or preventing successful nidation by impairing blood flow to the overlying endometrium where an embryo attempts to implant. It is estimated that 5–10% of infertile women have fibroids and that fibroids are present in 1–2.4% of infertile women with no other obvious cause of infertility.³⁰⁹ For an in depth review of myomas, reproductive function, and pregnancy outcome, the reader is referred to published summaries.^{287, 309, 310, 311}

Despite the general consensus that myomas rarely cause infertility, conception rates following myomectomy average approximately 50% (Table 16).^{307, 308, 312, 313, 314, 315, 316, 317, 318, 319, 320, 321, 322, 323, 324, 325, 326, 327, 328, 329} Perhaps this apparent paradox can be explained by concomitant treatment of other tuboperitoneal disease at time of myomectomy (e.g., endometriosis, tubal obstruction, pelvic adhesions). In fact, in women with myoma but without infertility factors, Berkeley and colleagues noted conception in only one of six women (16%).³¹² Assessment of the use of CO₂ laser in abdominal myomectomy has been limited. McLaughlin³¹⁵ noted a subsequent pregnancy rate of 33% and described an approximately one third reduction in blood loss with the CO₂ laser as compared to nonlaser technique.

Table 16. Pregnancy outcome after treatment of uterine fibroids 

† Randomized studies.

One of the most dreadful complications of myomectomy is spontaneous uterine rupture during pregnancy or in labor. Although uterine rupture after abdominal myomectomy is not a common occurrence, its incidence with laparoscopic myomectomy is reported to be as high as 1%.³¹⁹ There are only a few randomized controlled studies that compare laparoscopic myomectomy with conventional abdominal myomectomy. In the only randomized and double blinded study, Holzer and colleagues reported significantly less pain in the first 72 hours following surgery in patients having myomectomy by laparoscopy compared to laparotomy.³³² In this study 40 women undergoing myomectomy were randomized to laparoscopic (19 patients) or abdominal (21 patients) approach. All patients had a similar abdominal surgical dressing so that both patients and observers could not tell which procedure was utilized. Four randomized trials compared the short term outcome of laparoscopic and abdominal myomectomy.^{318, 333, 334, 335} Compared to laparotomy, laparoscopic myomectomy was associated with longer operative time, less intraoperative blood loss and postoperative pain, shorter postoperative ileus, less incidence of postoperative febrile morbidity along with shorter hospital stay and recovery time. Only two of these trails compared reproductive outcome in infertile women undergoing myomectomy and demonstrated no difference in pregnancy rate between laparoscopic and abdominal myomectomy groups (total number of patients was 267 in both trails).^{317, 318}

A potential complication of myomectomy is the postoperative development of pelvic adhesions, with its associated effects on fertility. Tulandi and co-workers performed a second look laparoscopy on 26 infertile women 6 weeks after undergoing abdominal myomectomy for large uterine fibroids. In all, 94% of women with posterior uterine wall incisions and 56% of women with fundal or anterior incisions developed adnexal adhesions.³³⁶ Women with posterior incisions had more dense adnexal adhesions. The incidence of adhesions after laparoscopic myomectomy is reported to be less than that with laparotomy, with several case series reporting a 28–66% incidence rate on second look laparoscopy.^{337, 338, 339, 340} In a case–control study, Bulletti and his group compared the development of adhesions postmyomectomy performed laparoscopically (14 patients) or abdominally (14 patients). On second look laparoscopy they documented adhesion formation in 29% of patients in the laparoscopy group and 64% of patients in the laparotomy group.³⁴¹ However, in randomized controlled studies investigating the use of surgical adjuvants (barriers) to prevent postlaparoscopic myomectomy adhesion development, the incidence of adhesions in the control group (where no adjuvant was used) ranged between 59% and 88% which is close to what is reported in abdominal myomectomy. With the use of adhesion prevention adjuvant the incidence of adhesions ranged from 28% to 40%.^{342, 343, 344} There seems to be no difference in the recurrence rate of fibroids postlaparoscopic or abdominal myomectomy. Rossetti and coworkers randomized 81 patients to laparoscopic or abdominal myomectomy and found a recurrence rate of 27% and 23%, respectively, on 40 months follow up.³⁴⁵

