Chapter 39: GUIDED PROCEDURES USING TRANSVAGINAL, TRANSPERINEAL, AND TRANSRECTAL SONOGRAPHY

Previously, ultrasound-guided procedures were essentially exclusively performed abdominally. With the recent improvement in vaginal ultrasound equipment and technique, vaginal ultrasonographically guided puncture procedures have replaced abdominally guided puncture in many cases. Certainly, procedures such as amniocentesis will always be performed abdominally, but marked changes in assisted reproductive technologies, specifically in vitro fertilization and human egg retrieval, have paved the way for vaginally guided techniques. The aim of this chapter is to describe the various vaginally guided puncture procedures, both those being presently performed and those being studied experimentally, and the experience gained thus far.

Ever since Smith and Bartrum¹ performed percutaneous aspiration of intraabdominal abscesses in 1974 and Gerzof et al used an abdominal catheter placed sonographically to drain purulent pelvic collections,^2,3 ultrasound-guided puncture procedures have been used to achieve both diagnostic and therapeutic goals. The advantages of these procedures over traditional surgical means are multiple: ease of technical mastery, accurate needle placement, rare injury to adjacent organs, portability, low cost, speed of administration, and, probably above all, patient satisfaction and comfort. Possible risks, although quite rare, include bleeding, infection, inadvertent puncture of organs, and, in the case of multifetal reductions, miscarriage.

Puncture procedures necessarily traverse three-dimensional space, not only the two-dimensional view of an ultrasound machine screen. This third dimension, because of the properties of ultrasound, is thinnest at the focal range of the probe and is inversely proportional to the operating frequency of the transducer crystal. This concept is called the slice thickness artifact and means that the operator must take the third dimension of the image into account: At times the tip of the needle used for puncture procedures appears to be within the structure at which it is aimed, but in reality it is in front of or behind the structure imaged.

When puncture procedures are performed abdominally, 1 of 2 techniques is employed: needle guide or free hand. Often, an operator uses a transabdominal needle guide when a needle has to be directed into a body part. With increasing experience, the free-hand approach has been successfully practiced, with the ease of handling the needle and keeping it in the scanning plane. The limitation of this technique is that occasionally the transverse section of the needle, excluding the tip, is what is imaged; the actual site of the needle tip may be unknown to the operator. When the puncture procedure is being performed abdominally, a quick readjustment of the scanning plane may be done to find the needle tip.

When puncture procedures are performed transvaginally, there is limited mobility of the probe and the needle, making the free-hand approach cumbersome. A fixed needle guide attached to the probe shaft allows easier visualization of the entire length of the needle within the scanning plane and better control for exact needle placement. An automated spring-loaded puncture device (Labotect, Gottingen, Germany) is frequently used in our units (Figure 39-1). This device, when mated to the shaft of the vaginal probe, provides extreme accuracy and precision, and its high-velocity release makes the procedure virtually painless: No anesthesia or analgesia is used. A software-generated directional dotted double line functions as a biopsy guide and is displayed on the monitor. This guide measures the actual depth necessary for effective needle penetration; the depth setting on the puncture device is then adjusted accordingly. A trigger on the device is released, and the needle (available in gauges 16, 19, and 21) is propelled into the desired structure, where the aspiration or injection can then be performed. Historically, this technique was first used for ovum pickup in assisted reproductive technology programs,⁴ but it was quickly abandoned secondary to the need for reloading and reshooting for each new follicle aspirated.

The automated puncture device is used where extreme accuracy is needed for any needle placement controlled and guided by the transvaginal probe.^{5,6,7,8,9, and 10} The accuracy of the needle placement tested in vitro by our laboratory was found to be between 1 and 2 mm. Manual needle introduction is less accurate and potentially more painful as the relatively slower forward motion of the needle frequently displaces mobile targets rather than immediately penetrating them. In most cases, where manual needle insertion is used, analgesia, and sometimes local anesthesia, is used via intravenous meperidine (25 to 50 mg) and diazepam (5 to 10 mg), with local injection of 2 to 3 mL of 1% lidocaine through the needle guide into the vaginal mucosa at the intended puncture site.

A general guideline of the technique employed for the majority of the common puncture procedures performed will be described, although several procedures follow different protocols, and will be discussed separately.

Punctures are usually performed with the guidance of 5.0 to 7.5 MHz vaginal transducer probes, with a needle guide that is attached to the shaft of the probe. Several companies offer disposable needle guides and attachments, but we use all stainless steel equipment that is sterilized (autoclaved) after each procedure. A software-generated fixed "biopsy guide" line is displayed on the ultrasound monitor screen, which marks the path of the entering needle. Most of these biopsy guides provide the operator with a centimeter scale with which to measure needle depth (inset of Figure 39-1). Needle gauges ranging from 14 to 21 are employed, depending on the nature of the procedure itself; the narrowest (thinnest) possible needle able to perform the desired task should be used. For better imaging, the "zoom" feature of the equipment should also be used as frequently as possible.

Recently, transrectal and occasionally transperineal sonography have been found to be valuable to guide difficult dilation and curettage (D&C), intracavitary tandem placement, and cerclage placement. The basic technique and application of this modality is discussed at the end of this chapter, after the puncture procedures are discussed in detail individually.

Patients should be counseled before the procedure and asked to sign routinely used (hospital) consent forms. Specifically designed consent forms are usually used in specialized procedures, such as multifetal reduction and puncture of ectopic pregnancies.

The procedures should be documented by still images, videotape recordings, or both. After the final withdrawal of the needle, the pelvic structures and cul-de-sac must be observed sonographically for approximately 10 minutes, or as necessary, to detect possible internal bleeding. The patients are then rescanned after a 2- to 3-hour observation period, to check for internal bleeding or previously undetected complications. A detailed written report describing the procedure should be promptly completed.

A short description of the more commonly performed transvaginally directed punctures (see list) will be described, as will the indications and specific attributes.

1. Transvaginal puncture/catheterization procedures in assisted reproduction

2. Punctures of ovarian cysts

3. Multifetal pregnancy reductions

4. Treatment of ectopic pregnancy

5. Drainage of pelvic (fluid) contents

6. Culdocentesis

7. Coelocentesis

Transvaginal Puncture/Catheterization Procedures in Assisted Reproduction

Oocyte Retrieval

The first transvaginal puncture performed was oocyte retrieval, after transabdominally and transurethrally directed transvesicle follicle aspiration had been routinely employed^{11,12,13, and 14} for this procedure. Transvaginally performed oocyte retrieval is now the technique of choice worldwide; the technique consists of a needle guide attached to the vaginal probe enabling guided needle placement under real-time observation (Figure 39-2).^15,16 Some programs use the automated spring-loaded puncture device that is attached to the shaft of the vaginal probe.⁴

Complications arising from puncture technique are rare, but have been described intermittently in the literature,^{17,18,19, and 20} such as perforation of a pelvic vein or postretrieval pelvic infection. In 1996, Coccia et al²¹ described a complication of acute abdomen following dermoid cyst rupture in a patient who had immediately previously undergone transvaginally guided oocyte retrieval. Other complications such as vaginal perforation and rectus sheath hematoma have also been described.^22,23

Ludwig et al prospectively reviewed their experience with over 1000 oocyte retrievals, analyzing the perioperative and postoperative complications. They found that no complications were caused by sedation or general anesthesia.²⁴ Vaginal bleeding was observed in 2.8% of procedures, but no cases of intra-abdominal bleeding. A single ureteral lesion occurred. No case of documented pelvic infection was noted, but one case of unexplained fever did occur. Severe ovarian hyperstimulation syndrome (OHSS) occurred in 2.7% of the cases. Although most patients tolerated the oocyte retrieval well, 3% of the patients experienced severe to very severe pain after the procedure, and 2% of patients were still suffering from severe pain 2 days after the procedure. The pain level increased with the number of oocytes retrieved.

As pelvic fluid volume is known to be associated with pain, possible intraperitoneal bleeding, and development of OHSS, Shalev et al evaluated the fluid volume in the pelvis immediately after and 3 to 5 days after transvaginal sonography (TVS)-guided oocyte aspiration.²⁵ Pelvic blood volume was measured by three-dimensional (3D) ultrasound examination 3 to 5 hours after aspiration and prior to embryo transfer (2 to 3 days after retrieval). When all study variables were held constant, the number of oocytes and serum estradiol level proved to be significant predictors of the amount of fluid in the pelvis. If two-dimensional (2D) ultrasound was used, the best predictor of fluid volume after aspiration was the width of the fluid scanned in the pelvis. This group concluded that the amount of blood in the pelvis following transvaginally guided oocyte retrieval was within acceptable clinical limits, which has important implications for the improvement of postprocedural care.

