Cervical dilation before first-trimester surgical abortion (<14 weeks' gestation)

When cervical dilation is needed, most North American providers employ rigid dilation alone with steel Pratt dilators (Fig. 1a ) or the plastic equivalent, Denniston dilators []. The Pratt dilator is characterized by a gradual taper at the end of the instrument and comes in sizes ranging from 9 to 79 F. For most pregnant women, dilation can be initiated easily with a 17-F dilator. Each French unit refers to the circumference of the dilator in millimeters. To obtain the diameter of the dilator in millimeters, the French unit is divided by Pi (approximately 3). By comparison, Hegar dilators (Fig. 1b) have a blunt end and come in sizes ranging from 1 to 26 mm in diameter. Hegar dilators increase in size more rapidly than tapered dilators, potentially reducing the time necessary for dilation but also requiring more force [, ]. It is easier to sense the loss of resistance of the internal os with Hegar dilators than with Pratt dilators []. No trials have compared the safety and efficacy of Pratt and Hegar dilators. A minority of providers report using the Hegar dilator [].

Cervical priming can be accomplished mechanically with osmotic dilators that absorb moisture from the cervix and slowly expand to dilate the cervical os or biochemically with prostaglandin analogues or progesterone antagonists. The commercially available options for cervical priming include osmotic dilators, laminaria and Dilapan-S, as well as pharmacological agents such as misoprostol (PGE1), gemeprost (PGE1), dinoprostone (PGE2) and mifepristone. Other agents for cervical priming include nitric oxide donors (e.g., isosorbide mononitrate, glyceryl trinitrate?), which are currently under study. As of 2002, 18% of North American providers routinely used misoprostol for cervical ripening prior to 11 weeks' gestation, and 16% used it for multiparous women. Beginning at 11 to 12 weeks of gestation, 25% used misoprostol for nulliparous women and 20% for multiparous women [].

Two types of osmotic dilators are available in the United States: laminaria tents made of seaweed and synthetic dilators. All osmotic dilators require a trained provider and a speculum examination for insertion.

Laminaria tents are dried compressed stalks of hygroscopic seaweed (Laminaria digitata, Laminaria japonicum) that absorb water from the cervical stroma, swelling to three or four times their dry diameter overnight. Laminaria tents are available in multiple dry diameters. Laminaria apply radial force to the walls of the cervical canal and also induce the local production of prostaglandins to promote dilation [, ].

Laminaria are safe; adverse events following their use occur infrequently. Since bacterial spores can remain despite treatment of the laminaria with ethylene oxide or irradiation for sterilization, clinicians have worried about the risk of infection with laminaria. Case reports have described bacteremia following laminaria placement in the second trimester [, ]. Researchers theorize that laminaria insertion can facilitate the transfer of cervical or vaginal flora into the uterine cavity and cause an ascending infection []. However, a randomized trial comparing laminaria and rigid dilation in first-trimester abortion found no significant difference in rates of postabortion infection []. In the first trimester, all osmotic devices are generally removed no later than 24 h after insertion to reduce the possibility of infection. Because laminarias derive from a natural material, their use rarely results in hypersensitivity reactions (urticaria, angioedema, respiratory distress) and anaphylaxis in women with previous exposure [, , ]. The true incidence of this reaction is unknown. Investigators believe that the mechanism for this allergic reaction is IgE-mediated [].

Use of laminaria before first-trimester surgical abortion may pose an access barrier: laminaria acts slowly and often require an extra visit. To address the limitations of laminaria, investigators created two synthetic dilators, Dilapan-S (GelMed International, Czech Republic) and Lamicel® (Medtronic, Mystic, CT, USA), which are sterile and swell more rapidly than laminaria. An additional advantage of synthetic dilators is the consistency of length and shape, which leads to more predictable results.

Dilapan-S is a rod-shaped hydrophilic dilator made from polyacrylate-based hydrogel (Hypan) available in the following dimensions: 3 mm×55 mm, 4 mm×55 mm and 4 mm×65 mm. Dilapan-S absorbs moisture from the cervical tissue; in 4 h, the 3-mm diameter swells to an average of 8 to 9 mm, and the 4-mm diameter swells to an average of 10 to 11 mm, according to the manufacturer. In the United States, the Dilapan-S product label recommends one device placed up to 4 h prior to suction abortion in gestations of less than 16 weeks. In other countries, Dilapan-S is not subject to limitations on number of devices, indication or duration of use. Therefore, US providers commonly use Dilapan-S in an off-label fashion. Problems from breakage due to entrapment with the original version of Dilapan were resolved with release of a new formulation in 1998 internationally and in 2002 in the United States []. The reformulated Dilapan, named Dilapan-S, has not generated any reports of breakage []. Advantages of Dilapan-S are its significant radial force and rapid swelling, making it ideal for same-day procedures.

