Although surgical methods of abortion have changed little in the past 50 years, the expanding interest in pharmacologic agents has both increased reproductive options and broadened the range of possible complications. Even with optimum use of medical-based protocols, the need for surgery to complete a failed procedure or to manage complications requires a thorough knowledge of operative technique.

The Centers for Disease Control and Prevention (CDC) define induced abortion as a procedure to terminate a suspected or known intrauterine pregnancy and to produce a nonviable fetus at any gestational age (Koonin, 1999). In 2008, approximately 50 percent of more than 6.5 million pregnancies in the United States were unintended, defined as either mistimed or unwanted. Of these, 40 percent—excluding spontaneous losses—ended in legal abortion (Finer, 2014). The national rate of unintended pregnancies has fallen, which is perhaps due to more effective contraception. Still, approximately 20 percent of sexually active, reproductive-aged women in the United States who do not use effective contraception account for more than 40 percent of unintended pregnancies (Gold, 2009). Thus, most of these conceptions, and therefore induced abortions, are preventable.

Beginning in 1969, the number and selected characteristics of women who underwent legal abortions were monitored on an annual basis by the CDC. Since 1990, when the number of legal abortions in the United States peaked at 1.43 million, the absolute number has declined. In 2002, the total was 854,122, and by 2011, this had fallen to 730,322 (Pazol, 2014). At the same time, estimates world-wide exceed 40 million (Sedgh, 2007). Data are also collected regarding patient demographics and include age, race, marital status, state of residency, and gestational age. In addition, the recorded method—curettage, medical, instillation, and other—is gathered.

Despite this, many of the associated demographic and abortion-method data are not reported. The CDC depends on cooperation between abortion providers and the state agencies to supply statistics. Thus, their summary data are likely underestimated, although data trends should be reliable. Indeed, the CDC reports that for 2011, their totals are approximately 70 percent of estimates recorded by the Guttmacher Institute. This privately funded agency also collects and publishes analyses of abortion data obtained through periodic direct contact with identified abortion providers (Jones, 2014). This implies that in 2011, approximately 1.04 million legal abortions were performed in the United States.

Notable trends include declines in proportions of adolescent pregnancies and increased proportions of parous women and women of color selecting abortion (Jones, 2009). Poor women were identified as having the greatest absolute increase in the abortion rate between 2000 and 2008. Overall, an estimated 30 percent of women aged 15 to 44 years in 2008 will have had an abortion by age 45 (Jones, 2011a).

The CDC also computes the abortion rate, which is the number of abortions per 1000 women of reproductive age. In contrast, the abortion ratio is the number of abortions per 1000 live births. The abortion rate is a crude estimate of the population incidence of undesired pregnancies, whereas the ratio better approximates the proportion of conceptions that are undesired. Since 2002, in the United States, both the abortion rate and ratio have been declining. In 2011, the abortion ratio was approximately 219 per 1000 live births, whereas the rate was approximately 13.9 per 1000 reproductive-aged women (Pazol, 2014).

The observed declines in the reported abortion number, rate, and ratio is not easily explained and likely involves multiple contributing, interrelated factors (Jones, 2009). Possibilities include more effective contraception use, changing attitudes toward abortion, decreasing availability of providers, and increasing legal and economic barriers. Still unclear is whether legal abortion should be viewed as a reportable event such as births and fetal deaths or as a health-related issue akin to other procedures.

In spite of these and other limitations, accurate abortion statistics help estimate abortion practices. Groups with higher abortion rates can be targeted for improved health-care services and effective contraception for the prevention of unintended pregnancies. Preventable causes of abortion-associated morbidity and mortality may also be identified and addressed.