Intrauterine Transfer of Fertilized Oocytes

Although this procedure cannot strictly be considered a puncture, it is mentioned because TVS is occasionally used for directing embryo transfers. Many centers do not use ultrasonography in this setting, and if they do, they often use transabdominal sonography as the preferred modality, being easier to perform and less interfering with the external os manipulations.²⁶ Rarely performed, embryo transfer directly through the myometrium into the uterine cavity is assisted by using the transvaginal probe.^27,28

Tubal Catheterization and/or Embryo Transfer

The fallopian tubes may be reached through a transvaginally guided catheter advanced through the cervix,²⁹ and both diagnostic and therapeutic techniques are employed: Diagnosis of tubal patency via injecting and observing fluid passage into the pelvis may be done,³⁰ and fertilized ova may be carried into the ampullary portion of the tube.^31,32 One group prefers the transabdominally guided route and was dissatisfied with the transvaginal approach.³³

Puncture of Ovarian Cysts

Puncturing an ovarian cyst is probably the simplest of all the puncture procedures, and it is one of the more common indications for the technique. Technically, the center of the cyst is targeted and the needle is placed and maintained in the middle of the slowly shrinking sonolucency while direct sonographic visualization is maintained throughout the entire aspiration (Figure 39-3). Technically, this technique has few complications and is well tolerated. The aspirated fluid is usually submitted for cytologic evaluation, although a negative result may in fact be a false-negative result.

Recently, more reproductive endocrinologists have used selective follicular aspiration during the selection phase of infertile women with polycystic ovary syndrome (PCOS) under controlled ovarian hyperstimulation in patients undergoing in vitro fertilization (IVF).³⁴ No cycles with moderate or severe OHSS occurred in the ultrasound-guided transvaginal follicular aspiration group, while 9 of 13 patients had OHSS in the untreated group. The oocyte maturation rate, fertilization rate, and cleavage rate in the aspirated group were all significantly higher than the untreated group. They conclude that follicular aspiration during the selection phase can significantly decrease severe and moderate OHSS prevalence in patients with PCOS undergoing controlled ovarian hyperstimulation.

Several investigators have reported series of successful punctures of ovarian cysts and symptomatic benign cystic pelvic lesions, such as paraovarian cysts and peritoneal inclusion cysts, without complications.^{35,36,37, and 38} The concern of inadvertent tumor cell spillage from an unsuspected malignant ovarian cyst prevents many from using this technique more frequently.

Recurrence and persistence of ovarian cysts are another concern, although they appear to occur in approximately 30% of cases. Weinraub et al³⁹ and Ron-El et al³⁸ followed a group of 35 women with ovarian cyst puncture and performed a second aspiration for recurrence in 14 of the women. Weinraub et al found that 3 parameters proved important in predicting the outcome of aspiration: patient age and the location and maximum diameter of the cyst. The older the patient and the larger the cyst, the poorer the prognosis for cure that was suggested. They concluded that aspiration of large cysts should be avoided in older women.³⁸ Bret et al published 2 papers using TVS for aspiration of ovarian cysts and found a high recurrence rate after cyst aspiration in pre- and postmenopausal women (48% and 80%, respectively).^40,41

Caspi et al investigated the recurrence rate of 107 sonographically guided aspirated ovarian cysts from 1 to 6 years after puncture.⁴² In 42 cases, the initial cyst aspiration constituted definitive therapy. In 65 cases, the cyst recurred (complete recurrence), and in this group 35 women had subsequent surgical management. Twenty-seven women with cysts smaller than 5 cm (incomplete recurrence) were allocated to the follow-up group. In 65% of women overall, surgery was avoided and no major complication was encountered.⁴²

In the cardiovascular radiology literature, transvaginal aspiration of ovarian cysts has been utilized as a viable alternative to surgery in patients who are poor operative candidates due to their cardiovascular disease. Long-term follow-up has been obtained. Duke et al⁴³ followed 24 women with ovarian cysts who underwent 34 transvaginal drainages over a 6-year period. Long-term follow-up of patients revealed a recurrence rate of 75%, but when recurrences do occur, repeat transvaginal aspirations may be considered in the symptomatic patient who would not be a suitable surgical candidate.

Recently, attention has turned to the possibility of using combined ultrasound-guided puncture of cystic adnexal masses and cytologic analysis of the aspirates to prevent or at least limit surgical intervention.^44,45 Our group studied 43 cystic ovarian masses that were sonographically characterized using the Sassone scoring system and then aspirated the cysts if they exhibited benign-appearing morphology.^45,46 All 25 cysts that were benign by TVS, histology, or both were also cytologically benign, as were an additional 11 cysts that were not resected. Seven cytologically and histologically malignant cysts, which were not aspirated but resected and evaluated postoperatively, had high TVS scores, which were suspicious for malignancy. The combination of TVS and aspiration cytology was considered valuable by the group, especially in obviating an operative procedure in the diagnosis and management of cysts having low TVS scores and benign cytology.⁴⁵ The concern of inadvertent ovarian cancer tumor spread has caused several gynecologists to oppose any ovarian cyst puncture procedure, and this has caused controversy over use of the procedure.⁴⁷ This should be considered an investigational procedure at present.

Another combination approach in the management of ovarian cysts includes both sonographically guided drainage and subsequent laparoscopic excision.⁴⁸ This technique is particularly useful in very large, unilocular cysts, which would be contraindicated for the laparoscopic approach alone. Nagele and Magos, 2 gynecologists who specialize in minimally invasive therapy and endoscopy, have employed this combined technique in a patient with a 24 3 10 3 20 cm unilocular cyst. Transabdominally guided insertion of a Veress needle into the center of the cyst through the patient's umbilicus was performed and the needle was attached to a suction pump: 3100 cc of fluid was drained off. As the cyst had shrunk to one-third its original size, standard laparoscopic techniques were then employed and the cyst wall was removed without incident.⁴⁸

Several recent authors have used combined transvaginal aspiration and instillation of a sclerosing agent in the hopes of decreasing the known high recurrence rate.^49,50 Although most sclerosing agents were used in the setting of endometriomas and endometriosis, Kukura et al described a puncture technique for use in simple-type cysts.⁴⁹ After complete emptying of the cyst, they injected an amount of sterile 96% alcohol equaling 50% of the volume of the cyst. They left the alcohol within the cyst for 5 minutes, and then aspirated completely. They treated 152 patients in this manner whose cysts contained a cyst volume of fluid between 40 and 140 mL. After 3 months recurrence occurred in 8 patients (5%) in most of whom the volume of the cyst had been more than 100 mL. Mesogitis et al from Greece evaluated the technique of TVS-guided aspiration with methotrexate injection in the management of 162 female patients with simple or endometriotic ovarian cysts. Of 162 patients, 148 were available for follow-up. At follow-up ultrasound, cysts had disappeared in 124 patients (83.8%) and persisted in 16.2%. Cyst diameter proved to be a significant prognostic factor for cyst resolution (P = .01) No major complications were observed.

The aspiration of endometriotic cysts is also an area of controversy, and is considered to be relatively contraindicated, although recently more work has been undertaken in the subset of patients undergoing assisted reproductive technologies. De Crespigny et al aspirated 6 endometriomas among 28 punctured ovarian cysts without incidence.⁵¹ Aboulghar et al aspirated 21 ovarian and endometriotic cysts using a transvaginally guided puncture technique and found a low incidence of recurrence and improvement in symptoms.⁵² Technically, the procedure is simple, although patient selection is particularly important. Indications for the aspiration of endometriotic ovarian cysts include avoiding impairment of folliculogenesis,⁵³ following sequential follicle growth, improving follicular retrieval techniques,⁵⁴ and the improvement of symptoms associated with large endometriotic cysts and adhesions. Certainly, the recurrence rate within 3 months is high, quoted between 66% and 83%, although often these recurrences do not occur until after the IVF program has been completed. Messalli et al from Italy used ultrasound-guided aspiration and alcohol sclerosis in patients with endometriomata.⁵⁵ In 9 patients, no recurrence was observed after 21 months, and in only 1 case recurrence was noted after 6 months.⁵⁵ Sclerotherapy with 5% tetracycline has been found to be a simple alternative to potentially complex surgical treatment of ovarian endometriomas before in vitro fertilization.⁵⁶ Fisch and Sher performed sclerotherapy rather than laparoscopy for 32 women in their infertility practice. Tetracycline 5% was instilled into the cyst after the endometriotic contents were completely aspirated, approximately 5 to 10 mL, and then removed. Ultrasound was performed 6 weeks later to assess the efficacy of treatment. They used resolution of the endometrioma and subsequent IVF pregnancy rate as their main outcome measures. Complete resolution was observed in 75% (24 of 32) patients at follow-up exam. Repeat aspiration was required in 8 patients and repeat treatment with tetracycline was required in 2. Only 1 patient in the group did not ultimately respond. IVF was then performed in 28 of these patients; ongoing gestation resulted in 16 (57%) from the next cycle. These authors then concluded that sclerotherapy with 5% tetracycline is a simple, effective, and in their limited series, safe alternative to surgical intervention for treatment of endometriomas before IVF. Agostini et al sclerosed endometriotic cysts with methotrexate after aspiration and found a 28.6% recurrence rate.⁵⁷

