Tests of Tubal Patency

Hysterosalpingography

In 1910, Rindfleisch injected a watery paste of bismuth into the uterine cavity of a young woman in an attempt to diagnose a possible pregnancy. The roentgenogram obtained demonstrated the uterine cavity and the left fallopian tube. In 1913, Rubin and Cary demonstrated tubal patency by roentgenographic examination after intrauterine injection of Collargol, a silver salt. The first iodine preparation, Lipiodol, was used in 1925.¹ In his original use of this technique, Rubin used an oily contrast medium. Oil-based media remained popular until the 1960s. Their popularity was based on their ability to demonstrate a good delayed film when image-intensified fluoroscopy was not universally available. Because most hysterosalpingography (HSG) is now performed by image-intensified fluoroscopy, many experts² concluded that oil-based media should no longer be used, due to the availability of what are perceived to be safer water-soluble preparations. To these authors, the advantage of the good delayed film was less important because fluoroscopic observation and short-term delay films gave equal information. Reevaluation of data on the use of these various preparations indicates that oil granuloma formation subsequent to the use of oil contrast medium is highly unlikely and of little clinical consequence when it does occur.^{3, 4, 5} There are conflicting data to indicate that cumulative pregnancy rates increase and that there is a reduced time to conception following the use of oil-based contrast.⁶ Nonetheless, the manufacture of oil-based contrast for hysterosalpingography (Ethiodol) was discontinued in 2010.

TECHNIQUE

HSG is an outpatient procedure. It should be performed in the proliferative phase of the cycle to reduce the possibility of radiation exposure of an early embryo. In addition, it obviates the unlikely possibility of inducing an ectopic pregnancy by performing the procedure after fertilization and altering embryo transport through the fallopian tube. The procedure should not be carried out during the active stage of any pelvic infection. Antibiotic prophylaxis is recommended (doxycycline 100 mg bid for 5 days) with a previous history of pelvic infection or if a hydrosalpinx is demonstrated during the study.⁶ PostHSG pelvic infections have been reported as high as 3.1%.⁷ The procedure usually requires no sedation. While providing analgesics prior to the procedure would seem advisable, a recent Cochrane database analysis failed to confirm any benefit from the use of pre-procedure or intraprocedure analgesics. Postprocedure analgesics were found helpful.⁸ The gynecologist is well advised to inform the patient that she will experience discomfort. A positive and comforting attitude by the gynecologist, combined with careful attention to technical details, will leave the patient with only minor and transient discomfort.

Contraindications to the procedure are pregnancy and active or recent pelvic infection. Performing the procedure in the proliferative phase of the cycle helps to avoid the possibility of radiation to an early gestation. The American College of Obstetricians and Gynecologists recommends antibiotic coverage in patients with a history of previous pelvic infection. A commonly prescribed antibiotic is doxycycline, 100 mg orally twice daily for 5 days.⁹

The patient is placed on the x-ray table in the lithotomy position. A radiolucent speculum is used to expose the cervix. If a radiolucent speculum is not available, a metal speculum may be used temporarily. After adequately cannulating the cervical canal, the speculum should be removed so as not to obscure visualization of the cervical canal during injection of contrast. The cervix is cleansed with an antiseptic solution and the anterior lip of the cervix grasped with a single toothed tenaculum. A cannula fitted with a rubber acorn tip is applied into the cervical canal. The tip of the cannula should not protrude further than 1 cm into the cervical canal. The acorn tip should be fitted so as to prevent reflux of the contrast medium. A locking device on the cannula can be used to fix the cannula to the tenaculum, allowing the two to be moved in tandem during fluoroscopy. It also decreases the likelihood of leakage of the medium from the cervical canal.

Warming the contrast medium decreases patient discomfort.¹⁰ The warmed contrast medium is slowly introduced through the cannula into the cervical canal. The cannula should be completely filled with the contrast medium before injection to avoid introducing air into the endometrial cavity. This would not only create confusing artifacts but also could possibly obstruct the fallopian tubes, giving a false-positive finding. Water-soluble medium is preferred by some clinicians to avoid the perceived hazards of intraabdominal oil (e.g. embolization and granuloma formation should the medium be trapped in pockets of intraperitoneal adhesions). However, these results are unlikely and of no clinical significance when they occur.^{3, 4, 5} It has been noted that some thyroid function tests may remain altered for several months after injection of the oil-based iodine preparations.² Proponents of the use of oil-based media point out the advantage of the more intense image obtained.

The procedure is best observed under image-intensified fluoroscopy. If this is not available, the following procedure is recommended. The dye may be injected without manometric control, although such control has been suggested by some. Most gynecologists think it is sufficient to feel the resistance of the dye and to assess the patient's subjective complaint of discomfort to gauge the rate and pressure of contrast injection. About 1 mL of the contrast should be introduced into the endometrial cavity, and then an x-ray should be obtained. The film is then developed and used to decide whether further injection is necessary or hazardous. This process is repeated until the tubes are filled and intraperitoneal spill is obtained. About 4–5 mL of contrast is usually required. If an oil-based medium has been used, a follow-up film 24 hours later may be obtained.

When the procedure is monitored under image-intensified fluoroscopy, the contrast is observed during its slow, steady injection. Spot films may be obtained during the injection. It is most important to observe the endometrial cavity during the early stages of filling to visualize endometrial defects that may be obscured as the cavity is completely filled with contrast. Defects that may be obliterated by full expansion of the uterus are endometrial polyps, intrauterine synechiae, and small myomata. The contrast material may then be observed traversing the fallopian tubes and ultimately spilling into the peritoneal cavity. Alternatively, the flow of dye may be obstructed anywhere along the course of the tube, or it may enter a large hydrosalpinx, simulating spill into the abdominal cavity.