The availability of safe, legal procedures is the most important preventable factor affecting abortion-related deaths (Cates, 1976; Grimes, 1979). Maternal death from legal abortion is rare, and rates since 1978 are less than 1 death per 100,000 reported legal abortions (Pazol, 2014). In 2010, the last year for which CDC data are currently available, 10 deaths related to legal abortion were recorded. Other reports confirm that the incidence of maternal death and other complications increases with advancing gestational age (Diedrich, 2009). From data of the Abortion Mortality Surveillance System, the maternal mortality rate between 1988 and 1997 for procedures performed at ≤8 weeks’ gestation was 0.1 deaths per 100,000 abortions. In the second trimester, the relative risk for mortality increased substantially with a 14-fold increase at 13 to 15 weeks, a 29-fold increase at 16 to 20 weeks, and a 77-fold increased risk at ≥21 weeks’ gestation (Bartlett, 2004). Causes of abortion-related death can be classified as direct, which include hemorrhage, infection, embolic (thrombotic, amnionic fluid, or air), and anesthetic-related. Indirect deaths stem mainly from cardiac and cerebral vascular events. Of deaths, 80 percent were due to direct causes, of which hemorrhage and infection accounted for approximately half of cases. Anesthetic and embolic complications each accounted for approximately 15 percent.

Approximately 70 percent of legal abortions are performed by first-trimester suction dilatation and curettage (D & C). The epidemiology of abortion morbidity and mortality principally reflects complications related to this procedure (Pazol, 2014). A continuing pregnancy has an associated death rate of 8 to 9 per 100,000 live births. A legal abortion is markedly safer than this, and a recent estimate was only 0.6 deaths per 100,000 abortions (Raymond, 2012). As gestational age advances and abortion methods change to dilatation and evacuation (D & E) or medical induction, the observed death rate increases and approaches baseline pregnancy-associated mortality rates. Second-trimester procedures generally require greater levels of anesthesia, which may explain a portion of the observed increase.

Although induced abortion is usually the result of an unintended pregnancy, numerous social, economic, and personal pressures influence the decision (Torres, 1988). Other concerns include the effects of pregnancy on a woman’s health or issues related to fetal well-being. Some women with chronic medical diseases choose pregnancy termination only after learning that continuing the pregnancy poses significant health risks. That said, the effects of pregnancy on preexisting medical disease are highly variable, and in most conditions, they form a continuum of risk based in large part on the disease type and current severity. The point at which the risks to a patient’s health (or life) become unacceptable is often obscure, and risks can only be estimated. Moreover, accepted guidelines for recommending pregnancy termination do not currently exist. Nevertheless, for most chronic medical conditions, there are sufficient data to inform clinical decisions based on the likelihood and magnitude of pregnancy-associated risks. The importance of accurate preconceptional counseling cannot be overemphasized.

To assess fetal well-being, prenatal diagnosis is increasingly elected, and significant structural and genetic conditions affect at least 3 to 5 percent of all births. As a result, an increasing number of women are faced with the decision of abortion based on neonatal health concerns. In some cases a presumed “lethal” condition is found, although the use of this term has been appropriately disparaged. Instead, a physiologic appraisal and counseling are preferred (Wilkinson, 2012). For fetal conditions not associated with neonatal survival, induced abortion merely changes the outcome timing. However, nonlethal conditions are more common and are generally associated with wide variations in clinical outcome. Rarely can diagnosis accurately depict the postnatal course. Thus, patients must usually make decisions based on incomplete information. Moreover, prenatal diagnosis is often not completed until the second trimester, a time when pregnancy terminations incur more risk and expense and become less available.

Provider Numbers

Based on a 1996 survey of abortion providers, the Guttmacher Institute reported that 32 percent of reproductive-aged women in the United States lived in the 86 percent of counties that had no recognized provider (Henshaw, 1998). In total, 70 percent of abortions were performed in outpatient specialty clinics compared with only 7 percent in hospitals. These findings confirm that, since abortion legalization in the United States in 1973, hospital-based services have mostly been replaced by outpatient centers. These centers can provide a broader and more cost-effective range of reproductive health services than can either hospitals or physician offices. Regrettably, this evolution has fostered the marginalization of abortion providers and decreased accessibility to hospital-based residency training. Because abortion is an emotionally charged issue, it has attained significant political and legal status, which threatens to transcend its medical aspects.