Combination modalities have also been recently employed for the treatment of endometriosis. Acien et al performed a double-blind randomized controlled trial to evaluate the results of ultrasound-guided aspiration of endometriomas under the effect of gonadotropin-releasing hormone (GnRH) analogues and a possible additional beneficial effect by leaving recombinant interleukin-2 (rIL-2) in the cysts.⁵⁸ Twenty-four women with endometriosis-related symptoms and TVS evidence of endometriomas greater than 3 cm who were initially sent to this group for laparotomy and conservative surgery were included. They found moderate clinical results after treatment with drainage plus GnRH analogues and significantly improved results in women having received rIL-2 intracystically. Although the rates of recurrence were similar in the 2 groups, the time until recurrence was significantly greater when rIL was used. Although a subset of patients in both groups did need definitive surgery during the following 30 6 12.7 months, this group did conclude that transvaginal ultrasound-guided puncture and aspiration of endometriomas under suppressive therapy with GnRH analogues do have some value for endometriosis treatment, and that rIL-2 left in these cysts increases these beneficial effects significantly.

Another area of controversy is the clinical benefit of aspirating "simple" ovarian and adnexal cysts. Ovarian cysts with simple architecture have a sonographic appearance that includes thin and smooth walls with no septations or irregularities along the inner surface and are filled with sonolucent fluid. The natural history of these cysts has been studied extensively in both pre- and postmenopausal women. An important review paper investigating these cysts in postmenopausal women used transabdominal and transvaginal ultrasonography in 184 asymptomatic postmenopausal women.⁵⁹ Forty-nine women with 72 cysts were sequentially followed sonographically over a period of 3 to 23 months. Thirty-eight of the 72 cysts (53%) disappeared completely, 20 cysts (28%) maintained their size, 8 cysts (11%) enlarged, 2 cysts (3%) reduced in size by 3 mm or more, and 4 cysts (6%) both increased and decreased in size at different follow-up examinations. In this study, no relationship between presence or activity of cysts and type or length of hormone replacement or time since menopause was found. The important conclusion of this study is that simple adnexal cysts smaller than 3 mm in postmenopausal women are most likely benign and may be managed conservatively with serial ultrasound examinations. Goldstein evaluated the conservative management approach in small cystic masses in postmenopausal women and agreed with the possibility of conservative management in selected cases.⁶⁰

There is also clinical evidence that a significant percentage of simple ovarian cysts resolve spontaneously and require no treatment. An excellent randomized trial was conducted by Zanetta et al where 278 women with simple cysts were included.⁶¹ This group compared the results of needle aspiration and simple observation for a 6-month follow-up period, where the main outcome measures were resolution of the cyst or the development of a malignancy. The rate of resolution was 46% with aspiration and 44.6% with observation alone; only the diameter of the cyst was a significant independent prognostic factor for resolution. Of the patients being followed, no adverse sequelae or evidence of malignant transformation was noted from 10 to 58 months after initial diagnosis of the cyst. They concluded that a short expectant management period is not risky.

In conclusion, ultrasound-guided aspiration of ovarian cysts offers a simple outpatient strategy for immediate pain relief and effective treatment in a majority of cases. The important criteria for selection include size and location of the cyst, morphologic characterization by sonography, and the patient's age.

Multifetal Pregnancy Reduction

The increasing use of ovulation-induction medications and the increasing number of assisted reproductive technology programs over the past 10 years have resulted in a recent dramatic increase in multifetal pregnancies of high order. It is known that the risk of pregnancy complication is proportional to the number of fetuses and, conversely, that the probability of achieving a term pregnancy is inversely proportional to the number of fetuses.^{62,63,64,65,66, and 67} Fetal and maternal complication rates for multifetal gestations are high: Mothers of triplets have a 35% risk of postpartum hemorrhage and at least a 20% rate of preeclampsia.⁶⁴ An extensive study from the Oxford Perinatal Epidemiological Unit compared mortality rates among infants born of multifetal pregnancies between 1975 and 1983, and it may be concluded that the risk increases dramatically in all orders after twins.⁶⁵ In addition to decreasing gestational age with increasing fetuses, the concomitant decrease in birth weight among multifetal pregnancies is striking: The mean gestational duration for twins is 260 days (37 weeks), for triplets is 247 days (35 weeks), and for singletons is 281 days (40 weeks).⁶⁶ The mean gestational age for quadruplets was only 29 weeks in a series where both the time of ovulation and time of delivery were accurately known.⁶³

The puncture procedure known as multifetal pregnancy reduction (MFPR) was developed to reduce the number of fetuses in utero to a lower, more desirable number, and it was first performed as a transcervical ultrasound-guided suction aspiration.⁶⁸ The transabdominal ultrasound-guided technique was first developed by French physicians in the early 1980s⁶⁹ and was soon adopted by many others.^{70,71,72,73,74,75,76,77, and 78} Once the widespread use of transvaginally guided ovum aspiration for in vitro fertilization and embryo transfer was in place, these techniques were successfully modified and applied for multifetal reduction procedures.^{6,7,10,79,80,81,82, and 83}

This technique of first-trimester multifetal pregnancy reduction performed for reduction of fetal number must be contrasted to second trimester selective fetal reduction, which is uniquely performed for anomalous fetuses, usually as one of a set of twins or, less commonly, triplets. This selective technique is almost always performed with transabdominal ultrasound guidance, and often after a second- trimester ultrasound anatomic survey reveals a fetal structural defect, or second-trimester genetic amniocentesis reveals a chromosomal or biochemical abnormality. The scope of this section focuses only on first-trimester transvaginal and transabdominal multifetal pregnancy reduction performed for fetal number.

Technically, the transabdominal or transvaginal MFPR is accomplished in the following manner. The patient first undergoes an extensive counseling session and a "baseline" mapping scan to evaluate the viability and chorionicity of each of the fetuses (Figure 39-4A). The counseling session includes information given to the patient regarding the technique, the complication and pregnancy loss rates, and outcome measures, such as final gestational age and route of delivery. In either of the approaches (transabdominal or transvaginal), the area of the fetal heartbeat is sought; once the desired location has been chosen, the needle tip is placed, either manually or using the automated device (Figure 39-4B). In our transvaginal approach, a fixed biopsy guide is visible on the ultrasound monitor screen to assist in accurate needle placement. Approximately 0.5 to 1 mL of 2 mEq/mL potassium chloride (KCl) solution is injected slowly through the 20- or 21-gauge needle for cessation of fetal heart activity; the heartbeat for each injected fetus is then observed sonographically for 5 to 10 minutes to ensure complete cessation. If the desired technique is aspiration of early 6- to 7-week embryos, 17- to 18-gauge needles are preferred.⁸⁴ The patient is then observed in the ultrasound unit for several hours and rescanned before discharge and rescanned after the procedure according to the protocol of the program. The resolution of the reduced fetus and chorionic sac is monitored intermittently throughout the first 2 trimesters or for as long as it is visualized. Occasionally, a subchorionic hematoma is seen along the line of needle penetration, but it is self-limiting (Figure 39-5); if the placental sites of the remaining fetuses are not compromised, no further complication occurs.