The observation of cornual obstruction should be taken with reservation because this is occasionally due to functional spasm. Various drug regimens have been suggested for eliminating this spasm, such as the use of anticholinergic drugs, beta-mimetic drugs, tranquilizers, antispasmodics, spasmolytics such as amyl nitrite, and sedatives. It has also been suggested that general anesthesia may relieve the uterotubal spasm. However, none of these has consistently worked. The cornual closure cannot be accepted until confirmed by additional procedures such as selective cannulation or laparoscopy. Other techniques for introduction of the dye have been suggested. A radiopaque suction cannula of the Semm variety that fixes to the cervix and introduces the dye at the external os is available. This is a recommended procedure that allows good visualization of the cervical canal and prevents obstruction by the cannula. Other available commercial devices deliver the dye into the uterine cavity, but they have the disadvantage of obscuring detail of the cervical canal.

If there appears to be bilateral cornual obstruction, the possibility of cornual spasm or small amounts of debris obstructing the small intramural portion of the tube must be considered. Under these circumstances, the appearance of the endocervical canal is not of immediate importance, and a device that allows increased intrauterine pressure can be used. Devices that inflate a balloon at the internal and external cervical os are available; these allow such pressures to be used carefully in an effort to overcome tubal spasm or to force out small amounts of intraluminal debris.

A pediatric Foley catheter for the introduction of the dye has been used for many years.¹¹ In this procedure, a No. 5 pediatric Foley catheter is introduced into the cervical canal. The balloon is then insufflated with about 2 mL of water, which obstructs the cervical canal. Contrast is then introduced through the Foley catheter. The advantage of this procedure reportedly is decreased pain and bleeding that may be caused by the tenaculum. The patient may be more comfortably moved into the various positions for better visualization. I have used this procedure intermittently and find it to be a most acceptable alternative to the use of the acorn-fitted metal cannula. It is particularly suitable for defining such anomalies as uterus didelphys. Its disadvantage is that it does not allow uterine manipulation. When the bulb is allowed to enter the endometrial cavity, it completely obscures the detail of the endocervical canal and also obscures the anatomy of the endometrial cavity. There are commercially available HSG insufflation cannulas that can be used to reproduce the Foley catheter technique. When using these devices, it is frequently necessary to allow the bulb to enter the endometrial cavity to obtain the traction necessary to manipulate the uterus into a good frontal position. When this is done, the endometrial cavity must be viewed fluoroscopically as the bulb is being deflated and before the evacuation of the contrast to observe the detail of the endometrial cavity adequately.

Generally, the choice of contrast medium is between an oil-based medium and a water-based medium. Initially, oil-based media were the media of choice because of the intense contrast provided and because of the ability to obtain good delayed films. Subsequently, this medium was less commonly used because of concern that the intraperitoneal oil could be associated with intraperitoneal granuloma. Miyamoto and associates¹² compared water-soluble contrast with Lipiodol, an oil-soluble material administered intraperitoneally to rabbits at doses of 100 mg of iodine per kilogram. Retention in the body was investigated by x-ray imaging, plasma kinetics, and urinary and fecal excretion. Irritability in the abdomen was investigated by histologic examination. The water-based material was entirely excreted into the urine within 2 days. The oil contrast was excreted with a half-life of 50 days and retained for 21 days in the abdomen. The water-soluble material induced no inflammatory reaction, whereas the oil-based product induced an abdominal inflammatory reaction, including granuloma formation. Although this has been demonstrated in experimental animals, it has not been documented in women. A Cochrane database review failed to identify any clinically significant complications from the use of oil based media.^{3, 4, 5} Therefore, the oil-based product remained in use by many gynecologists because of the better contrast it provided. In addition, pregnancy rates after the use of oil contrast have been reported to be significantly higher than rates after water-based contrast,¹³ and decreased time to conception has also been reported.⁶ This is particularly so in cases of unexplained infertility. It was presumed that this effect is the result of dislodging intraluminal debris or disrupting minor degrees of perifimbrial adhesion. Such claims are difficult to evaluate because of the spontaneous pregnancy rate of about 30% that occurs generally in patients with unexplained infertility. Nonetheless, Savage Laboratories discontinued the manufacture of oil-based contrast in 2010 for economic considerations. Water-based contrast specifically designed for HSG is available (Sinografin). This material provides excellent imaging but is quite expensive. Other generally available water-based contrast media (e.g. Hypaque) may be used, although contrast is less than optimal.

A pneumoperitoneum performed at the time of HSG may be helpful to allow the demonstration of peritubal adhesions. Pneumoperitoneum can be achieved by insufflating CO₂ through the fallopian tubes before instilling contrast. Pneumoperitoneum can also be obtained with a cul-de-sac puncture using a Veress-type needle or by means of transabdominal injection, as in laparoscopy. Although this combination approach previously achieved some popularity, it is now rarely used since the less invasive technique of simultaneous ultrasonography usually is available. The combination of HSG and ultrasonography achieves imaging both of the intraluminal and extraluminal portions of the uterus. This is important in defining the extent of intrauterine defects such as submucous leiomyomata. It is also helpful in differentiating uterus didelphys from uterus bicornus.

Obstruction of the endocervical canal is occasionally encountered after cauterization, cryosurgery, or cone biopsy of the cervix. Under these circumstances, severe stenosis of the cervical os makes cannulation and contrast introduction difficult. Dilation of the cervical canal may be done before HSG. If necessary, angiodilators, as described by Yoder and Pfister,¹⁴ can be used.