Neuwirth^{323, 324} described hysteroscopic resection of submucous uterine fibroids with subsequent pregnancy in eight of 28 women, following reports showed higher pregnancy rates.^{325, 326, 327, 328, 329} The hysteroscopic procedure, in experienced hands, has the theoretical advantages of reducing hospitalization, operative morbidity, consideration of cesarean section for delivery, and the development of postoperative pelvic adhesions. However, it does not allow removal of nonsubmucosal fibroids. The latter is potentially an important factor because the rate of symptomatic recurrence of myoma has been reported to be 27% in women with isolated myomas and 59% with multiple myoma.³⁴⁶ In part, the recurrence rate probably varies with the diligence of the surgeon in removing “seedling” myomas at the initial procedure. The main complications of hysteroscopic myomectomy are uterine perforation and fluid overload. Fluid overload results from the absorption of the hypotonic media during operative hysteroscopy. The resulting hyponatremia, in its severe form, could be life threatening. To avoid fluid overload the surgeon should keep track of the fluid deficit while performing hysteroscopy. This is best done by using a closed fluid management system. The surgeon should conclude the procedure if the fluid deficit exceeds 1000 ml and the procedure should be immediately stopped if the deficit is 1500 ml. To minimize fluid overload, the distention pressure should be set at the lowest level feasible to perform the procedure and should not exceed the patient’s mean arterial pressure. While resecting a myoma the amount of fluid absorbed is small since myomas are made mainly of fibrous tissue and lack blood vessels within them. On the other hand, the amount absorbed increases significantly when normal myometrium is resected because of its increased content of blood vessels. Disruption of the endometrial integrity by performing curettage before hysteroscopy is discouraged since this would expose the terminal vessels and increase fluid absorption. When the fluid deficit starts to increase exponentially uterine perforation should be suspected. Although recent development in hysteroscopy in utilizing bipolar electric energy or use of special morcellator allows the use of electrolyte rich isotonic distension media, fluid deficit should be recorded. The surgeon should conclude the procedure when the deficit is 2000 ml and immediately stop if the deficit is at 2500 ml when electrolyte rich isotonic solution is used. Endometrial cavity adhesions, or Asherman’s syndrome, may develop if two opposing submucosal fibroids are hystreroscopically resected during the same procedure. Opposing fibroids are better treated separately by hysteroscopy to prevent contact of two opposing raw surfaces and development of adhesions. 

Uterine artery embolization (UAE) is another minimally invasive method of treating uterine fibroids. It involves the injection of micro-particles or sclerosing substances into one or both uterine arteries to block the blood supply to the fibroid. This would cause degeneration of the fibroid. Patients may experience postembolization syndrome (pain, fever, nausea, and vomiting) that requires hospitalization. Up to 1% of patients may develop pelvic infection. Future fertility with this procedure is uncertain and several series reported incidence of ovarian failure which could be up to 5–14%,^{347, 348} however, ovarian failure seems to predominantly affect older women.³⁴⁹ There are few series reporting on pregnancy after UAE. Pabon and colleagues recently reported on 100 women who underwent UAE, 57 of whom were seeking pregnancy. There were 11 pregnancies in ten patients (18%) with eight live births, seven term and one preterm, four vaginal deliveries, and four cesarean sections.³⁵⁰ In another study, 121 women with large intramural fibroids and desire to preserve fertility were randomized to UAE (58 patients) or myomectomy (63 patients, 42 laparoscopic and 21 abdominal). Over a period of at least a year, 26 women after UAE and 40 women after myomectomy tried to conceive. There were significantly fewer pregnancies and deliveries in the UAE group compared to the myomectomy group with 17 pregnancies and five deliveries vs. 33 pregnancies and 19 deliveries in the two groups, respectively.³⁵¹ There are few other small case series reporting adverse pregnancy outcome like increased preterm labor or intrauterine growth restriction after UAE. In view of the limited data present and lack of well designed studies about pregnancy outcome after UAE, this procedure should not be offered as an alternative to myomectomy to women who desire to preserve fertility.

GnRH analogues have also been used to treat leiomyoma. While these agents have been successful in reducing the size of the myoma during administration, cessation of treatment has been associated with re-enlargement.^{352, 353} The analogues have been used in conjunction with surgery to achieve preoperative shrinkage. However, the frequency with which this approach truly enhances the ease of surgery has not been established: some surgeons feel the myoma capsule is less well-defined following GnRH therapy, thereby making the surgery more difficult. One advantage of GnRH pretreatment that has been demonstrated is the reduction of blood loss; however, whether the volume of blood sample is clinically significant is unclear. In anemic patients, use of GnRH analogues may be used successfully as a temporizing measure to postpone the surgery while allowing the patient to build up her blood count and donate autologous blood for transfusion should it become necessary. It should be noted, however, that myomectomies are infrequently complicated by their large volume of hemorrhage requiring heterogeneous transfusion or the need to convert to an undesired hysterectomy.³⁵⁴