Between 1982 and 2000, the number of abortion providers in the United States had steadily declined. However, provider numbers through 2011 have become relatively stable, which may be attributable to an increase in clinicians who perform medical instead of surgical procedures. In 2008, approximately 1800 providers were reported by the Guttmacher Institute (Jones, 2011b). Reasons for the decline include perceived social stigma, economic reimbursement, decreased training opportunities, and even personal fears (American College of Obstetricians and Gynecologists, 2014c). These pressures have created a paradox in which most obstetrician-gynecologists favor access to legal abortion, but most are unwilling to perform them.

Attempts at reversing the trends in provider attrition have met obstacles. The American College of Obstetricians and Gynecologists (2014d) has long been a supporter of legal abortion prior to fetal viability, and the Accreditation Council for Graduate Medical Education has mandated since 1996 that residency curricula include induced abortion. That said, specific training in surgical methods, especially second-trimester D & E, may be lacking (American College of Obstetricians and Gynecologists, 2014a; Eastwood, 2006). Clearly abortion training in residency improves many skills needed for practice, including first-trimester sonography, management of incomplete and missed abortions or fetal death, and provision of effective analgesia.

Limitations to Access

In 1990, the National Abortion Federation and the American College of Obstetricians and Gynecologists sponsored a symposium on abortion provision (Grimes, 1992). This group identified many disincentives to abortion access, most of which had no simple solutions. Increasingly, federal and state legislatures have sought to limit legal abortion on many levels. The most widely imposed restriction is economic. Most states have emulated the federal restriction on abortion reimbursement by prohibiting public funding. Women with federally funded medical care are usually not covered for abortion services unless they are performed for serious maternal health concerns. Other common restrictions involve parental involvement in the abortion decision, mandatory counseling with state-imposed content, waiting periods, and the requirement to be given the opportunity to view sonographic images.

Much of the current state legislative activity seeks to emphasize the potential restrictions permitted by Roe v. Wade for postviable pregnancies, originally described as third-trimester cases. Presumably, the 1973 court recognized that third-trimester abortions might reasonably be restricted because of legitimate concerns for neonatal survival, but acknowledged that a procedure performed for the life or health of the mother could not be banned.

Other legislation has focused on procedural restrictions such as those associated with partial birth abortions, also known as intact dilatation and extraction (D & X) (Epner, 1998). Partial birth abortion is a lay term that is not recognized by the College or other medical authorities. According to a policy statement by the American College of Obstetricians and Gynecologists (1997), intact D & X contains these four elements: (1) deliberate cervical dilatation, (2) conversion of the fetus to a footling breech, (3) breech extraction of the body (but not the cranium), and (4) evacuation of the intracranial contents to aid the delivery of a dead but otherwise intact fetus.

However, because all of the included steps are part of established obstetric techniques, they must be performed in the precise sequence to meet criteria for an intact D & X. The criminalization of poorly defined abortion methods has harmful ramifications for abortion providers and carries penalties that far exceed commonly encountered civil liabilities (Kassirer, 1997). The American College of Obstetricians and Gynecologists (1997) has acknowledged that intact D & X may be the best or most appropriate procedure in a particular circumstance and that the decision should be left to the patient and her doctor.

In 2003, a law banning partial birth abortion was signed and subsequently upheld by the Supreme Court in 2007 (Gonzales v. Carhart). In response, the American College of Obstetricians and Gynecologists condemned the ruling. States responded by passing laws restricting this procedure, and as of this writing, 32 have passed laws banning it. Although some have been enjoined by state courts, others permit an exception for the mother’s life. Because of the indefinite nature of these legal proceedings, periodic surveillance of local policies is essential.

Thus, much of the contemporary debate involves the definition of “viability,” an imprecise term that lacks evidence-based definitions. More recent court rulings have wrestled with legal definitions of viability. Although imperfect, gestational age and fetal weight estimation are still the most accurate predictors of neonatal survival. Because the capacity for neonatal survival spans a biologic continuum of probability across gestation and is greatly influenced by the presence and severity of fetal anomalies and the level of neonatal care provided, it is unlikely that a single gestational age threshold could ever be universally accepted. Nevertheless, recent expert opinion now suggests that viability may be reasonably assumed as early as 23 weeks’ gestation (Raju, 2014).