FINDINGS

Although a detailed discussion of interpreting HSGs is beyond the scope of this chapter, it should be understood that demonstrating tubal patency is not completely reliable. An apparent blockage should be regarded only as suggestive, because about 10% of patients with apparent bilateral blockage by HSG have become pregnant without further therapy.¹⁵ Even patency may be questionable, because a large hydrosalpinx may mimic intraperitoneal spill of contrast material and frequently will not demonstrate peritubal adhesions.¹

In Siegler's series¹ of 1000 patients, 58% had bilateral patent fallopian tubes, 20.5% had cornual obstruction, 12.6% had distal obstruction, and 5.7% had midtubal obstruction. Topkins¹⁶ found fimbrial occlusion to be the most common reason for obstruction. Sanfilippo and colleagues¹⁷ reported that 54.3% of HSGs performed in infertile couples demonstrated bilateral tubal patency. Tubal abnormalities were demonstrable in the remainder. Cornual obstruction was noted in 5.3%; the largest number of obstructions occurred at the fimbria. Gabos¹⁸ found bilateral cornual obstruction in 13% of patients, and Hutchins¹⁹ in 15%. In a study by Ostry,²⁰ 27 of 209 patients demonstrating bilateral cornual obstruction spontaneously achieved pregnancy.

Using laparoscopy as the gold standard, hysterosalpingography has been reported to be only 65% sensitive and 83% specific in demonstrating tubal pathology. The interobserver reliability for proximal tubal occlusion was 85%, for distal tubal occlusion was 69%, for hydrosalpinx was 64%, and for peritubal adhesions 55%. The reported intraobserver reliability for proximal tubal occlusion was 89%, for distal tubal occlusion was 72%, for hydrosalpinx was 68%, and for peritubal adhesions was 65%.⁵

CONTRAINDICATIONS AND COMPLICATIONS

Some patients are intolerant of HSG as an outpatient procedure. Under these circumstances, the test may be performed under conscious sedation although other imaging methods, such as HyCoSy, should be considered. HSG can be performed under general anesthesia. Under these circumstances, however, the gynecologist may want to proceed to combination of hysteroscopy with laparoscopy as a more comprehensive alternative. 

HSG is contraindicated in patients who exhibit evidence of active or subacute pelvic infection. In addition, active uterine bleeding, recent pregnancy, and curettage should also be considered contraindications. Even in the absence of these contraindications, one should not hesitate to abandon the procedure if the patient experiences unusual pain during its performance.

Generally, only insignificant complications occur secondary to hysterosalpingography. The postprocedure infection rate has been reported as high as 3.1%.⁷ All contrast agents have been demonstrated to elicit local inflammatory reaction in the fallopian tubes. These reactions invariably resolve within 2 weeks of exposure.²¹ There have been ten deaths reported in the literature.^{18, 22} All but one of these occurred in early use of the procedure and were associated with the use of oily media. In most instances, the deaths were due to anaphylaxis or embolization. The most recent report of a death, which occurred after the injection of water-soluble contrast medium, was published in 1986. There have been other complications such as allergic reaction, extreme pain, uterine rupture, tubal rupture, and hemorrhage. Cardiac arrhythmias and vagal responses resulting in severe abdominal pain and bradycardia have been reported. Patients have also noted chemical and toxic complications, such as granulomas and iatrogenic thyroid dysfunction. Intravasation of contrast has occurred, but this is of little concern with the use of water-soluble contrast.

It is generally reported that patients are exposed to less than one rad of irradiation during HSG (1 rad = 10 mGy). In a study by Sheikh and Yussman,²³ 29 patients undergoing HSG were monitored with a dosimeter in the upper vagina. The dosimeter measured between 75 and 500 millirads (0.75–5 mGy). The variation was due to the number of spot films obtained and the duration of fluoroscopic exposure. This work was confirmed by others,^{24, 25, 26} who have measured from 60 to 2500 millirads (0.6–25 mGy) during HSG.

Radiologic selective cannulation of the fallopian tube

Radiologically demonstrated cornual obstruction in the absence of other demonstrable tubal disease remains an enigma for the radiologist and the gynecologist. The differential diagnosis is tubal spasm, true mechanical blockage, such as salpingitis isthmica nodosa, or plugging of the 0.5-mm intramural lumen with mucus or inspissated uterine detritus. Occasionally, these obstructions were approached surgically. The cornu was resected or the tube transected, probed, and then reanastomosed. About two thirds of the fallopian tubes resected for proximal tubal obstruction fail to reveal luminal occlusion when examined microscopically.²⁷ Selective fluoroscopic transcervical fallopian tube catheterization to help clarify such situations was first performed by Thurmond and colleagues in 1988.²⁸