Lamicel, a polyvinyl alcohol polymer sponge impregnated with 450-mg magnesium sulfate and compressed to form a thin cylindrical tent, is no longer available in the United States. When placed in the cervical canal, Lamicel absorbs water from the cervical stroma and swells to four times its original size, transforming itself into a soft sponge [, ]. Lamicel's mechanism of action is biochemical rather than mechanical, exerting no outward pressure on the cervical canal []. Lamicel reduces the amount of force needed to dilate the cervix [, ] and is approved by the Food and Drug Administration for cervical priming for gestations of less than 16 weeks. Lamicel is significantly easier to remove than laminaria or Dilapan-S. Although Lamicel is the most expensive osmotic dilator, it often takes multiple laminaria to achieve the same priming effect that one Lamicel can produce in 4 h. Lamicel is 75-mm long and is available in either 3-mm or 5-mm dry diameters internationally.

Pharmacologic agents, such as prostaglandin analogues and progesterone antagonists, can be used for cervical priming in the first trimester. Misoprostol is the medication most commonly used for cervical priming []. Misoprostol is a PGE1 synthetic analogue marketed as an oral preparation to prevent and treat gastroduodenal damage induced by nonsteroidal antiinflammatory drugs (NSAIDs). E-series prostaglandins are preferred over F-series prostaglandins because they stimulate uterine smooth muscle more than intestinal or vascular smooth muscle and do not cause bronchoconstriction []. Misoprostol's advantages are its low cost, long shelf life and lack of need for refrigeration. Other prostaglandins, such as gemeprost and dinoprostone, are not used for cervical priming before surgical dilation in North America because they are more expensive but no better than misoprostol at cervical priming [, , , ]; in addition, they require refrigeration for transport and storage [].

Investigators have examined different routes of misoprostol administration. The ideal route must take into account not only efficacy but also patient and staff acceptance and convenience. Misoprostol can be administered orally, vaginally, sublingually, buccally or rectally. Pharmacokinetics studies comparing oral and vaginal administration (Fig. 2) have shown that vaginal misoprostol is associated with slower absorption, lower peak plasma levels and slower clearance, similar to an oral extended-release preparation [, , ]. Vaginal misoprostol is also associated with a greater overall exposure to the drug, area under the curve (AUC) and greater effects on the cervix and uterus []. There is no clinically significant difference between vaginal misoprostol that is administered dry and vaginal misoprostol moistened with water, saline or acetic acid [, , , ]. The rectal route of administration shows a pattern similar to vaginal administration but a lower AUC, including a significantly lower maximum peak concentration []. The sublingual route of administration has an AUC similar to vaginal administration but more rapid absorption and higher peak levels than either vaginal or oral administration []. This translates into higher rates of gastrointestinal side effects. The sublingual route also causes uterine contractions at a rate equivalent to vaginal administration []. The buccal route of administration shows a lower AUC, a lower peak concentration, and fewer side effects than sublingual administration []. The buccal route has a pattern of absorption similar to the vaginal route but produces lower serum levels overall. Nevertheless, the buccal route of administration produces uterine tone and activity similar to that resulting from vaginal administration. The buccal route of administration is also felt to be the least variable in terms of drug exposure and peak levels []. The administration of NSAIDs for pain relief does not alter the efficacy of misoprostol for cervical priming [].

The adverse effects caused by misoprostol have prompted research into other agents that can dilate the cervix effectively without causing abdominal pain, nausea, vomiting and diarrhea. Mifepristone is a progesterone antagonist that withdraws hormonal support of the pregnancy [, , , ], causing cervical softening []. One randomized trial suggests that mifepristone is as effective at cervical priming as misoprostol and better tolerated []. Nevertheless, mifepristone, while effective at cervical priming, is not often utilized because of its high cost and limited availability in many clinical settings. In addition, mifepristone is typically administered 24 h prior to a procedure, where misoprostol is administered 2–3 h prior. Both misoprostol and mifepristone are believed to exert their effects on the cervix via nitric oxide [, ]. Because nitric oxide inhibits contractions in the uterus but still induces cervical softening, nitric oxide donors have been investigated as potential cervical priming agents []. While nitric oxide donors (e.g., isosorbide mononitrate) are more effective than placebo or no treatment in dilating the cervix, a systematic review found that prostaglandins were superior to nitric oxide donors []. Furthermore, nitric oxide donors were associated with their own side effect profile of headache, palpitations and dizziness.