We first published our experience with MFPR in 1993, which included the results of our first 134 cases.⁷ The majority of these (94 of 134) were performed using the previously mentioned automated puncture device and a 21-gauge needle. Our conclusions from the study were that the data indicate a very low maternal complication rate (usually infectious morbidity) and an acceptable loss rate of the entire pregnancy, which compares favorably with the rates for MFPR performed transabdominally in other centers. First-trimester sonography before MFPR allows for a basic preliminary early anatomic fetal survey, which can pick up a wide variety of major and minor structural fetal defects that also may aid in fetal selection to reduce.⁸⁵

We expect that the prematurity rates of these reduced pregnancies would approach those of nontreated singletons, twins, and triplets,^67,68 although there is a continuous debate regarding the risk-to-benefit ratio of reducing twins or triplets. There is very little doubt that MFPR is beneficial in cases of 4 or more fetuses. Two well-designed studies have examined the outcomes of reduced triplets and twins as opposed to nonreduced triplets and nonreduced twins.^86,87

Macones et al⁸⁶ compared the perinatal outcome of triplet pregnancies reduced to twins with that of continuing twin and triplet pregnancies. According to the results, MFPR of triplets to twins yielded an improved outcome compared with nonreduced triplets, and a similar outcome when compared with nonreduced twins. Lipitz et al⁸⁷ compared the outcome of triplet pregnancies managed expectantly or reduced to twins by MFPR. There were lower incidences of prematurity, low and very low birth weight infants, pregnancy complications, and neonatal morbidity and mortality in the reduced group. They concluded that the reduction of triplets to twins resulted in an improved pregnancy outcome without excessive pregnancy loss when compared with the outcome of triplets managed expectantly.

A study by Smith-Levitin et al⁸⁸ compared the pregnancy complications and birth weights of remaining twin fetuses after MFPR with expectantly managed triplets and nonreduced twins. Fifty-nine twin pregnancies resulting from MFPR, 54 triplet pregnancies, and 88 twin pregnancies were retrospectively reviewed. Twins after MFPR and nonreduced twins were less likely to have preeclampsia than were triplets and were less likely to be delivered before 36 weeks. Reduced twins were similar to nonreduced twins in all parameters studied. They concluded that MFPR decreases the incidence of pregnancy complications, prolongs gestational age, and increases birth weight compared with expectantly managed triplet pregnancies.

As previously mentioned, the 2 major technical approaches in MFPR that have been developed are the transabdominal route and the transvaginal route, where the injection of KCl solution is performed through a needle guide or using an automated device. Occasionally, centers use the transcervical approach in much the same way. Much controversy exists on the superiority of route. By far, the technique most commonly used has been transabdominal needle insertion of KCl solution into the fetal thorax. The major difference in approach is gestational age at time of procedure: The transabdominal MFPR is usually performed at a gestational age that is 11½ to 2 weeks more advanced than that at which the transvaginal approach is used for both visualization and technical reasons. In a collaborative study evaluating the cumulative database among the world's largest centers, the differences in efficacy and complication rates by transabdominal, transcervical, and transvaginal MFPR were analyzed.^89,90 Data had shown that the transvaginal and transcervical routes appeared to have a higher loss and complication rate than that using transabdominal sonography (TAS); therefore, the technique using TVS has been abandoned except for very specific cases decided on an individualized basis.^90,91

Timor-Tritsch et al reported on 290 consecutive cases of MFPR and compared the TA versus the TV approach in 2004.⁹² Two hundred and three cases (70%) were performed transabdominally, 75 cases (25.9%) were performed transvaginally, and 12 cases (4.1%) used both routes. The indications for the transvaginal route were extreme obesity, abdominal scars, or if the lower fetus could not be reached transabdominally. As previously noted in other studies, the MFPR success rate was higher with the transabdominal route compared with the transvaginal route. Total pregnancy loss rates of MFPR were 3.5% (7 of 203 reductions) for the TA route and 13.3% (10 of 75 reductions) for the TV route (P = .004).

Significant differences in favor of the transabdominal route were observed for starting with triplets or finishing with a single fetus. They conclude that the transvaginal route should be reserved only for cases in which the transabdominal approach is hard or impossible to perform.⁹²

Recently, in 2001, Evans et al published the most recent data on improvement in outcomes of MFPR with increased experience.⁹¹ More than 3500 cases from 11 centers in 5 countries were evaluated: These centers all had extensive experience in the performance of the procedure. This group found that with increasing experience there has been considerable improvement in outcomes, with decreases in both rates of pregnancy loss and in prematurity. Overall loss rates correlated strongly with starting and finishing numbers, with improvement of outcome for smaller starting and finishing numbers. Birth weight discordance between surviving twins was increased with greater starting number, and overall the proportion of cases with a starting number greater than or equal to 5 decreased significantly from 23% to 12% over the last decade. This group clearly confirms that the technique is a safe and effective option for multifetal pregnancies, and that the outcome has improved significantly with experience. Reductions from triplets to twins and recently from quadruplets to twins carry outcomes as good as those of unreduced twin gestations.⁹¹

Pregnancy of very high order uses a special technique that is not often required. The groups led by Evans and Berkowitz have successfully reduced multifetal pregnancies (in 2 steps) having 9 and 8 fetuses, respectively. Monteagudo et al reported on a 12-to-2 MFPR that was successfully carried out in 3 steps over 2 weeks.⁹³ The concern of impaired coagulation in the presence of large volumes of nonviable fetal tissue has been addressed in a study in which no significant clinical or laboratory expression of coagulopathy was seen after MFPR.⁹⁴

Selection of the final number of fetuses is complex, and many factors must be included in the decision, including the potential for reasonable perinatal outcome. Based on the improved perinatal outcome for twins, the suggestion that 2 fetuses remain after the MFPR in quadruplet or triplet pregnancies seems reasonable. For these same reasons, twin pregnancies should not as a rule be offered selective reduction to singletons, although there will always be the rare case of twins where, for physical or emotional reasons, a reduction may be offered.

Many patients undergoing MFPR are candidates for genetic counseling and testing. Maternal serum α-fetoprotein levels are not useful in these patients.^95,96 Amniocentesis following MFPR has recently been studied in several articles in the literature. Selam et al performed a cohort study on 127 patients who underwent amniocentesis after MFPR and compared the pregnancy loss observed with 167 patients who did not have amniocentesis after MFPR.⁹⁷ This group found a pregnancy loss rate of 3.1% (4 of 127) in the amniocentesis group compared with 7.2% (12 of 167) in the control group. They concluded that amniocentesis after MFPR does not increase the risk of pregnancy loss—a conclusion that has been corroborated by a recently published study by McLean et al in 1998⁹⁸ and our own group.⁹⁹

Finally, a comment on the ethical issues raised by this procedure, which range from duty of preservation of life all the way down to less complex issues, such as nomenclature.¹⁰⁰ Berkowitz and Lynch suggested the term reduction of multifetal pregnancy, because this term illustrates the nature of the procedure and does not emphasize the "selective" aspect of the process.¹⁰⁰ The extent to which ethical issues are discussed with patients may be individualized, although common aspects that should be addressed include the preservation of the pregnancy and the duty not to destroy human life without necessity for that preservation. There are several well-written treatises addressing these moral issues by leaders in the field: Berkowitz and Lynch,¹⁰⁰ Evans et al,⁷⁵ and Hobbins.¹⁰¹ The consensus among these groups concludes that this procedure should be offered to patients as a modality to preserve a desired, viable pregnancy and that the physician must advise the modality that results in the least harm and most good for both mother and fetus.

Treatment of Ectopic Pregnancy

Historically, the treatment of ectopic pregnancy was surgical: The affected tube was resected. Recently, more conservative, nonsurgical approaches, such as close observation or parenteral or local injection of methotrexate, have gained popularity. The introduction of TVS, with a high-frequency, high-resolution probe, has greatly improved diagnostic accuracy of ectopic pregnancy. This accuracy then enables both the sonographer and the clinician several different approaches to the problem that were not available previously. Ectopic gestations may be classified,¹⁰² which then allows a reasonable algorithm for management regimens.

The injection of an "ectopic embryo" by a TVS-guided needle is technically similar to the previously described oocyte aspiration or MFPR and, if successful, saves the patient from a more invasive surgical procedure. Feichtinger and Kemeter first described the transvaginally guided needle puncture in this setting, where the affected tube, which contained a nonviable conception (no fetal heartbeat), was injected with 10 mg of methotrexate in 1 mL of solvent.^103,104 Transvaginally guided injection of ectopic gestations may be performed in tubal, cornual, and cervical locations, and the special aspects of each location will be expanded upon.