TECHNIQUE

The procedure is performed on an outpatient basis and requires only mild sedation. The precautions and contraindications are the same as for HSG. The patient is positioned and draped in the lithotomy position on the x-ray table. Some prefer to perform the procedure on a cardiac catheterization table to allow better positioning of the patient. The cervix is prepared as for HSG. A vacuum cup applied to the cervix allows manipulation of the cervix and also allows the introduction of a series of coaxial catheters and guide wires into the cervical canal. A preliminary HSG is obtained to confirm cornual occlusion. A 5.5 French curved catheter is introduced through the cervical vacuum cup into the uterine cavity and advanced into the uterine cornu. A J-wire may be used to guide the catheter to the region of the cornu. The J-wire is removed and replaced with a small-diameter guide wire tipped with soft platinum. An effort is made to introduce the guide wire into the intramural and isthmic portion of the fallopian tube. This is usually accomplished with short and gentle intermittent probing. When the guide wire is judged to have entered the fallopian tube, a 3 French Teflon catheter is advanced over the guide wire until an obstruction is encountered. The guide wire and catheter are alternately advanced until the catheter is judged to have entered the isthmic portion of the fallopian tube. The soft platinum tip can usually follow the tortuosity of the fallopian tube without difficulty, allowing the Teflon catheter to follow. The guide wire is then removed and contrast material injected through the catheter. If patency is not obtained, the coaxial guide wire and catheter can be advanced until an unremitting obstruction is encountered or until patency is obtained. Several attempts at positioning the catheters may be necessary if there is a question of proper placement. Once patency is obtained or the procedure is abandoned on one side, the process is repeated at the other tubal ostium.

RESULTS

In their initial series, Thurmond and associates²⁸ reported that 10 women conceived out of 35 women who had patent fallopian tubes after selective catheterization. One woman aborted twice and another had an ectopic pregnancy. The eight remaining patients had ongoing pregnancies at the time of that publication. In a follow-up publication in 1990, Thurmond and Rosch²⁹ reported that they were able to obtain patency in at least one tube in 86 of 100 patients in whom selective fluoroscopic tubal cannulation was attempted. Twenty-six intrauterine pregnancies occurred in these 86 patients. Among these 100 patients, 20 were identified who had bilateral cornual obstruction with no other demonstrable cause for infertility. Tubal patency was obtained in 19 of these women. Nine patients conceived, and all the pregnancies were intrauterine. Eight of the 10 patients who did not conceive had subsequent HSG. Of these, four again showed bilateral cornual occlusion.

In a further series reported in 1995,³⁰ the same group reported their experience with selective cannulation in cases of salpingitis isthmica nodosa. In this series, 47 of 65 tubes with proximal obstruction were recanalized successfully. Among the 19 women who were able to conceive via these recanalized salpingitis isthmica nodosa tubes, there were six live births and two tubal pregnancies.

Thompson and co-workers¹³ reported their results in 47 tubes with cornual obstruction. Of these, 11 (23.9%) were demonstrated to be patent on repeat HSG. An additional six (13%) obtained patency by simply injecting contrast after selectively wedging the obstructed ostium with a Teflon catheter. Nine of the remaining 30 tubes had successful wire recanalization (31%). One of these demonstrated distal obstruction. Other authors have also reported patency rates between 30% and 40% after selective cannulation, achieving pregnancy rates of 15% among those who obtained patency and ectopic pregnancy rates of 25% in patients who conceived.^{31, 32}

In institutions where interventional radiologists are available and interested, this technique has proven to be increasingly useful, particularly if bilateral cornual obstruction is identified in the absence of other causes for infertility. In addition, this technique may also be attempted before tubal microsurgery or in vitro fertilization in patients with salpingitis isthmica nodosa.³⁰

Saline infusion sonography and hysterosalpingo-contrast sonography

The ultrasonic observation of uterine architecture, endometrial lining and endometrial cavity interfaces following the instillation of saline into the endometrial cavity has been termed saline infusion sonography (SIS). The attempt to delineate the fallopian tube lumen during the instillation of various scintillation contrasts has been given many names. These include hysterosalpingo-contrast sonography (HyCoSy), hysterosonosalpingography, contrast infusion sonography, Doppler hysterosalpingography, salpingosonography, and sonohysterography. Irrespective of name, the purpose of the procedure is to accomplish the sonographic delineation of the endometrial cavity and demonstrate patency of the fallopian tubes. The contrast medium distends the endometrial cavity and provides a fluid interface that allows the identification of endometrial abnormalities such as polyps, intrauterine adhesions, leiomyomata, and septa.  Unlike HSG, HyCoSy generally can be performed in the gynecologist's office using available ultrasound equipment. The procedure has certain advantages over HSG. The major advantage is that it allows visualization of extrauterine pelvic structures as well as more clearly delineating intramural and some intraluminal abnormalities. It is generally faster, less painful and less expensive than HSG. It easily allows both anterior and saggital views and avoids radiation exposure. While there is significant controversy, it is generally reported to be as accurate as HSG in demonstrating tubal patency (see section on comparative studies). 

TECHNIQUE

HyCoSy has the same contraindications as HSG. No premedication is required because the procedure generally produces little discomfort. The patient, who has previously emptied her bladder, is placed on the ultrasound table in the lithotomy position. A routine vaginal ultrasound examination is performed, systematically observing the cervix, the myometrium, the endometrial stripe, the ovaries, and the cul-de-sac. A speculum then is placed into the vagina and the cervix cleansed with an antiseptic solution. At this point, one of a number of catheters may be placed into the endocervical canal. An insemination catheter, a No. 5 pediatric Foley catheter, or one of a variety of specially designed intrauterine catheters may be used. A tenaculum is rarely necessary. If a catheter with a bulb is used, it should just be in the cervical canal so as not to obscure the endometrial cavity. At this point, the speculum is removed and the vaginal transducer reintroduced either anterior (in the case of an anteverted uterus) or posterior (in the case of a retroverted uterus) to the intrauterine cannula. Sterile saline is then slowly introduced through the cannula to distend the endometrial cavity. During instillation, the uterus is scanned longitudinally. The transducer is then turned 90° to scan the uterus in a transverse manner.