Most providers in North America continue to use rigid dilators for cervical dilation without preoperative priming [, , ]. They do so because the risk of uterine perforation or cervical laceration with first-trimester suction aspiration is very small, and each method of cervical priming is associated with side effects and additional inconvenience for the patient. Previous international guidelines for cervical priming have been published, including recommendations from the World Health Organization (WHO) and the Royal College of Obstetricians and Gynaecologists (RCOG). WHO recommends cervical preparation prior to surgical abortion for all pregnancies of 12 to 14 weeks' gestation. The guidelines note that cervical preparation may be considered for women at any gestational age, in particular those at high risk for cervical injury or uterine perforation, such as adolescents and women with cervical anomalies or previous cervical surgery []. The RCOG guidelines state that cervical priming should be considered for all women in the first trimester, recommending that misoprostol 400 mcg be administered vaginally 3 h prior to surgery or sublingually 2 to 3 h prior to surgery. They note that cervical preparation is likely “particularly beneficial where risk factors for cervical injury or uterine perforation exist, such as [in] adolescents aged less than 17 years, advanced gestational age (particularly among parous women), cervical anomalies or previous surgery, or when a less experienced surgeon is operating” [].

There is no consensus among providers regarding the desired width of dilation. Frequently, in early first-trimester procedures (<8 weeks), no dilation is required to insert the desired cannula, especially in multiparous women. According to a survey of North American providers, approximately half report dilating the cervix to a diameter in millimeters equal to the gestational age in weeks. An additional 37% of providers dilate to 1 to 2 mm greater than the number of gestational weeks. The remainder dilates to 1 to 3 mm less than the gestational age, using a smaller cannula to evacuate the pregnancy []. Most providers dilate the cervix incrementally, but some skip dilator sizes if the internal os is compliant [].

The Marie Stopes Procedure performed in some parts of the United Kingdom differs markedly from North American practices. In Marie Stopes outpatient centers, no rigid dilation or cervical priming is used for first-trimester surgical abortion []. Marie Stopes providers employ atraumatic tenaculums, flexible cannulas and 1% lidocaine gel for local anesthesia. They use the smallest cannula size possible for a given gestation to evacuate the uterine contents (<6 weeks, 4-mm cannula; 6 to 7 weeks, 5-mm cannula; 8 to 9 weeks, 6-mm cannula; and 10 to 12 weeks, 7-mm cannula). If any dilation is necessary, cannulas of the appropriate size are employed as dilators. In rare cases, os finders are needed to ascertain the path of the cervical canal. This method has not been compared to other techniques in any clinical trials.

When evaluating cervical priming agents, important clinical outcomes are whether adequate cervical dilatation can be achieved to complete the procedure as planned and whether complications are reduced []. Most studies have evaluated some combination of baseline cervical dilatation, need for further mechanical dilation, duration of the procedure, subjective assessment of ease of dilation, force required for dilation measured by cervical tonometer, intraoperative blood loss, premature passage of fetal or chorionic tissue, side effects (preoperative bleeding, fever, pain, nausea, vomiting and diarrhea), acceptability and complication rates [, , , , , , , , , , , , , , , , , , , , , , , , , , , , , ]. Primary outcomes on which sample sizes are calculated, however, tend to be baseline cervical dilatation or force required to reach a certain diameter of dilatation. These primary outcomes illustrate only whether or not cervical priming works to soften and dilate the cervix.

Another important outcome measure is patient-centered, the frequency of side effects and time spent waiting for the abortion procedure. Rarely are the side effects of cervical priming, and their acceptability to women is the primary objective of a study. Women generally prefer 1-day procedures to 2-day procedures and prefer misoprostol to laminaria for cervical priming [, ]. Most women find both the oral and vaginal routes of misoprostol acceptable [, , ]. The benefit of passive dilation of the cervix prior to the surgical procedure in terms of decreased operative time and improved ease of procedure must be balanced against the patient's individual circumstances, such as side effects experienced while waiting []. The frequency of side effects has varied in cervical priming trials. Often those showing no difference lack the power to detect an effect. Those that do show significant results find that the placebo arm experiences significantly fewer side effects than the misoprostol arm. Medication abortion trials have demonstrated that self-administration of vaginal misoprostol at home, or in the clinic, is acceptable to women [, , , ].