Tubal Pregnancy (Salpingocentesis)

Since the original case report by Feichtinger and Kemeter, many centers have used salpingocentesis as an additional tool in the treatment of ectopic pregnancy, and several of these centers have reported on their findings,^{5,103,104,105,106,107,108,109,110,111,112,113,114,115,116,117,118, and 119.}

Monteagudo et al described a series of consecutive cases of live ectopic pregnancies managed with ultrasound-guided local injection of methotrexate (MTX) or potassium chloride (KCl) in 2005.¹²⁰ Eighteen consecutive women with live and unruptured ectopic pregnancies (tubal, cornual, or cervical) were treated with transvaginally guided puncture and injection of the ectopic pregnancy. The "tubal ring" containing the ectopic embryo was imaged sonographically, and the needle was introduced by using the automated puncture device into the area of the beating fetal heart and following the puncture path line on the screen. Either MTX or KCl was injected, producing immediate cessation of fetal cardiac activity. The mean gestational age in this series was 6 weeks 6 days and the mean initial β human chorionic gonadotropin (hCG) level was 33,412 mIU/mL. Ten ectopic gestational sacs were injected with KCl and 8 were injected with MTX. There was no difference in time to resolution of the ectopic pregnancies between those injected with KCl and those with MTX. The group concluded that unruptured live ectopic pregnancies of many types can be successfully managed without surgical intervention through local injection of KCl or MTX.

Halperin et al described a set of 12 patients with ectopic pregnancy and fetal cardiac activity, which were managed conservatively with combined local and systemic injection of methotrexate during January 2000 to July 2002.¹²¹ This group was then compared those results to the outcome of 53 patients who had ectopic pregnancy without fetal cardiac activity, and who were treated only by systemic injection of methotrexate during the same period of time.¹²¹ The success rate was similar for the ectopic pregnancies with or without fetal cardiac activity: The success rate was 91.6% (11 of 12) in the group of patients with ectopic fetal cardiac activity and 90.5% (48 of 53) in the group of patients with ectopic pregnancy but without fetal cardiac activity. As for long-term outcome and patency of the tubes, this group reported that 6 of 8 of these patients diagnosed with ectopic fetal cardiac activity, who desired to become pregnant, succeeded in conceiving within 6 months following the combined local and systemic injection of methotrexate.

Several centers have reported series of cases where local puncture and injection of KCl, methotrexate, or both were performed into tubal ectopic pregnancies. Merz et al¹²² treated 30 patients with local injection of methotrexate under ultrasound guidance and was successful in 25 of the 30 (83%). Eighteen of these patients required only a single injection, whereas 7 patients required a second treatment (intramuscular methotrexate) secondary to plateauing hCG levels. There were 5 failures requiring surgical intervention. Subsequent hysterosalpingogram was performed 4 to 6 months after the local treatment and showed tubal patency on both sides in 86% of the patients examined. Mesogitis et al¹²³ reported on the transabdominal approach in the management of 26 patients with tubal ectopic pregnancy: Methotrexate was injected percutaneously into the gestational sac under free-hand abdominal ultrasound guidance using a 22-gauge needle. Complete resolution of the ectopic was noted in all patients with total regression of trophoblastic tissue and no adverse reactions. In 4 cases with increasing β-hCG levels, a second local injection of methotrexate was needed. Perhaps the largest series published is that by Darai et al¹²⁴ where transvaginally guided intratubal injection of methotrexate was performed in 100 patients between 1993 and 1998. Patients were not excluded on the basis of the size of the mass, the term of the ectopic pregnancy, or the initial β-hCG level. The only exclusions were uncertain diagnoses, evidence of rupture, or signs of hemoperitoneum on ultrasound. Complete resolution was obtained in 78 of the 100 cases; of these, 66 only needed 1 intratubal injection and 12 required an additional intramuscular methotrexate injection (1 mg/kg). In this study, local treatment was only successful in 66% of the cases. In cases where embryonal heartbeats were observed sonographically, the success rate of the procedure was 40.9% (9 of 22 cases). In the absence of cardiac activity or when the ultrasound examination showed no embryo, the success rate achieved was 84.6% (66 of 78 cases), which was a statistically significant difference, although many centers do not employ this technique in the absence of a viable fetus with visible cardiac activity.

In a well-designed study by Fernandez et al¹²⁵ published in 1995, a comparison of local injection of methotrexate under ultrasound guidance to laparoscopic salpingotomy for conservative management of ectopic pregnancy was performed. Inclusion criteria were stringent and included a viable ectopic pregnancy visualized by ultrasound and meeting criteria using a scoring system with 6 variables: gestational age, hCG level, progesterone level, abdominal pain, volume of hemoperitoneum, and diameter of hematosalpinx. The success rates, defined by hCG levels returning to normal, were 19 of 20 cases in both groups. In selected cases, methotrexate therapy appeared to be safe and efficient, as was conservative treatment by laparoscopy.

Recurrence of tubal pregnancy after local injection of methotrexate has also been reported and must be considered whenever pregnancy occurs after any conservative tubal therapy. Tarnelle et al¹²⁶ described a case report where a second tubal pregnancy occurred in the same tube as the first one, which had been previously treated with local methotrexate injection. A hysterosalpingogram performed 3 months after the initial methotrexate injection was normal. The recurrent tubal pregnancy was treated by laparoscopic salpingectomy, and the pathology report confirmed unruptured tubal pregnancy. Histologically, no damage to the tube could be attributed to the methotrexate injection, and the investigators concluded that methotrexate has no deleterious effect on tubal epithelium, although recurrent ectopic pregnancy can occur despite a patent tube.

A large series of locally treated ectopic pregnancies to date was described by Gjelland et al¹²⁷ and these patients' reproductive outcome were followed up in a report by Hordnes.¹²⁸ Initially, 80 patients were randomized into 2 groups: local injection of hypertonic (50%) glucose into the gestational sac by TVS or by laparoscopy. Overall, 53 of the 80 women (66.3%) were successfully treated: 82% in the sonography group versus 51% in the laparoscopy group. The follow-up study reported the subsequent intrauterine pregnancy rate of 58% and repeat ectopic pregnancy rate of 16.7%.

It is difficult to make any conclusions on the cumulative experience from the many centers that perform these transabdominal or transvaginal puncture procedures in treating ectopic pregnancies. Although the success rates differ, from 60% to 100%, it is almost impossible to compare the results secondary to different gestational ages, techniques, solutions, and follow-up protocols. It seems that methotrexate injection may be the best choice. Some of the reported failures may not match the strict definition of a failed treatment; several groups, including our own, were quick to intervene if a slowly decaying β-hCG level was observed, or a small accumulation of blood was sonographically visualized in the pelvis. Several conclusions may be made: (1) a slow decrease of the serum β-hCG levels is often observed, especially if KCl is used; (2) severe lower abdominal cramping is a common finding, and the decidual cast may be expelled; and (3) tubal abortions with or without injection frequently cause pelvic blood accumulation and may not be proportional to patient discomfort. Relative contraindications to the procedure appear to include a gestational age greater than 8 completed weeks, a tubal diameter larger than 2 cm, or both.

Cornual Pregnancy

Although cornual pregnancies occur in only 2% to 4% of ectopic gestations, they may create a surgical emergency with great morbidity or even mortality for the patient. A necessary aspect of cornual pregnancy is the thin, although abundantly vascularized, myometrial layer that surrounds this gestation. Rupture of a cornual pregnancy causes rapid hemorrhage and hemodynamic instability. The surgical management is cornual resection, although in many cases hysterectomy is unavoidable. However, unruptured cornual pregnancy could conceivably be managed in the same manner as the injected tubal ectopic pregnancies previously described.

One of the complications of such a local puncture is bleeding from the puncture site. In 1996, Timor-Tritsch et al first reported on a potentially safer route for puncture and injection of cornual ectopic pregnancies.¹²⁹ In this technique, the needle is led into the cornual ectopic pregnancy by first traversing the myometrium and approaching the gestational sac from the medial aspect.¹²⁹ After methotrexate, or if a live heterotopic intrauterine pregnancy is present, KCl solution is injected, the needle is extracted, and potentially less chance for rupture or bleeding occurs. After completing the procedure, the puncture area is observed for 5 to 10 minutes to detect any immediate postprocedure bleeding, if it occurs (Figure 39-6).¹²⁹

Other investigators have described their case reports or very small series of local treatment of cornual pregnancies. Batioglu et al¹³⁰ reported on the successful treatment of a patient with a cornual pregnancy that was documented by both ultrasound and laparoscopy. The patient received 2 doses of methotrexate injected into the gestational sac, which was simultaneously directed by both sonography and laparoscopy. Although the patient's β-hCG levels rose initially, they subsequently fell to nonpregnant levels within 27 days and no further treatment was necessary. Hoek et al¹³¹ reported on their treatment of a patient with an advanced interstitial pregnancy diagnosed by sonography and confirmed by laparoscopy. The patient's treatment consisted of systemic methotrexate followed by local injection of methotrexate into the gestational sac.

Although this unorthodox approach to cornual pregnancies appears to carry a low risk for the patient and avoids extensive surgical manipulations, the small numbers of patients managed in this manner thus far can only suggest that more study is necessary.