Patient evaluation should include a history, physical examination, appropriate laboratory studies, and counseling. Beyond establishing that the decision is voluntary, counseling must include informed consent. At a minimum, the patient should know the diagnosis, purpose of the procedure, risks and possible complications, alternative treatments, and the likelihood of successful treatment (American College of Obstetricians and Gynecologists, 2012).

Counseling requires a nonjudgmental, informed, and objective attitude. Using open-ended questions, clinicians encourage the woman to express her own feelings and the feelings of those close to her regarding her decision, in the context of her current life situation and her plans for future childbearing and contraception. Providing accurate information can alleviate common fears regarding abortion and later reproductive health. In some cases, women experience guilt from various underlying beliefs and perceptions. These feelings should be normalized, even if they cannot be completely dispelled (Adler, 1992). Legal restrictions, particularly those imposed on adolescents, may further heighten the psychologic discomfort, and counselors may effectively mediate the process of parental notification, when applicable. Pain management and postprocedure contraception is also discussed.

History and Physical Examination

The complete history focuses on the timing and reliability of the last normal menstrual period, because accurate gestational age determination is crucial to selecting the correct method. Other important historical elements for preoperative care are listed in Chapter 18 (p. 291). Women with conditions severe enough to warrant a medically indicated pregnancy termination may require additional evaluation to optimize their status. Although stable and well-controlled chronic medical conditions can be managed in outpatient centers, indications for inpatient procedures are numerous and require appreciable judgment (Guiahi, 2012). Women with known bleeding disorders or who are receiving anticoagulation represent a particularly high-risk group. In these women, termination using an abortifacient is avoided because the timing of the passage of conception products and the amount of associated hemorrhage are unpredictable.

Physical examination ideally obtains vital signs and weight and assesses cardiopulmonary findings. A sterile speculum is inserted to seek cervical pathology that might increase surgical complications. These include active cervicitis, deformities, and past trauma. Bimanual pelvic examination should ascertain uterine size, detect other pelvic pathology, and determine the position of the uterus with respect to the cervix. The latter helps direct the insertion angle of cervical dilators and curettes. Women with suspected pelvic pathology or a discrepancy between the menstrual date and uterine size should undergo sonography. Imaging can be used to exclude multifetal gestations, uterine anomalies, or ectopic pregnancies. Maternal obesity or other obstacles to a complete pelvic examination should also prompt sonographic imaging. If an embryonic or fetal demise is found, this may remove sources of guilt and anxiety and may enhance third-party reimbursement. Although routine use of sonography has been questioned, it is widely performed (Kulier, 2011a; O’Connell, 2009). In women with prior cesarean deliveries, sonography localizes the placental implantation site to exclude abnormal placentation, which increases bleeding risks. Some cases, such as cesarean scar ectopic pregnancy, require special care at facilities equipped to manage their potentially significant complications (Chap. 8, p. 128).

Laboratory Testing

Few laboratory studies beyond a reliable urine or serum pregnancy test are required in healthy, young women. At minimum, hematocrit, Rh status, indirect Coombs, and urinalysis are obtained. Cultures of the cervix are appropriate if active cervicitis is suspected, and routine testing of women younger than 25 years has been recommended by the Centers for Disease Control and Prevention (2015). Rh immunoglobulin is indicated for an Rh-negative woman, unless she is already isoimmunized or the paternal Rh-negative status is confirmed. Testing for human immunodeficiency virus is also considered (Gupta, 1998).

Methods of first- and second-trimester abortion are either primary surgical or medical. Surgical methods generally require mechanical cervical dilatation sufficient to pass the necessary instruments. Pharmacologic cervical “ripening,” that is, softening, may serve as a useful adjunct.

Primary medical regimens combine the effects of pharmacologic or mechanical cervical ripening and the stimulation of uterine activity. Pharmacologic regimens are favored in the first trimester. In later gestations, a separate phase of cervical ripening may be desirable and can employ both mechanical and pharmacologic methods. However, two phases increase procedural time.