Ideally, the efficacy of cervical priming should be defined by its impact on complication rates. Whether or not cervical priming is worth the added inconvenience and side effects depends on whether it makes the procedure safer. Given the rarity of complications with first-trimester abortion, however, only a few studies have had sufficient power to evaluate a difference in complication rates using cervical priming [, , , ].

Immediate complications from suction abortion include uterine perforation, hemorrhage, cervical laceration (deep or superficial) and incomplete abortion. Risk factors for major complications differ according to the specific complication. For example, increasing parity is a risk factor for uterine perforation but not for cervical injury. This finding may reflect differential changes in the myometrium and cervical stroma after delivery [, ]. The increased risk of perforation observed with higher parity may also be related to the frequency of prior cesarean deliveries, but past studies do not always specify delivery mode. A recent large WHO multicenter trial of misoprostol versus placebo in first-trimester abortion did not find that parity was associated with the rate of acute or delayed complications, although the two deep cervical tears from dilation and four uterine perforations all occurred in the parous group []. Evidence for higher complication rates in adolescents is more compelling [, , ]. When the effect of parity is controlled for, adolescents have a higher risk of cervical injury, especially at gestations of more than 12 weeks, than do adult women [, ]. It appears that young age is not a proxy for nulliparity. Rather, adolescents (age ≤17) have small, physiologically immature cervices that may be more difficult to dilate than those of adult women, regardless of obstetric history []. The WHO trial cited above did not include enough adolescents to evaluate differing complication rates.

Existing data provide evidence of an increasing risk of immediate complications as gestational age increases, even in the first trimester [, , , , , ]. Because studies differed in their reference groups, categories of gestational age and whether or not second-trimester procedures were included, they provide no clear evidence to guide decisions about when cervical priming should begin []. Taken together, however, these studies suggest that the risk of complications increases after 9 weeks, accelerating at 12 and 13 weeks' gestation. Published data are insufficient to support any comment on the risk of complications specifically in women with prior cervical conization, cesarean section or cervical stenosis or in women who are obese or have acutely anteflexed or retroflexed uteri [, ].

In sum, women late in the first trimester and adolescents are more likely to benefit from cervical priming. Women with uterine anomalies or known cervical stenosis might also benefit [].

The insertion of one medium laminaria effectively dilates the cervix to at least 9 mm in more than 75% of women, reducing the need for rigid dilation [, , , , , ]. The dilatation achieved with laminaria tents increases with the number of tents used and with gestational age [, ]. Laminaria show some effect after 4 h, but more time is necessary for them to induce prostaglandin release and reach their maximum diameter [, ]. The diameter of each laminaria tent increases by 25% if left in place for 4 h and by 90% if left in place overnight [, ]. The use of laminaria prior to first-trimester surgical abortion, in certain clinical settings, decreases the incidence of cervical lacerations and, to a lesser extent, uterine perforations [, , , ].

Bokstrom and Wiqvist [] studied the use of Dilapan for cervical priming in women from weeks 10 to 12 of gestation. They found that a 3-mm Dilapan achieved 8 mm of dilatation at 3 to 4 h and 10 mm of dilatation at 16 to 20 h, while a 4-mm Dilapan reached 8.5 mm and 11.3 mm of dilatation, respectively. The additional dilation gained with the longer duration of retention might not be clinically significant in the first trimester.

Lamicel similarly swells rapidly, and its administration 4 h preoperatively is as effective as 16 h at increasing baseline cervical dilatation []. In fact, studies indicate that the total force required for dilation drops rapidly after 2 h of Lamicel placement and then plateaus []. One randomized trial of 629 women showed that Lamicel achieved a higher mean dilation than placebo (8.2 mm vs. 5.8 mm, p<0.001). However, the study failed to show that priming with Lamicel significantly reduced the rates of uterine perforation, infection or recurettage. However, the perforation rate in the Lamicel group was 0.4% versus 1.7% in the control group (nonsignificant), and the study may not have had sufficient power to detect a difference in complication rates [].