Cervical Pregnancy

Cervical pregnancy is the rarest type of ectopic pregnancy, although when undiagnosed or treated too late, it often develops into an acute obstetric emergency. The true incidence of cervical pregnancy is not known because of misdiagnosis and underreporting of cases. The etiology is unknown. However, a history of using an intrauterine device, previous D&C, previous cesarean section, and a history of in vitro fertilization treatments is frequently encountered. Massive hemorrhage is common, and often drastic surgical measures such as vigorous D&C or hysterectomy are the only management options available.

Rubin's criteria¹³² were the first in establishing the diagnosis of cervical pregnancy and are still used today. Duckman¹³³ and Flanagan and Walsh¹³⁴ attempted to define clinical criteria for diagnosis. Ultrasound-based criteria were first suggested by Kobayashi et al¹³⁵ in 1969, but it was not until 1978 that the first cervical pregnancy was diagnosed sonographically.¹³⁶ Since that time, ultrasound has become a key element in diagnosis.^137,138 Our suggestion⁹ for the sonographic diagnosis of cervical pregnancy includes findings that the uterine cavity is empty, the cervical canal is barrel shaped and distended, and the entire gestational sac containing live fetus and placenta is located below the internal os (Figure 39-7). The history, management, and outcome of cervical pregnancy over the past 60 years reflect the improvement in diagnosis and critical care treatment since 1970, after which no further deaths were reported. Treatment has historically been hysterectomy, although cervical suture, bilateral internal iliac artery ligation, descending uterine artery ligation, cervical tamponade, and cervicotomy, alone or in combination with massive blood transfusions, have been performed in an attempt at uterine conservation.¹³⁹

In 1982, Tanaka et al first used intramuscular methotrexate to treat tubal ectopic pregnancies,¹⁴⁰ and this was successfully repeated by others.^{141,142, and 143} The first use of intramuscular methotrexate in the setting of cervical pregnancy was by Farabow et al¹⁴⁴ in 1983 and a combination of methotrexate both intracervically and intramuscularly was reported by Palti et al in 1989.¹⁴⁵ Kaplan et al¹⁴⁶ in 1990, used transabdominal ultrasound guidance to inject methotrexate into the amniotic cavity of a cervical pregnancy with good outcome.

Because TVS is now widely used in the diagnosis of ectopic pregnancy in all locations, our group evaluated transvaginally guided methotrexate injection into viable cervical pregnancies.⁹ Five viable cervical pregnancies at 6 to 8 weeks' gestational age were treated, with 3 of those using the automated puncture device with a 21-gauge needle, as previously described. All 5 cases were successful, and no complication occurred. The serum β-hCG returned to nonpregnant levels within 120 days.

Four additional cases of TVS-guided puncture and injection of viable cervical pregnancies before 71½ weeks were performed by Frate et al.¹⁴⁷ This group injected KCl solution and all cases were successful. Two other groups reported on their experiences in local injection of cornual pregnancy in single case reports: Nomiyama et al¹⁴⁸ and Pretzsch et al.¹⁴⁹ In the case of Nomiyama et al, the treatment was successful; for Pretzsch et al, the patient required postinjection vaginal hysterectomy 5 weeks after treatment due to very heavy vaginal bleeding, although β-hCG levels were noted to be low at the time.¹⁴⁸

Monteagudo et al also reported on our experience with a patient who was found to have a heterotopic pregnancy after in vitro fertilization treatment: 1 viable intrauterine gestation and 1 viable cervical pregnancy.¹⁵⁰ The cervical pregnancy was directly injected with KCl solution under transvaginal ultrasound guidance and the patient closely followed with serial sonograms for the duration of her pregnancy. Although the size of the now nonviable cervical pregnancy mass grew in size during the pregnancy, no further treatment was undertaken and the patient had a primary cesarean section performed at 34 weeks of gestation. Intramuscular methotrexate was given postoperatively, and after a prolonged period of time the cervix returned to normal; at 1 year after the patient's cesarean section, the cervix was normal in appearance both grossly and sonographically.¹⁵⁰

Jeng et al reported on their experience using transvaginal ultrasound-guided treatment of cervical pregnancy in 2007.¹⁵¹ They prospectively reviewed all cases of cervical pregnancies treated conservatively through transvaginal ultrasound-guided therapy at their institution. Thrity-eight cases were identified during the time period of 1993 to 2004. All cases were managed with transvaginal intra-amniotic and intrachorionic injection of 50 mg methotrexate under ultrasound guidance. An additional intracardiac fetal injection of 2 mL KCl was given for those cervical pregnancies with documented cardiac activity. Gestational age ranged from 5.4 to 11 weeks in this series. All cervical pregnancies were successfully aborted, with an average resolution of the cervical mass in 49 days. A mean 4.5-year follow-up showed that, of 21 patients who desired pregnancy, 18 had achieved subsequent successful pregnancies. This group concluded that cervical pregnancies can be successfully managed without surgical intervention through local injection of methotrexate and KCl, and found that this treatment not only ablates the ectopic pregnancy but also preserves the uterus for subsequent pregnancies.¹⁵¹

Ferrara et al described a single case and reviewed the literature for successful management of cervical pregnancies by sonographically guided transvaginal local injection in 2007.¹⁵² Their case concerned a patient who had a recurrent cervical pregnancy, and then they performed a Medline English language search, which identified 90 cases of cervical pregnancy treated with local therapy. A review of the cases identified a mean gestational age at diagnosis of 7.5 weeks. Vaginal bleeding was the most common presenting sign (79%). The most common risk factor encountered was a history of curettage (69%) followed by previous cesarean delivery (35%). An additional dose of methotrexate was needed in 6% of the cases and bleeding requiring alternate procedures was present in 5% of cases, with 1 of those requiring hysterectomy. There were no complications in 81% of cases. As in earilier studies and reviews, this group concluded that conservative management of cervical pregnancy using local injection has a low complication rate and a high efficacy for cure.

Cesarean Section Scar Pregnancy

The recent meteoric rise of repeat cesarean sections in the Western world has been associated with an increase in multiple complications of placentation in subsequent pregnancies, such as placenta accreta and its subtypes. Embryo implantation in a previous cesarean scar is another complication being noted with increasing frequency. Transvaginal sonography and color-flow Doppler provides high diagnostic accuracy with low false positives. Early diagnosis may offer alterative treatment options including transvaginally guided puncture of the pregnancy with direct injection of methotrexate or KCl.

The incidence is still unknown, although Jurkovic et al in 2003 estimated a prevalence of 1:1800 in his local population of women attending their early pregnancy assessment unit.¹⁵³ Seow et al¹⁵⁴ identified a rate of 0.15% in women with a previous cesarean section.¹⁵⁴ IVF-associated heterotopic cesarean section scar pregnancies have also been described as part of a set of twins and triplets, and have their own unique treatment plans and complications.

Systemic administration of methotrexate has been used, with the same rationale as its successful use in tubal ectopic pregnancy. Cesarean section scar pregnancy has been shown to respond well to this treatment, especially in those with β-hCG levels less than 5000 mIU/mL. Local injection of embryocides have recently been reported, including direct intra-amniotic and/or intrachorionic injection of methotrexate, potassium chloride, hyperosmolar glucose, and crystalline trichosanthin.^{151,152,153,154,155,156,157,158,159,160,161, and 162} Although most centers use a 20- or 22-gauge needle for injection, as in other ectopic procedures, one center¹⁵⁴ uses a 16-gauge double lumen oocyte retrieval needle to ensure better aspiration of the trophoblastic tissue via one lumen and injection of methotrexate through the other. Combined puncture procedures and systemic medical treatment protocols have also been promoted, as well as medical treatment combined with direct surgical sac removal.¹⁶³ Of course, standard surgical treatment options exist and may need to occur when other more conservative measures have failed.

Doubilet et al described their clinical experience from 1992 to 2003 regarding sonographically guided treatment of unusual ectopic pregnancies.¹⁶⁴ They described their report as including ectopics that are heterotopic pregnancies and pregnancies occurring at ectopic locations other than the extracornual portion of the fallopian tube. Twenty-seven cases were identified, which included 18 cervical, 6 cornual, 1 tubal heterotopic, and 2 cesarean section (CS) scar implantations. Guidance was via TVS in all cervical pregnancies, 3 of 6 cornual pregnancies, and the tubal heterotopic pregnancy, and via TAS in 3 cornual ectopics. One of the CS scar pregnancies was treated by TVS-guided KCl injection and the other was treated by TAS-guided dilatation and evacuation. This group found that treatment was successful in 25 of the 27 patients, including all 23 patients with an ectopic pregnancy and no concomitant intrauterine pregnancy (IUP). Four patients had IUP and ectopic pregnancies concomitantly, and in 3 the intrauterine gestation resulted in liveborn infants, and in the fourth the IUP was electively terminated. Again, in this study, posttreatment follow-up showed good results, with 8 of the 27 patients having subsequent intrauterine pregnancies.