First-Trimester Dilatation and Curettage

In the early first-trimester, vacuum aspiration is by far the most common abortion method. For pregnancies ≤7 weeks’ gestation, small flexible cannulas (4 to 6 mm) may be selected. Often, cervical dilatation is not required at this early stage. This procedure has been termed menstrual extraction or menstrual regulation. It has been performed after a missed menses but without biochemical or sonographic confirmation of a pregnancy. Required instruments include a speculum, tenaculum, 50-mL manual vacuum syringe, and appropriate cannulas (Fig. 9-1). Although preoperative sedation is not usually necessary, preoperative administration of a nonsteroidal antiinflammatory drug (NSAID) may help reduce pain.

After the patient is placed in dorsal lithotomy position, a speculum is inserted. The cervix is visualized and cleaned with povidone-iodine or chlorhexidine gluconate solution or simply irrigated with saline (Achilles, 2011). In parous women, a small flexible catheter will usually pass without much discomfort. In nulliparas, a paracervical block, described below, may be needed to reduce pain. Injection of a local anesthetic at the site of tenaculum placement may also be considered.

After the catheter reaches the uterine fundus, the vacuum syringe is primed, and the catheter is slowly rotated and withdrawn. Confirmation of tissue consistent with products of conception is crucial to all surgical abortions. Failure to obtain such tissue should prompt transvaginal sonography and a measurement of a quantitative human chorionic gonadotropin (hCG) level to exclude a possible ectopic gestation (Edwards, 1997). The optimal suction cannula for early curettage is unclear. Flexible cannulas have two ports and smaller diameters, which may cause them to clog more easily. A 7-mm rigid cannula may permit intact extraction of the gestational sac, which aids postoperative confirmation.

For women who present early in pregnancy, this method appears to be safe and as effective as procedures performed later in the first trimester. The lower gestational age threshold for this method has not been established. If attention to detail and appropriate follow-up is implemented, there may not be an absolute lower limit.

For gestations aged ≥8 weeks, curettage usually requires some cervical dilatation, especially in nulliparas. This is performed intraoperatively with tapered dilators such as Hank dilators or Denniston (Pratt design) autoclavable dilators (Fig. 9-2). Preoperative osmotic dilators or, alternatively, prostaglandins administered via oral, buccal, or vaginal routes are also options and are discussed on page 141. Required cannulas have a larger diameter (Fig. 9-3).

Intravenous (IV) access is considered whenever a paracervical block is placed because of risks for anesthetic agent toxicity and for vagal reactions. Access may also be used to provide light conscious sedation. Common agents include midazolam (Versed) and fentanyl (Sublimaze). Nitrous oxide in concentrations of up to 50 percent is another option. Conscious sedation is defined as a minimally depressed level of consciousness that still permits a woman to maintain her airway and to appropriately respond to both physical and verbal stimulation. That said, agents and dosages that provide light conscious sedation in one patient may lead to much deeper levels, even to respiratory compromise, in others.

In practice, conscious sedation forms a continuum. Three levels are recognized—minimal, moderate, and deep sedation—categorized by responsiveness, spontaneous ventilation, cardiovascular function, and airway control. Patients with underlying medical problems are more likely to experience unanticipated effects. Accurate and continuous monitoring of the patient’s level of consciousness is the single most important aspect of safe conscious sedation. Moderate or deep sedation is ideally avoided and may unexpectedly precipitate the need for intubation. Thus, incremental dosing is preferred to a single-dose-fits-all approach. For these reasons, facilities providing conscious sedation should have equipment, medications, and personnel on hand to assist the primary operator if needed. Guidelines for the administration of conscious sedation are available from the American Society of Anesthesiologists (2002).

During the procedure, initial steps mirror those described for earlier gestations. Previously placed osmotic dilators can be removed at this time. After the paracervical block has set, a single-tooth tenaculum is applied for countertraction. This also straightens the endocervical canal and helps avert uterine perforation. The tenaculum may be oriented horizontally and incorporate tissue on the anterior ectocervix. Some prefer a vertical orientation, with the posterior tooth inside the external os. The tenaculum may also be placed on the posterior cervical lip in analogous fashion, and in cases of significant retroflexion, it provides superior traction forces for straightening the canal. In these cases, the curved dilators and curettes are also inserted with a posterior trajectory to match canal and cavity topology.