All types of osmotic dilators have advantages and disadvantages. Investigators have conducted a number of comparative trials testing their efficacy for cervical priming before first-trimester surgical abortion. Randomized trials have demonstrated that one Dilapan™ device outperforms a similar diameter laminaria with 4 h of use, but the two are equivalent at 6 h [, ]. A recent trial of Dilapan-S compared with 400 mcg of buccal misoprostol 3 to 4 h prior to surgical abortion at 12 to 15 weeks showed similar baseline cervical dilation (10.8 mm vs. 10.2 mm, p=0.065) []. There is a trend toward more cramping and difficult removals with Dilapan and laminaria than with Lamicel. Some surgeons report anecdotally that combining Dilapan with laminaria makes removal easier []; however, no evidence is available in the literature to support this practice. Finally, case reports of breakage, entrapment and displacement into the uterine cavity have involved both Dilapan and laminaria but not Lamicel [, , ]. Since the release of the reformulated version, however, Dilapan-S has not been associated with these adverse events [].

Important characteristics of osmotic dilators include convenience and side effects as well as effectiveness. The evidence indicates that Dilapan-S is superior to laminaria for short treatment periods (4 h) but that laminaria would achieve the same dilatation if allowed more time (at least 6 h). It also appears that Dilapan-S and multiple laminaria are superior to Lamicel for any given duration of time in causing wider initial dilatation [, ]. However, the ability to subsequently achieve the desired dilation easily with rigid dilators is comparable with all three methods. Therefore, a same-day procedure could more easily be accomplished with Dilapan-S or Lamicel than with laminaria. One Lamicel is approximately twice the cost of one Dilapan-S, and in turn, one Dilapan-S is twice the cost of one laminaria []. However, multiple laminaria are often required to achieve the same dilation as one Lamicel or one Dilapan-S []. To date, no published studies comparing Dilapan-S or Lamicel with a placebo have been large enough to detect a difference in complication rates.

Several randomized trials have compared misoprostol and placebo or no therapy for cervical priming before first-trimester surgical abortion and found that misoprostol increases baseline dilation and facilitates further dilation. These studies analyzed vaginal doses of 100 to 750 mcg [, , , , , , , , , ], oral doses of 400 to 600 mcg [, , ] and a sublingual dose of 400 mcg [, , ]. Almost all of these studies demonstrated an increased baseline cervical dilatation with misoprostol, and some also found a greater subjective ease of dilation [, , , , , ], lower measured cumulative force with dilation [, ], shorter procedure time [, , , , , , ] or lower estimated blood loss [, , , ]. The majority of studies measured baseline cervical dilation with Hegar dilators, but a few used a cervical tonometer to measure the force required for cervical dilation. Many of these studies reported more cramping (although mild) with misoprostol than with placebo [, , , , , ]. In general, differences between misoprostol and placebo in terms of operative time and blood loss are statistically, but not clinically, significant. The effectiveness of misoprostol is not influenced by whether manual vacuum aspiration or electric suction is used [].

In the largest trial to date, the WHO randomized 4972 women in 14 countries to either 400 mcg of misoprostol or placebo vaginally 3 h prior to surgical abortion up to 12 0/7 weeks of pregnancy []. In the misoprostol arm, there was a reduced risk of superficial damage to the cervix due to tenaculum tears [3 vs. 12, relative risk (RR): 0.25, 95% confidence interval (CI): 0.07–0.89]. Only 2 of 14 study sites reported this complication, which suggests that it may be related to provider experience or technique. There was no difference in the risk of deeper cervical tears due to dilation or uterine perforation between the two groups []. The risk of the delayed complication of incomplete abortion with or without the need for recurettage was lower in the misoprostol arm than in the placebo arm (19/2427 vs. 55/2431, RR: 0.35 95% CI: 0.21–0.58) []. The investigators attributed this difference to (a) the ability of providers to use larger cannula sizes for procedures in the misoprostol arm, thus facilitating evacuation of the uterus and (b) a larger cervical diameter postprocedure, which eased emptying of the cavity of any retained products of conception. However, it is also possible that the reduction in incomplete abortion observed with misoprostol was due to its uterotonic effects, assisting with postprocedure expulsion of any retained tissue. Of interest, this difference was more pronounced in multiparous women.

If misoprostol is used for cervical priming and the woman is unable to undergo surgical abortion as planned, then she is at some risk for expelling the pregnancy and theoretically a risk of congenital anomalies should she continue the pregnancy. Therefore, before receiving misoprostol, all women should give informed consent for the abortion procedure and be adequately screened by appropriately trained personnel.