In conclusion, sonographically guided puncture procedures as treatment for ectopic pregnancies are safe and effective alternatives to surgical therapy, and preserve the uterus and tubes for subsequent pregnancies. Transvaginal sonography performed early in the gestation will identify and classify ectopic pregnancy, and this in turn will allow the selection of the best therapeutic approach.

Pelvic Drainage Procedures

Pelvic drainage using ultrasonography has a long tradition within obstetrics and gynecology. This procedure was first performed by transabdominal ultrasound guidance of a needle or catheter through the vagina.^165,166 The proximity of transvaginal ultrasound tranducers to the cul-de-sac creates an ideal setting for guided aspiration of fluid collections under continuous observation. Septations and loculated fluid pockets are accurately imaged, and sequential aspiration of each loculation may occur with redirection of the needle.

One of 2 techniques for pelvic fluid aspiration is usually performed using TVS guidance: simple aspiration of contents (Figure 39-8) or aspiration plus placement of a catheter to remain in situ for several days. Catheter placement ensures continued drainage and is often used in the treatment of pelvic abscesses. The abscesses are initially drained by a 14-gauge needle, and then a thin guidewire is introduced through the needle while still in place. The needle is then extracted, and a flexible plastic catheter with perforations is slid over the guidewire. The guidewire is then removed, and the catheter is attached to a drainage bag that is fixed to the patient's thigh. The patient is rescanned 2 and 4 days after the procedure, and the catheter is extracted when no further fluid is drained and the repeat transvaginal ultrasound shows no additional pelvic fluid in the cul de sac.

Peritoneal inclusion cyst is the general term used within obstetrics and gynecology for accumulated pelvic fluid entrapped by peritoneal adhesions. They may also be referred to as peritoneal pseudocysts, mesothelial inclusion cysts, and inflammatory cysts of the pelvic peritoneum. Although the pathophysiology is not well understood, a general theory has been proposed. Normal circulation of pelvic fluid produced by active ovaries during ovulation is absorbed by the peritoneum. When the peritoneum is injured, its absorption properties are decreased, and fluid is trapped by peritoneal adhesions. Thirty percent to 100% of women with documented pelvic inclusion cysts have a history of abdominal or pelvic surgery. Despite surgical resection of pelvic inclusion cysts, the risk of recurrence is estimated at 30% to 50%.¹⁶⁷

Povena et al¹⁶⁸ described a set of indications and techniques for the sonographically guided drainage of pelvic fluid collections. They recommended the technique only for large fluid collections (>5 cm), in the case of persistent collection, or when the patient presents with clinical features of infection. They suggested that drainage should be made through a simple percutaneous needle puncture or through a catheter that is kept in place for several days. They recommended the technique highly and found that it guaranteed a high percentage of success and complete resolution with a low percentage of complications.

Transvaginal ultrasound-guided drainage has been shown to be successful in the treatment of postoperative fluid collections, so several groups have attempted to treat women with pelvic inclusion cysts in a similar manner. Sohaey et al¹⁶⁹ performed adhesiotomy with needle aspiration of the cyst in 1 of 7 patients in which some form of conservative therapy was used.

Transvaginal or transrectal sonographically guided drainage for treatment of pelvic abscesses and infected masses in patients who have failed intravenous antibiotic therapies have become more prevalent in the last 10 years. Prior to these guided procedures, most of these women went straight to laparotomy with the steep surgical morbidity it entailed. All authors conclude that endovaginally guided sonographic drainage is a safe and effective procedure for the treatment of pelvic masses and should be considered the route of choice for draining collections not amenable to percutaneous drainage.

In one of the largest studies to date, Gjelland et al evaluated the effectiveness and safety of transvaginal ultrasound-guided aspiration together with antibiotic therapy for treatment of tubo-ovarian abscess in Norway, published in 2005.¹⁷⁰ A total of 449 transvaginal aspirations were performed on 302 women. A total of 282 women (over 93%) were successfully treated for transvaginal aspiration of purulent fluid, together with antibiotic therapy. In the other 20 women (6.6%), surgery was performed after failure of aspiration and antibiotics. No procedure-related complications were diagnosed, and they concluded that this combination modality is an effective and safe treatment regimen for tubo-ovarian abscess, and that the high success rate indicates that it should be a first-line procedure. Other smaller studies confirm these findings.^{171,172,173,174,175,176,177, and 178}

Lee et al in 2002 reviewed their findings on 22 patients with suspected pelvic abscesses refractory to antibiotic therapy who underwent single-step transvaginal pelvic aspiration or drainage.¹⁷⁴ The drainage was successful in 19 of the 22 patients (86%). In the 3 patients that failed aspiration or in which drainage failed, all went on to have surgery despite undergoing repeated drainage procedures. Drainage catheters that were placed in a subset of these patients were left in place an average of 3.7 days. This group found that aspiration alone resulted in a 100% success rate, whereas drainage with catheter placement had only an 80% success rate, possibly due to the severity of the infection, or size and location of the infected collection.

Sudakoff et al reviewed the technique, patient selection, pre- and postprocedural care, and monitoring aspects of transrectal or transvaginal ultrasound-guided drainage.¹⁷⁶

Ho et al in 2008 reported on their experience in Korea with 15 patients who underwent image-guided transvaginal drainage of pelvic abscesses and fluid collections and found the overall success rate to be 87% without any procedure-related complications.¹⁷⁹

Endocavitary 3D ultrasonographic assistance has also been utilized as assistance for transvaginal or transrectal drainage of pelvic fluid collections. Rose et al published their first experience determining the feasibility of using 3D ultrasonography to assist in planning and performing endocavitary drainage of deep pelvic fluid collections.¹⁸⁰ Review of images and medical records of 16 patients in whom endocavitary 3D US was used was performed. Three-dimensional US was assessed regarding its ability to display the relevant structures, whether new information was provided compared with pelvic computed tomography (CT) and conventional 2D US displays, and whether this information altered drainage techniques. This group found that 3D US added information in 11 of 16 patients (69%) that in turn resulted in adjustment of interventional technique in half (8 of 16 patients). Specific features of 3D US that provided new information included the simultaneous display of 3 orthagonal US images, display of reconstructed US image plane orientations not possible with 2D US, and the ability to interactively scroll images through complex structures to assess for communication between the loculations. They concluded that the addition of 3D is feasible and usually adds new and clinically important information.

Drainage of pelvic fluid collections of any etiology is an easy procedure with little morbidity and high patient tolerance. If no catheter is left in place, it may be performed in an office or emergency room setting.

Culdocentesis

The technique for culdocentesis is essentially the same as that for pelvic fluid collections because culdocentesis is a particular type of pelvic fluid aspiration. Due to the accuracy of TVS, this procedure is rarely employed currently. Previously, it was often performed in the emergency room setting when ectopic pregnancy is suspected: Fluid-containing fragments of blood clots or bloody fluid that does not clot gives rise to the possibility of hemoperitoneum secondary to a leaking or ruptured ectopic pregnancy. Culdocentesis may be inconclusive in patients with pelvic inflammatory disease or extensive endometriosis, where the cul-de-sac may be obliterated. The failure to aspirate nonclotting blood from the cul-de-sac does not exclude the diagnosis of hemoperitoneum or ectopic pregnancy, whether ruptured or unruptured. If culdocentesis is performed before TVS, air may be introduced inadvertently into the cul-de-sac, creating an ultrasonic curtain, which then prevents an adequate sonographic study to establish the diagnosis (Figure 39-9).

Noninvasive techniques, such as TVS, used in conjunction with rapid β-hCG testing create fast and accurate diagnoses without the use of the painful and often inconclusive procedure. Vermesh et al¹⁸¹ reevaluated the role of culdocentesis in the management of ectopic pregnancy and reconfirmed previous reports establishing the lack of correlation between actual tubal status and culdocentesis result. Patients in whom the ectopic gestation could not be ruled out by means of TVS and pregnancy testing are probably best managed by laparoscopy.

The availability of TVS and a readily available machine with vaginal probe, especially in emergency rooms, has decreased and may perhaps eliminate the need for diagnostic culdocentesis.