With paracervical blockade, pain transmitted through sensory and parasympathetic fibers in Frankenhäuser plexus is blocked. These fibers innervate the cervix and upper vagina and enter the uterus at the level of the internal os. With countertraction applied by the tenaculum, an operator injects 4 to 5 mL of anesthetic to a depth of about 1.5 cm into the cervical stroma and at the level of the cervicovaginal junction. Injections are at 4 and 8 o’clock positions and near the insertion sites of the uterosacral ligaments (Fig. 9-4). As another method, injection to a depth of about 3 cm directly into the cervix may be more effective. This is completed at the 3, 5, 7, and 9 o’clock positions to deliver a total of 16 to 20 mL. However, significantly greater resistance is expected at the greater depth (Cetin, 1997; Wiebe, 1992).

Intravascular injection is avoided by careful needle aspiration. Either 1-percent lidocaine or 1- to 2-percent chloroprocaine may be selected. Chloroprocaine, an ester, has a faster onset of action and is safer after inadvertent intravascular injection. Some practitioners add 3 to 4 mL of 8.4-percent sodium bicarbonate to 30 mL of lidocaine to reduce acidity and injection discomfort. This may also hasten onset of the block. Addition of vasopressin (Pitressin)—10 units to 30 to 50 mL of the local anesthetic—decreases blood loss and eases cervical dilatation (Phillips, 1997; Schulz, 1985). A standard vial of Pitressin contains 20 units.

Progressive dilatation of the cervix is performed with a set of tapered, curved dilators with the operator’s hand stabilized on the perineum (Fig. 9-5). Dilators with a gentle taper (e.g., Hank or Pratt types) may be easier to use than Hegar dilators. Hank dilators also have a raised band on the shank, which provides a visual clue to prevent overinsertion. Dilators are sized in consecutive numbers, which represent either their circumference in mm (Hank or Pratt types) or their diameter in mm (Denniston or Hegar types). In general, the cervix is dilated to an opening size slightly larger than the chosen cannula, because as the abortion proceeds, physiologic narrowing of the canal ensues. An important aspect of curettage is the tactile assessment of insertion force and the location of the uterine fundus, which marks the anatomic boundary for instrument passage. Constriction at the internal os interferes with this feedback. Thus, a second dilatation and larger dilators are used as needed, or the curettage may be finished with a smaller cannula.

Creation of a false tract such as that shown in Figure 9-6 is a complication of dilatation, especially if small-gauge dilators are initially required. It may result in uterine perforation. At least in a training institution, the adjunctive use of intraoperative sonography is instructive and potentially protective. It provides important visual feedback as to the optimal path for the dilators and curette. Sonography also reinforces desirable tactile feedback to guide insertion forces and to halt insertion past the uterine fundus. Sonography also helps to identify the location of fetal parts during D & E as later described. It can confirm an empty cavity or identify the reaccumulation of blood. In sum, lower rates of uterine perforation and fewer repeat procedures for retained products have been reported with intraoperative sonography use (Acharya, 2004; Darney, 1989).

If a rigid cannula is chosen, its diameter in millimeters should be equal to (or 1 mm less) than the gestational age in weeks. Both curved and straight rigid cannulas are available, and the choice is primarily based on operator preference and canal shape (see Fig. 9-3). If flexible cannulas are preferred, a 5-mm Karman cannula may be used up to 7 weeks’ gestation; a 6-mm up to 10 weeks’ gestation; a 7-mm at 11 weeks’ gestation; and an 8-mm at 12 weeks’ gestation.

After adequate dilatation is confirmed, the cannula is inserted into the uterine cavity as shown in Figure 9-7. Next, 50 to 60 mm Hg suction is applied. After the appropriate vacuum level is attained, the abortion proceeds by slowly rotating and withdrawing the cannula (Fig. 9-8). Because of the risk of perforation with rigid cannulas, these are not inserted with suction applied. With flexible cannulas, a slow, gentle in-and-out motion may be applied, taking care to rotate the cannula only on the outward movement lest the tip be avulsed. Excessive bleeding after catheter withdrawal is treated with a second aspiration. At times, the catheter may become clogged, and this may require catheter removal. Alternatively, rapid application and removal of the suction by the sliding suction control on the handle, while keeping the cannula stationary, can also be effective. After suctioning is finished, some recommend routine sharp curettage to confirm complete tissue removal. This, however, is probably unnecessary and may increase bleeding (Darney, 1987b). Once only scant bleeding is noted, instruments are withdrawn. A bimanual examination is repeated to document a firm, involuting uterus.