One additional consideration in placing women with uterine fibroids on GnRH analogues is that it may actually be a leiomyosarcoma. While a rare occurrence, this has occurred and can be suggested by myomas that fail to shrink in size or actually enlarge during GnRH therapy.^{355, 356} If leiomyosarcoma is present, it is usually (95% of the time) the only or the largest myoma present. Lastly, it appears that GnRH therapy should not significantly alter the ability of the pathologist to differentiate a benign myoma from a malignant leiomyosarcoma.³⁵⁷

Endometrial polyps

Endometrial polyps are benign local overgrowths of the endometrium and are commonly associated with abnormal uterine bleeding. The association between polyps and infertility is still unclear. The polyps may play a factor by mechanically interfering with sperm transport or produce inhibitory factors that affect embryo implantation.²⁸⁷ Endometrial polyps can be identified on hysteroscopy in up to 16% of women with otherwise unexplained infertility.³⁵⁸ In infertile patients with endometriosis the rate is even higher (47%).³⁵⁸ Hysteroscopy is the gold standard method for the diagnosis of endometrial polyps. Polyps can be diagnosed by transvaginal ultrasound, hysterosapingography or saline infusion sonohysterography (SIS), with the latter reported to be more sensitive. A meta-analysis of published studies about SIS in uterine cavity evaluation reported a pooled sensitivity of 95% and specificity of 88%.³⁵⁹ A total of 215 infertile patients with endometrial polyps undergoing IUI were randomized to have hysteroscopic polypectomy or hysteroscopy with biopsy of polyp before initiation of treatment. Over the period of four IUI cycles, the pregnancy rate was significantly higher in the polypectomy group (63.4%) compared to the biopsy group (28.2%). Of interest, 65% of the pregnancies in the polypectomy group occurred spontaneously before the first IUI cycle.³⁶⁰ Other observational studies reported a 50–78% postpolypectomy spontaneous pregnancy rate in infertile women with endometrial polyps.^{361, 362, 363, 364} The location of the endometrial polyp^365, but not the size or number³⁶¹ seems to affect the favorable impact of polypectomy on infertility, with uterotubal junction polyps having the most deleterious effect.

The main objective in performing reproductive pelvic surgery in infertile patients is to restore pelvic anatomy and have a patent and functional pathway for the sperm to reach the ovulated oocyte for fertilization and transport of the embryo back into the endometrial cavity for implantation. With ART and particularly IVF with embryo transfer the need for a patent pathway is bypassed. In depth discussion of ART is beyond the scope of this chapter, however, ART has witnessed dramatic improvement in the past two decades due to advances in embryo culture media and improvement in controlled ovarian hyperstimulation protocols. The utilization of IVF is on the rise with more than 100,000 cycles performed annually in the United States and a current per cycle success rate often between 40 and 50% in some patient groups. Embryo freezing and thawing has also improved yielding pregnancy rates comparable to those of fresh cycles of IVF. With such a high per cycle pregnancy rate, IVF is becoming the mainstay of treatment for a variety of infertility causes, mainly severe male factor infertility. Whether IVF should become the primary method of treatment for tuboperitoneal cause of infertility is controversial. If the surgical intervention for tuboperitoneal infertility is successful in restoring fertility the patient can attempt more than one pregnancy following the procedure which may turn out to be more cost effective compared to IVF where the patient may need to undergo treatment every time she desires a pregnancy with all the associated financial and emotional burdens. However, with IVF becoming more accessible and perceived less invasive than surgery more patients with tuboperitoneal disease are resorting to it as their primary approach in the treatment of infertility. This may result in less performance of reconstructive tubal microsurgery which in turn affects the training of the new generation of gynecologic surgeons, thereby eliminating this credible surgical option in most centers. To date there is no randomized controlled study comparing surgical treatment of tuboperitoneal cause of infertility with ART.

Two categories of challenges confront the gynecologic surgeon. The first is to examine the previous success of surgical treatment of tuboperitoneal disease and intrauterine lesions, and to compare them with the efficacy and safety of new innovations involving different operative approaches (laparotomy, laparoscopy or robotics) and instrumentation that continue to be introduced into the performance of reproductive pelvic surgery. Additionally, the efficacy of these innovations will have to be compared with nonsurgical means of treatment, namely medical therapy (e.g., danazol and GnRH analogues for endometriosis, methotrexate for ectopic pregnancies) or use of ART. The second challenge is to recognize specific areas where improvement is needed and to develop approaches, devices, drugs, and instruments necessary to overcome them. Such areas include maximizing pregnancy outcome, identifying accurate noninvasive methods of assessing intraperitoneal endometriosis and adhesions, and continuing to reduce the operative morbidity associated with gynecologic surgery.