The endometrial cavity can be well outlined using saline as the hypoechoic contrast. However, the fallopian tubes cannot be consistently or accurately visualized using this technique. Randolph and co-workers³³ originally described the technique of instilling up to 200 mL of saline through the intrauterine cannula. The fluid was noted to collect in the cul-de-sac in the presence of tubal patency. The distal tube frequently could be seen silhouetted in the cul-de-sac fluid. In the presence of obstructed tubes, the cavity expanded without accumulation of cul-de-sac fluid. When compared with ultrasound, patency was correctly diagnosed in 91% of patients, although side-specific patency could be diagnosed only in 67% of patients.

Several contrasts have been used in an attempt to better delineate the fallopian tubes. The easiest of these is the simple addition of air bubbles to the saline. This can be accomplished by shaking the saline solution to create the air bubbles. This same concept is accomplished by injecting a small amount of air into the uterine cavity prior to injecting the saline. The air bubble is then followed through the endometrial cavity and out the fallopian tubes if they are patent. The air injection must always follow a preliminary examination of the endometrial cavity in order not to create intraluminal artifacts. A variation on this technique is the alternate installation of air and saline accomplished by the use of a stopcock fitted to separate syringes of air and saline.  More recently, the Femvue Sono Tubal Evaluation System (Femasys Inc), which introduces air and saline in a more controlled manner, has become available.

Specific contrast fluids have been developed in an effort to delineate the fallopian tubes. Echovist-200 (Schering AG, Berlin) is a suspension of galactose monosaccharide microparticles in an aqueous solution. The suspension is prepared immediately before use and instilled in the same manner as saline. To diagnose the patency of each tube, the contrast is observed to emerge into the cul-de-sac at the end of the fimbria. As an alternative, one may observe a steady intraluminal flow of scintillations lasting at least 120 seconds in at least one imaged tubal section. The scintillations can be observed to flow over the ovaries. This medium causes some degree of discomfort in half of the patients and occasionally causes enough discomfort to require analgesia.³⁴ When compared with results obtained by HSG in the detection of tubal patency, Echovist showed excellent correlation when tubes were patent but overdiagnosed tubal obstruction. Albunex (Molecular Biosystems, Inc., San Diego, California, USA), a suspension of human serum albumin microspheres which previously had been tested for tubal instillation, is no longer being marketed.

HyCoSy combined with SIS has excellent sensitivity and specificity for diagnosing intrauterine abnormalities, particularly endometrial polyps and submucous leiomyomata.^{35, 36} The kappa analysis for unilateral patency and bilateral occlusion is reported as 0.61, making it somewhat less accurate than HSG.⁵ Interestingly, the accuracy of diagnosing left tube patency is significantly lower than the diagnosis of right tube patency. It may be helpful in the detection of the less common combination of proximal and distal blockage.³⁷ The addition of color Doppler sonography, using saline as the contrast agent, increases the accuracy of assessing fallopian tubes. Using laparoscopy as the standard, the sensitivity is greater than 90% and the specificity is greater than 80% in the diagnosis of tubal occlusion.^{38, 39, 40, 41}

Hysteroscopy

The diagnosis of tubal problems using the hysteroscope was, for a long time, considered inappropriate. This was because of the concern that the complex angulation of the intramural portion of the fallopian tube would be disrupted by efforts to visualize or probe the interior of the fallopian tube hysteroscopically. The recognition that two thirds of cornual obstructions failed to demonstrate anatomic defects provided the impetus for searching out new methods of evaluating the cornual portion of the fallopian tube. Thurmond and Rosch's pioneering work on fluoroscopic selective cornual catheterization provided the impetus to overcome this reluctance to approach the intramural portion of the fallopian tube hysteroscopically.

If HSG or HyCoSy demonstrate consistent bilateral cornual occlusion, efforts can be made to cannulate the cornua selectively. It is not currently possible, however, to determine the extent of distal disease hysteroscopically. In two thirds of such cases, no anatomic abnormality can be demonstrated. The remaining third have salpingitis isthmica nodosa, concurrent distal disease, endometriosis, or another anatomic abnormality not immediately apparent with radiologic techniques. Concomitant laparoscopy, advisable in most of these cases, allows the physician to determine whether the occlusion is confined to the cornu or involves significantly more of the fallopian tube. One can then assess whether the obstruction can be repaired hysteroscopically or will require a laparoscopic or laparotomy approach. One can also assess, at this time, whether the patient would be better served by in vitro fertilization.

Prior to performing cornual cannulation, cervical culture and culture for sexually transmitted diseases, IgG and IgM for chlamydia and a  negative pregnancy test should be obrtained. Prophylactic antibiotics should be administered. The procedure is performed in much the same manner as the radiologic procedure. Any operating hysteroscope can be used. The hysteroscopy is performed in the usual manner. Any distension medium can be used, although sorbitol is most often selected. The instruments available for hysteroscopic cannulation of the cornu are similar to those for the radiologic procedure. There is no need for the suction cannula or the J-wire because the procedure is performed under direct observation.

The ostia of the tubes are identified. A 5.5 French polyethylene cannula with a slight curve at its end is introduced through the operating channel of the hysteroscope. The catheter usually can be manipulated into the tubal ostium with little difficulty. If the operator encounters difficulty placing the end of the catheter into the ostium, the curvature of the catheter end can be changed with the use of a trocar placed through the lumen of the catheter. When the catheter is adequately wedged into the ostium, dye is injected using a syringe attached to the Luer-Lok adapter on the catheter. Frequently, this maneuver alone serves to demonstrate the patency of the fallopian tube previously been thought to be occluded.