Coelocentesis

Amniocentesis and chorionic villus sampling (CVS) have been used in prenatal diagnosis for chromosomal aneuploidies and the majority of single gene defects. Although CVS offers earlier diagnosis than amniocentesis, the procedure is generally not performed before 10 weeks of gestation because of the concern of fetal limb defects. During the first trimester of pregnancy, the amniotic cavity is surrounded by the extraembryonic coelom, which reaches its maximum volume at 7 to 9 weeks and then subsequently disappears at around 13 weeks of gestation. Transvaginal aspiration of coelomic fluid under ultrasound guidance may be performed as early as 6 weeks of gestation, although the groups performing this technique at present find that the ideal time is from 7 to 10 weeks of gestation.

The puncture of the extraembryonic coelom (Figure 39-10) by a thin needle transabdominally or transvaginally under TVS guidance is still experimental at present. The technique was first described by Jurkovic et al in 1993,¹⁸² and they were able to diagnose fetal sex by fluorescent in situ hybridization and polymerase chain reaction (PCR) on coelomic fluid sampled at 8 to 10 weeks of gestation. The limitation of the technique was that cytogenetic analysis always failed due to unsuccessful culture. Cruger et al improved the method in 1996,¹⁸³ and showed that at 6 to 10 weeks of gestation they were able to establish coelomic cells in culture and perform cytogenetic analysis.

It appears to be technically feasible, as reported by one group who performed 58 such punctures before elective termination in singleton pregnancies, at 7 to 10 weeks of gestation.¹⁸⁴ In all cases, the β-globulin DNA was successfully amplified, and 3 of 5 fetuses, where the maternal hemoglobin phenotype was HbAS, were found to be heterozygous for the sickle cell gene (HbAS). Normal β-globulin genotype was detected in the remaining 55. These results establish the feasibility of analyzing early fetal coelomic fluid for the identification of sickle cell anemia and perhaps other single human gene disorders. A recent study has established that highly sensitive fluorescent PCR can provide rapid, reliable, and accurate multiple genetic diagnoses (sexing and single gene diagnosis) from coelomic cells.¹⁸⁵ This technique is rapidly becoming the method of choice: it is highly reliable and highly accurate.

Other analyses of early coelomic fluid by transvaginally guided puncture have added to our knowledge base regarding the biochemical makeup of the early gestational unit. Campbell et al analyzed the levels of various intracellular substances including sodium, potassium, bicarbonate, chloride, urea, bilirubin, protein, albumin, and glucose in the coelomic and amniotic fluid compartments of early pregnancies.^{186,187,188,189, and 190} All differences in concentration were significant and showed that sodium, potassium, and bicarbonate were found in higher concentrations in the amniotic fluid, whereas chloride, urea, bilirubin, protein, albumin, creatinine, calcium, phosphate, and glucose were present in higher concentrations in the coelomic fluid. Amniotic and extraembryonic coelomic fluids have widely differing biochemical composition, although the significance of these differences is presently unexplained.¹⁸⁸

Recent advancements in the technique and assays have revolutionized our existing knowledge base and the possibilities for clinical importance of coelocentesis in the future. In a 2003 paper by Jauniaux et al,¹⁸⁵ quantitative fluorescent (QF) PCR amplification has been utilized for the rapid prenatal diagnosis of chromosome disorders on coelomic cells. QF PCR amplification using several markers for chromosomes X, Y, 21, 18, and 13 was successfully achieved on all 17 serial samples of exo-coelomic fluid (ECF), placental tissue, and maternal blood. Complex analyses of maternal blood samples and chorionic tissues allowed the distinction of fetal from maternal patterns and the identification of maternal contamination of the ECF samples. Prenatal detection of fetal gender was successful in all cases. When tested with autosomal primers, 7 samples were found to contain exclusively fetal DNA. Eight samples contained small amounts of maternal DNA that did not interfere with the PCR analysis. They conclude that QF PCR requires very small volumes of sample compared with cell culture, suggesting that coelocentesis may prove useful for very early prenatal diagnosis.

In Prenatal Diagnosis in 2005, Chatzimeletiou et al¹⁹¹ described their initial experience with aneuploidy screening in coelomic samples using fluorescent in situ hybridization (FISH). Coelocentesis was performed in 12 singleton pregnancies at 6 to 9 weeks of gestation immediately before surgical termination of pregnancy. Fluorescence probes for chromosomes 3, 7, 9, 13, 16, 17, 18, 21, 22, X, and Y were applied on uncultured coelomic fluid samples and placental tissue. They found that successful analysis by FISH was possible in all cases, and the results of the coelomic fluid were concordant with those from the analysis of placental tissue. They avoided problems with limited cell numbers by the application of a second layer of FISH, which has also enabled accurate identification of maternal cell contamination in male samples. Since these pregnancies were immediately terminated, assessment of risk of fetal loss could not be performed.

Santolaya-Forgas¹⁹² and his group tried to answer the question of loss after coelocentesis procedure by using a baboon model. Under aseptic conditions, 9 animals underwent coelocentesis under continuous transvaginal ultrasound guidance to avoid the amniotic or yolk sacs. They concluded that extracoelomic fluid could be easily aspirated using a 20-gauge needle. One pregnancy loss was detected within 3 days of the procedure. No complications occurred in the remaining 8 pregnancies. They believe that in the baboon model, the coelocentesis procedure is technically simple and presents a relatively low risk to mother and fetus if a 20-gauge needle is used and the amount of aspirated extracoelomic fluid is less than 3 cm³.

The risk for ongoing pregnancies after coelocentesis is unknown and must still be assessed. From a purely theoretical point of view, it would seem that the technique may be safer than the alternatives: CVS and amniocentesis. The amniotic membrane is not punctured in coelocentesis, resulting in a smaller risk than with amniocentesis, and no evidence of fetal limb anomalies have been observed, as has for CVS.

Ross et al¹⁹³ reported on the short-term safety of coelocentesis by examining the miscarriage rate within the week following the procedure. The miscarriage rate was 5% in the control group and 25% after coelocentesis. This group concluded that the short-term miscarriage rate is high after the procedure and increases with maternal age, is inversely related to gestational age, and is higher in women who experience bleeding in early pregnancy. The miscarriage rate after coelocentesis could not be explained by the proportion of blood-stained samples or the number of uterine entries, which are parameters known to affect the outcome of CVS and amniocentesis.

Transrectal sonography (TRS) can provide guidance for difficult D&C, intracavitary tandem placement, and cerclage placement. The biplane transrectal probe is recommended because one can image in the sagittal plane as well as axially, thereby confirming the position of a dilator or tandem in 2 planes (Figures 39-11 and 39-12). Transrectal sonography is particularly helpful in patients whose external cervical os cannot be adequately visualized due to extensive cervical cancer or vaginal stenosis, for example.

The biplane transrectal probe is covered with a condom, and fluid is introduced within the condom. The covered probe is introduced into the rectum after K-Y jelly is spread over the tip and anterior surface of the condom. This provides optimal contact with the rectal wall. The cervix is usually imaged initially in the longitudinal plane followed by selected images in the axial orientation.

If one is interested in localization of the external os, the endocervical canal can be identified on longitudinal views (Figure 39-13). Once the dilator is introduced into the cervix, its central location can be confirmed with the axial scan (Figure 39-14).

We have had experience with more than 25 cases where TRS was useful in guiding D&C, intrauterine tandem placement, or cerclage suture placement.¹⁹⁴ It can also be used for guiding biopsy or aspiration (Figures 39-15,39-16, and 39-17).¹⁹⁵

If it is used intraoperatively, the transducer is held by the operator and a sterile towel is placed over the operator's hand. The retractors are placed to the sidewalls of the vagina, allowing ultrasound to be transmitted through the rectum in a long-axis plane. The study should be videotaped for documentation purposes.

We have observed a few cases of uterine perforation (Figure 39-16). The dilator, tandem, or both could be identified as being extrauterine in both cases. For cerclage placement, the relative position of the sutures within the cervix can be identified.

In some cases, a transperineal approach is needed (Figure 39-17). This is particularly true in patients who have undergone pelvic exoneration. In some cases, the transvaginal probe with a needle guide can be used for guided aspiration.

In this chapter, both presently conventional and experimental transvaginally performed puncture procedures have been described and evaluated. The low procedure-related complication rates, the high degree of accuracy, and the high patient acceptance are advantages well worth considering; in some cases abdominal or vaginal surgery may be avoided or pregnancy outcome improved. This experience encourages consideration of this treatment modality in cases evaluated individually. Some of them may be performed in an office or emergency room, whereas others must be confined to hospital settings under carefully monitored protocols.

Our group has found the automated puncture device indispensable. The extreme accuracy of needle placement and lack of discomfort experienced by the patient have revolutionized our ability to better care for our patients who require these procedures.