Tissue that has accumulated in the gauze-lined suction canister can be resuspended in saline and examined (Fig. 9-9). Continuing pregnancy is reported in 0.1 to 0.3 percent of cases after first-trimester D & C (Binkin, 1984). The risk increases with declining gestational age. In some cases, a multifetal gestation may be the underlying cause. At gestational ages less than 9 weeks, portions of the gestational sac and trophoblastic tissue are identified. Beyond this time, fetal parts can also be identified. If the expected tissue is not recognized, then ectopic pregnancy, molar gestation, or failed abortion is suspected, and the patient is evaluated accordingly (Edwards, 1997). Use of prophylactic antibiotics is indicated for all suction abortions, as described on page 149 (Low, 2012; Sawaya, 1996).

Second-Trimester Dilatation and Evacuation

In the midtrimester, D & E is by far the most commonly reported abortion method (American College of Obstetricians and Gynecologists, 2015). But, the newer medical induction regimens have not been adequately compared with D & E in contemporary randomized trials. Thus, the choice of initial method, especially beyond 16 weeks’ gestation, is best left to operator or institutional preference. It is well established that D & E, especially when performed in the latter part of the second trimester, requires special training, skills, and instruments.

Compared with curettage, D & E is technically more complex for several reasons. These include the need for more advanced cervical dilatation, the use of special instruments to remove the fetal parts, and the requirement for additional analgesia. A paracervical block alone is unlikely to provide sufficient pain relief. Thus, IV access followed by conscious sedation or regional analgesia is strongly considered. As mentioned in the previous section, vasopressin, as a component of paracervical blockade, reduces intraoperative blood loss (Peterson, 1983).

In most cases, and in essentially all cases beyond 16 weeks’ gestation, preoperative osmotic dilators or other forms of cervical ripening are necessary. Without preliminary cervical preparation, sufficient dilatation to a diameter ≥2 cm is unlikely without using excessive force with the dilator. Adequate dilatation is required to pass the grasping forceps through the internal os and to remove all the fetal parts. Prior to starting the evacuation, sufficient cervical dilatation must be confirmed. If cervical dilatation is inadequate, additional osmotic dilators can be placed and the procedure delayed for at least 4 to 6 hours or rescheduled for the next day (Hern, 1994). Care is taken to leave the membranes intact during osmotic dilation. If the chorioamnion is inadvertently ruptured, the osmotic dilators are still placed but removed within 12 hours. The procedure is then completed.

Once dilatation is deemed adequate, the cervix is grasped with a tenaculum, and a cannula is passed into the uterine cavity to rupture the membranes and evacuate as much amnionic fluid as possible by applied suction. The cannula is always advanced fully into the uterine cavity before suction is applied. Up to 15 weeks’ gestation, a 14-mm cannula will usually suffice, whereas gestations ≥16 weeks require a 16-mm size. During suctioning of the amnionic fluid, some membranes and placental tissue are often removed. However, complete placental removal is generally delayed until after the fetal parts are taken.

After the amnionic fluid and cannula have been removed, the operator extracts the fetal parts with forceps (Fig. 9-10). Once inserted, the forceps should be opened widely and one blade inserted along the anterior uterine wall to grasp the fetal parts. When fetal parts are felt between the blades on closing the handles, these are removed with a firm grasp and a gentle twisting motion. The calvarium is the most likely part to be left behind. If it cannot be grasped with forceps, some clinicians administer oxytocin, which may induce its descent toward the internal os and thereby aid its retrieval.

Complete placental removal generally requires one or more passes with the 14- to 16-mm suction curette. Before concluding the procedure, the fetal parts are assembled to confirm a complete abortion. Similarly, the quantity of placental tissue is visually assessed. If fetal parts remain in the uterus, gentle probing with a small, sharp curette may locate and retrieve them. Conversely, they made be identified with sonography and then removed.