If the dye fails to traverse the cornu as demonstrated by laparoscopic observation, a coaxial set consisting of a guide wire and an overlying 3 French Teflon catheter is introduced through the larger catheter. The soft guide wire is manipulated through the cornu using gentle probing. Once the guide wire has entered the isthmic portion of the tube, the 3 French catheter is advanced. The alternating progression of the guide wire and the catheter ultimately allows cannulation past the cornu into the isthmus of the fallopian tube. The guide wire is removed and dye injected through the cannula to confirm patency.

Several studies have been published using this direct cannulation of the fallopian tube.^{27, 42, 43} Successful cannulation occurred in more than 70% of cases, and 60–75% of those cases remained patent. Pregnancy rates after this procedure have been high, but the number of cases reported to date have been too few to assess its therapeutic value. At present, it is prudent to consider selective cannulation of the fallopian tube as primarily diagnostic.

Tubal endoscopy

Endoscopy of the fallopian tube was initially reported in 1987.⁴⁴ As initially described, access was gained only to the ampullary portion of the fallopian tube. A 3-mm rigid telescope was passed along the operating channel of a laparoscope and into the fimbrial opening of the fallopian tube to allow assessment of the infundibulum and ampulla. Lesions of the mucosa were detected that were not apparent by HSG or laparoscopy. Similar observations of the distal fallopian tube have been accomplished by introducing a pediatric fiberoptic cystoscope through a 5-mm second puncture site at the time of laparoscopy. Currently, flexible fiberoptic or rigid salpingoscopes 1.8–2.8 mm in diameter with camera attachments and irrigating channels are available from several manufacturers.

TECHNIQUE OF SALPINGOSCOPY

The salpingoscope is introduced into the peritoneal cavity at the time of laparoscopy through a 3-mm reducer placed in a 5-mm second puncture cannula. The fallopian tube is stabilized by grasping the antimesenteric serosa immediately before the fimbria. In patients with an obstructed fallopian tube, a small opening can be obtained with scissors, laser, or unipolar needle. The salpingoscope is then introduced into the ampullary portion of the fallopian tube, which is distended with heparinized Ringer's lactate solution delivered by a hand-held syringe.⁴⁵

The Surreys⁴⁵ correlated salpingoscopic with laparoscopic findings in a group of patients with tubal disease. They showed that findings correlated well when the patients were demonstrated to have moderate to severe disease by laparoscopy. However, when laparoscopy showed only minimal disease, salpingoscopy was able to demonstrate significant intraluminal pathology in many cases. Marana and co-workers⁴⁶ demonstrated that salpingoscopy had a greater prognostic value than the American Society for Reproductive Medicine (ASRM) classification of adnexal disease in patients undergoing tubal surgery. Salpingoscopy may also be useful in evaluating the remaining tube in patients who undergo salpingectomy for ectopic pregnancy. In a series of 18 such patients,⁴⁷ eight of 13 who demonstrated normal mucosa in the remaining tube conceived an intrauterine pregnancy. None of these had a recurrent ectopic pregnancy. In the five patients with demonstrated intra-ampullary adhesions, three had repeat ectopic pregnancies.

Fiberoptic endoscopes 0.5 mm in diameter (falloposcopes) have been tested for observation of the cornual and isthmic endothelium. The endoscope is introduced into the uterine cavity through the operating channel of a flexible hysteroscope. The flexible hysteroscope is used to aim the falloposcope into the ostium of the fallopian tube. The scope is then advanced into the isthmus of the fallopian tube by gentle probing. Other ingenious scopes that allow the lens system to advance as the outer sheath unfurls in the tube have been devised to allow observation of the mucosa of the isthmus. These observations are made as the scope is slowly withdrawn. At this point, little clinical correlation with the use of these scopes is available, and they are used primarily for investigation.

Laparoscopy

After its introduction into the US in the mid-1960s, laparoscopy achieved the status of the ultimate test for demonstrating tubal patency and anatomic pelvic abnormality. The more recent success of selective cannulation of the uterine cornu and the promise of such experimental techniques as salpingoscopy and falloposcopy serve only to increase the importance of laparoscopy in the evaluation of tubal disease. This is because laparoscopy has proven to be an integral part or a convenient accompaniment to the newer techniques. It is not the purpose of this chapter to detail the technique of performing laparoscopy. However, it should be noted that laparoscopy has become the most commonly performed of all gynecologic surgical procedures and in competent hands should be considered among the safest of intraabdominal operations.^{48, 49} Except for the complications common to anesthesia, the complication rate for laparoscopy is only slightly higher than for other tubal diagnostic procedures,⁵⁰  although the overall complication rate for laparoscopy worldwide is reported to vary from 0.2% to as high as 10.3%.⁵¹ It is certainly the most accurate means available for demonstrating the extent of pelvic inflammatory disease and external tubal damage. However, it should not be used as the final arbitor of tubal disease since 3% of patients with normal laparoscopic studies have demonstrated intrauterine abnormalities using other techniques. In addition, although lack of patency is apparent after chromotubation during laparoscopic observation, the point of obstruction and the extent of luminal obliteration cannot be observed. Laparoscopy is most effective when combined with HSG or other tests such as salpingoscopy or selective cannulation.

Laparotomy

Laparotomy is a highly accurate and certainly a direct approach to the evaluation of tubal patency. However, the “infertility laparotomy” of the past is no longer an acceptable procedure since laparoscopy provides a similar assessment with less morbidity. However, if laparotomy has been performed for other indications, the demonstration of tubal patency can be accomplished in the following manner. The cervix is occluded by digital manipulation, by a clamp such as the Buxton clamp, or by some other device designed for occluding the cervical canal during laparotomy. A 22-gauge 1.5 needle is inserted into the endometrial cavity through the fundus of the uterus. A dilute solution of indigo carmine dye is then introduced into the endometrial cavity by means of a finger-control syringe attached through intravenous extension tubing to the intrauterine needle. The dye can be observed flowing through the fallopian tube, confirming patency.

The difficulty with this procedure is that occasionally one has trouble entering the endometrial cavity through a transfundal puncture. As an alternative, a No. 5 pediatric Foley catheter may be placed in the endocervical canal before laparotomy for the purpose of later introducing the indigo carmine dye. Several commercially available devices may be used to introduce dye into the endometrial cavity to check tubal patency during a laparotomy. They fix into the endometrial cavity and are placed transvaginally before the laparotomy. However, these devices seem to have little advantage over a transfundal needle or the use of an intracervical pediatric Foley catheter.

For the most part, IVF has replaced surgical repair of damaged fallopian tubes. However, when the procedure is performed, it is frequently necessary to inject concentrated dye solution through the endometrial cavity to stain the proximal portion of the fallopian tube. A concentrated dye also may be introduced through the fimbrial ostium to stain the distal portion of the fallopian tube. Under the operating microscope, serial incisions can then be made until well-stained endosalpinx can be observed. The stained lumina can then be reanastomosed. A small Foley catheter adapter is a good means of introducing the dye through the fimbrial ostium.

Hysterosalpingography versus laparoscopy

Many studies have compared the findings observed at laparoscopy with those of HSG. A few studies have compared HSG, laparoscopy, and tubal insufflation. In 1981, Philipsen and Hansen⁵² demonstrated agreement between HSG and laparoscopy in 57.7% of a group of infertile patients. However, among 54 patients whose HSGs were interpreted as normal, 29.6% exhibited abnormalities at laparoscopy. Among the 224 patients with abnormal HSGs, 24.6% had normal internal genitalia demonstrated at laparoscopy. Predictably, the greatest discrepancy between HSG and laparoscopic diagnoses was in the demonstration of peritubal adhesions. However, intrauterine abnormalities were diagnosed by HSG in 2.4% of laparoscopically normal patients. In 1977, Hutchins¹⁹ reported conflicting results in up to 65% of cases between the two procedures. In a 1978 study of 121 patients by Servy and Tzingounis,⁵³ false-negative results were found in 29% of patients and false-positive results in 7.4% of patients by HSG. 

Laparoscopy can demonstrate pelvic abnormalities in addition to tubal disease, including endometriosis, polycystic ovaries, and uterine leiomyomata. In a 1980 study performed by Portuondo and co-workers,⁵⁴ 335 infertile patients were evaluated by HSG and laparoscopy; in addition, tubal insufflation was performed. The results among the three methods agreed in 65% of patients. False results were found in 18% of patients by tubal insufflation, 9% by HSG, and 6.8% by laparoscopy. Nontubal pelvic pathology was detected by laparoscopy in 17% of the 336 patients. Pelvic adhesions and endometriosis were the pathologic entities most commonly undetected by HSG. It was apparent that evaluation by both HSG and laparoscopy was needed in 30% of the infertile patients.

A meta-analysis of 20 studies comparing HSG with laparoscopy for tubal patency and peritubal adhesions was reported by Swart and co-workers.⁵⁵ A total of 4179 patients with infertility culled from these 20 studies were evaluated. For tubal patency, the reported sensitivity and specificity were 65% and 83%, respectively. The authors confirmed the usefulness of HSG for demonstrating tubal obstruction and its unreliability for demonstrating peritubal adhesions.

The inter- and intraobserver variability in the interpretation of HSG was addressed by Mol and colleagues.⁵⁶ Four observers evaluated 143 HSGs twice on proximal tubal obstruction, distal tubal obstruction, hydrosalpinx, and peritubal adhesions. Reproducibility between observers was almost perfect for proximal obstruction, substantial for distal obstruction and hydrosalpinx, and moderate for adhesions. Reproducibility within observers was perfect for proximal obstruction and substantial for distal obstruction, hydrosalpinx, and adhesions.⁵⁷

It is apparent to physicians experienced in evaluating infertile patients that HSG and laparoscopy have become procedures to be used as adjuncts.^{58, 59} Laparoscopy can confirm such diagnoses as endometriosis and can demonstrate peritubal adhesions if the fallopian tubes are judged to be patent by HSG. In some situations, technical malfunctions during laparoscopy create difficulty in chromotubation of the fallopian tubes. Under these circumstances, a prior HSG that demonstrated patency of the tubes is reassuring. With chronically damaged fallopian tubes due to pelvic inflammatory disease, HSG can demonstrate the point of proximal obstruction, whereas laparoscopy can demonstrate the extent of distal obstruction. The observation of a distal hydrosalpinx combined with a proximal obstruction is considered by most pelvic surgeons to offer a poor prognosis for surgical repair.⁶⁰

HSG before laparoscopy has been advised for the preoperative evaluation of previously ligated fallopian tubes to demonstrate the extent of intramural patency or proximal isthmic patency.⁶¹ With little proximal tube demonstrated by HSG, a reimplantation procedure rather than a microsurgical reanastomosis was occasionally performed. The prognosis was poorer under these circumstances, and cesarean section was more likely to be necessary should pregnancy occur. Under such circumstances, the patient may elect not to have a repair performed. A clearly identifiable amount of proximal fallopian tube observed by HSG, combined with an acceptable amount of distal fallopian tube observed laparoscopically, indicated that a microsurgical reanastomosis was feasible, and the prognosis thereby improved.