Hysterotomy and Hysterectomy

Except in unusual circumstances, hysterotomy is infrequently used for abortion. Prior placement of an abdominal cerclage with plans for continued cerclage retention or a large obstructing myoma may be legitimate indications for hysterotomy. In some cases, a spontaneous hysterotomy may occur with uterine rupture during midtrimester labor induction (p. 150). Possible indications for concurrent abortion/hysterectomy include some cases of cervical cancer, uncontrollable hemorrhage from abnormal placentation, and septic abortion.

Many first-trimester procedures are performed following preoperative cervical preparation. One Cochrane review of 51 trials of first-trimester surgical abortion included comparisons of 24 different cervical preparations (Kapp, 2010). These investigators concluded that a separate cervical ripening phase decreased the procedure length and the need for mechanical dilation. Agents included mifepristone, osmotic dilators, and any of four different prostaglandins. More recent placebo-controlled trials have confirmed the efficacy of preoperative ripening with misoprostol in the first trimester, albeit at the risk of more maternal symptoms. Of these, Meirik and colleagues (2012) showed that a single 400-μg oral dose of misoprostol was associated with lower rates of incomplete abortion and the need for repeat curettage. Using a similar dose administered vaginally, Mittal and associates (2011) observed decreased procedural time, blood loss, and need for mechanical dilation. The World Health Organization Technical and Policy Guidance on Safe Abortion and the Royal College of Obstetricians and Gynecologists recommend cervical preparation for procedures beyond 9 weeks’ gestation for nulliparas and for all women beyond 10 weeks (Lalitkumar, 2007). In the second trimester, cervical preparation is essential to prevent cervical and uterine injury. Various mechanical and pharmacologic methods are listed in Table 9-1.

Osmotic Dilators

Laminaria Tents

Laminaria species, specifically L japonica and L digitate, were one of the earliest cervical ripening devices. Made from desiccated seaweed stalks, laminaria induce cervical ripening primarily through mechanical distention of the endocervical canal. Secondary mechanisms are the release of endogenous prostaglandins and possibly disruption of weak collagen cross-linkages (Ye, 1982). After preparation of the cervix with an antiseptic solution, one or more laminaria tents are coated with sterile lubricating jelly. They are inserted into the cervical canal so that their proximal tip lies just past the internal os (Fig. 9-11). The tents expand most rapidly in the first 4 to 6 hours but reach maximum size at 24 hours. They attain a final diameter four times greater than their original (Wheeler, 1983). Placement of several smaller tents, rather than one large tent, may result in improved dilatation. This may also ease their later removal, as a single, large tent can form a dumbbell shape in the canal and lead to a more difficult and painful tent extraction. The clinician ideally places as many tents as will fit without causing patient discomfort. Gentle placement of several gauze sponges at the external os helps to prevent spontaneous expulsion. If placed too far into the endocervical canal, laminaria can become dislodged superiorly and become intrauterine. For this reason, the numbers of sponges and laminaria inserted are carefully tabulated and recorded in the chart. Laminaria, once placed, do not preclude the patient from ambulating, voiding, or stooling.

For midtrimester medical induction, laminaria are effective and significantly reduce the induction-to-delivery interval (Atlas, 1998; Stubblefield, 1975). In general, the value of laminaria increases with longer insertion-to-induction intervals, and the laminaria are ideally inserted the day prior to the procedure. Serial laminaria applications significantly increase cervical dilatation compared with a single application (Hern, 1994; Stubblefield, 1982). Indeed, repetitive laminaria application over several days can produce cervical dilatation of approximately 4 cm. This may create enough dilatation to preclude the need for additional forced dilatation in the operating room during D & E or further hasten labor in cases where labor induction is selected. For the latter, use of a vaginal prostaglandin and a single laminaria application prior to or concurrent with the uterotonic agent is generally sufficient. In sum, the need for shorter inductions must be weighed against the delay and patient inconvenience associated with 2 to 3 days of cervical preparation.