The necessity for HSG before tubal reanastomosis has been challenged. Tubal reimplantation has, for the most part, been superseded by cornual reanastomosis in the few instances where the cornua have been severely damaged by the sterilization procedure. Previously, such damage was created when laparoscopic sterilization was performed with unipolar cautery. Methods currently in use for sterilization (e.g. bipolar coagulation, clips, Silastic bands, and partial resection) should not damage the cornua, thus eliminating the need for HSG. This was clinically confirmed in a study by Groff and associates⁶² in which preoperative HSGs performed on 81 patients failed to add information that would alter the management of their sterilization reversal. HSG should no longer be considered a prerequisite for tubal reanastomosis.

Hysterosalpingography versus hysteroscopy

Hysteroscopy is now in the armanentarium of most gynecologists. In the past there was significant disagreement among clinicians regarding the relative usefulness of HSG and hysteroscopy. In 1986, Daly and co-workers⁶³ concluded that abnormalities reported on hysteroscopy were of little clinical importance and that previously undetected abnormalities were rare in the patient who had undergone adequate HSG. They did, however, report three cases in which small “strategically placed lesions” were found to explain the patients' previously unexplained infertility. They suggested that routine hysteroscopy be added when laparoscopy is performed. Their rationale was that hysteroscopy added little time and morbidity, and occasionally revealed an otherwise undetectable abnormality.

A 1987 comparative study⁶⁴ of 100 infertile patients reported that the sensitivity and specificity of HSG were 74.2% and 91.3%, respectively, when compared with hysteroscopy. The authors suggested that a complete study should include both procedures. This opinion was somewhat different to that of Fayez and colleagues,⁶⁵ who studied 400 infertile patients with both HSG and hysteroscopy. They concluded that carefully performed HSG revealed virtually all abnormalities later confirmed by hysteroscopy. However, the exact nature of the intrauterine lesion usually could be determined only by hysteroscopy.

A 1988 prospective multicenter study⁶⁶ of 619 patients was designed to assess the relative value of these procedures. Examination of the tubal ostia by hysteroscopy revealed 26 patients with unilateral tubocornual polyps and one with bilateral polyps. There were hysteroscopically demonstrable lesions in 20% of the patients with normal results on HSG. In 35% of the cases, there were false-positive findings by HSG. The abnormalities demonstrated by hysteroscopy were confined to the uterine cavity or the cornual area.

The obvious limitation of hysteroscopy in the absence of salpingoscopy is inadequate assessment of the fallopian tubes. The promise of newer techniques and general availability of salpingoscopy may ultimately make HSG as antiquated as tubal insufflation. At this time, however, the combination of HSG and hysteroscopy is necessary for adequate assessment of both the endometrial cavity and the fallopian tubes.

Hysterosalpingo-contrast sonography versus hysterosalpingography

Using saline as a contrast medium, patency of at least one tube, as indicated by the accumulation of fluid in the cul-de-sac, agreed with laparoscopy in 91% of patients. No patient with bilaterally obstructed fallopian tubes was interpreted as having cul-de-sac fluid. Side-specific tubal patency agreed with laparoscopy only 68% of the time. A hydrosalpinx could be demonstrated by the presence of a fluid-filled sac before the instillation of saline.³³ The use of color Doppler increases the correlation with HSG to 92%⁴⁰ and correlation with laparoscopy to 93%.³⁹ However, in the diagnosis of myometrial and endometrial lesions such as polyps and leiomyomata, the use of saline contrast ultrasonography equaled that of hysteroscopy and laparoscopy and was superior to that of HSG.^{35, 36} 

A meta-analysis of 18 studies compared HyCoSy using Echovist-200 with both HSG and laparoscopy. Concordance between HyCoSy and HSG in the individual studies ranged from 70% to 91%. The overall concordance for detecting tubal pathology was 83%.⁶⁷ This was confirmed by a subsequent extensive review that demonstrated an occlusion detection concordance of 100% and a patency detection concordance of 86% when compared with HSG.⁶⁸ HSG is slightly more accurate in detecting patency and HyCoSy slightly more accurate in detecting occlusion.⁶⁹

Hysterosalpingo-contrast sonography versus endoscopy

Holtz et al. reported a meta-analysis of three studies comparing laparoscopy with HyCoSy using Echovist-200. In reviewing the results on over 1000 patients. Results were confirmed in 68.7% patients.  HyCoSy showed false occlusion in 10.3% and false patency in 6.7% of patients.⁷⁰ More recent studies have confirmed that the accuracy of HyCoSy, using laparoscopy as the standard, varies from 70 to 89% in the diagnosis of both proximal and distal obstructions of the fallopian tube, and 65% in the diagnosis of peritubal disease.^{5, 71, 72} Using hysteroscopy as the comparative standard, HyCoSy was found to be over 93% sensitive and over 97% specific in demonstrating tubal patency.⁷³ A meta-analysis that included 20 comparisons between HyCoSy and laparoscopy reported a concordance varying from 80% to 90.9%.⁶⁷

Although HyCoSy does not quite reach the accuracy of HSG in the diagnosis of tubal disease, it is rapidly becoming the first-line method of tubal imaging because of its availability, convenience, comfort, and cost-effectiveness. It is particularly useful in secondarily diagnosing uterine abnormalities such as intramural leiomyomata, fusion defects and polyps. Its greatest failure is in the diagnosis of peritubal adhesions. Because of its ability to evaluate the uterine contour as well as tubal anatomy, it may be considered the imaging method of choice prior to in vitro fertilization.