Human Reproduction Update

Human Reproduction Update Advance Access originally published online on October 17, 2006
Human Reproduction Update 2007 13(1):37-52; doi:10.1093/humupd/dml049

This Article Abstract FREE Full Text (PDF) All Versions of this Article: 13/1/37    most recent dml049v1 Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Request Permissions Google Scholar Articles by Lalitkumar, S. Articles by Gemzell-Danielsson, K. Search for Related Content PubMed PubMed Citation Articles by Lalitkumar, S. Articles by Gemzell-Danielsson, K. Social Bookmarking          What's this?

© The Author 2006. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

Mid-trimester induced abortion: a review

S. Lalitkumar, M. Bygdeman and K. Gemzell-Danielsson¹

Department of Woman and Child Health, Division for Obstetrics and Gynaecology, Karolinska University Hospital/Karolinska Institute, Stockholm, Sweden

¹ To whom correspondence should be addressed at: Department of Woman and Child Health, Division for Obstetrics and Gynecology, Karolinska University Hospital/Karolinska Institutet, SE 17176 Stockholm, Sweden. E-mail: kristina.gemzell{at}ki.se

	Abstract

TOP Abstract Introduction Materials and methods Conclusions References

 
Mid-trimester abortion constitutes 10–15% of all induced abortion. The aim of this article is to provide a review of the current literature of mid-trimester methods of abortion with respect to efficacy, side effects and acceptability. There have been continuing efforts to improve the abortion technology in terms of effectiveness, technical ease of performance, acceptability and reduction of side effects and complications. During the last decade, medical methods for mid-trimester induced abortion have shown a considerable development and have become safe and more accessible. The combination of mifepristone and misoprostol is now an established and highly effective method for termination of pregnancy (TOP). Advantages and disadvantages of medical versus surgical methods are discussed. Randomized studies are lacking, and more studies on pain treatment and the safety of any method used in patients with a previous uterine scar are debated, and data are scarce. Pain management in abortion requires special attention. This review highlights the need for randomized studies to set guidelines for mid-trimester abortion methods in terms of safety and acceptability as well as for better analgesic regimens.

Key words: dilatation and evacuation / mid-trimester induced abortion / mifepristone / misoprostol / vacuum aspiration

	Introduction

TOP Abstract Introduction Materials and methods Conclusions References

 
Abortion is defined as ‘termination of pregnancy (TOP) by any means before the fetus is viable’. Viability is now considered to be reached at 23–24 weeks of gestation. Second trimester, or mid-trimester, is a period ranging from 13 to 28 weeks of gestation, which again is subdivided into an early period between 13 and 20 weeks and a late period between 20 and 28 weeks. In this review, we have limited late abortions up to 24 weeks gestation.

TOP by induced abortion is practised worldwide. Induced abortion, either elective or therapeutic termination of a viable pregnancy, is one of the most ancient procedures. Of the 210 million pregnancies that occur each year, >46 million (22%) end in induced abortions (Alan Guttmacher Institute, 1999). A majority (90%) of the terminations take place in the first trimester. Worldwide mid-trimester abortion constitutes 10–15% of all induced abortions but is responsible for two-thirds of all major complications (WHO, 1997). Although the majority of abortions are performed in the first trimester, there is still a gradual increase in second-trimester abortion because of the wide scale introduction of prenatal screening programs detecting women whose pregnancies are complicated by serious fetal abnormalities such as cardiovascular and skeletal malformation.

Over the last 20 years, there have been continuing efforts to improve the abortion technology in terms of effectiveness, decreasing rates of complications, technical ease of performance and acceptability. During this time, >20 countries have partially or fully liberalized their abortion laws (Berer, 2004). Today, in almost all countries, the law permits abortion to save a pregnant woman’s life. However, the requirement of legalization is no guarantee for a safe abortion. In many countries where abortion is illegal, as in Latin America, private physicians often perform safe abortions for relatively high medical fees, and the law is rarely enforced (Fathalla, 1997). The larger population, which cannot afford such fees, succumb to the cheaper ways of abortion practices that eventually risk their lives. This may be because of lack of safe services provided in the public sector and lack of access to referral units that exist. Today, in most cases, safe and efficient abortion services can be offered or improved by minor changes in existing healthcare facilities.

Aim of the study

The aim of this study is to provide a review of the current literature on mid-trimester methods of abortion with respect to efficacy, side effects and acceptability and also to provide evidence-based recommendations for safe regimen(s) for mid-trimester pregnancy terminations.

	Materials and methods

TOP Abstract Introduction Materials and methods Conclusions References

 
Search criteria

Electronic literature search (English) of MEDLINE database using following keywords: Mid-trimester induced abortion, second-trimester pregnancy termination, induced abortion, mifepristone, misoprostol, gemeprost, medical abortion in prior Caesarean section (CS) and pain management.

Types of studies
All randomized controlled trials, large case series with data on mid-trimester TOP.

Types of outcome measures
Primary outcomes:

1. Rate of complete abortion.
   
2. Failure to achieve complete abortion with intended method.
   
3. Induction-to-abortion interval.
   

Secondary outcomes:

1. Safety of the method/regimen used.
   
2. Acceptability of the method used.
   
3. Excessive blood loss either measured or estimated by a clinically relevant drop in haemoglobin.
   
4. Pain resulting from the procedure, reported by the women or measured by use of analgesics.
   
5. Post-abortion curettage required in women with medical abortion method.
   
6. Side effects such as pyrexia, nausea, vomiting and diarrhoea.
   
7. Uterine rupture.
   
8. Infectious morbidity.
   
9. Mortality.
   
10. Any other grave complication.
    

Background

Abortion dates back to the period of Socrates, Plato, Aristotle and Hippocrates (Anonymous, 1995). Different surgical and medical methods of abortion have been used since the early age. Surgical abortion is one of the oldest and most commonly practised techniques in many parts of the world. A matter of great concern was that there were no safe drugs for inducing an abortion. Since ancient time, women have used various herbs, salts, douches and purgatives, all with questionable success to achieve pregnancy termination (Riddle et al., 1993). In recent years, effective medical abortion methods with low morbidity have been emerging and become better accessible (Figure 1).

View larger version (18K): [in this window] [in a new window] [Download PowerPoint slide]   Figure 1. Second trimester methods of abortion.

 
In the early 1970s, the most commonly used methods were vacuum aspiration (VA), dilatation and curettage, sharp curettage, hysterotomy (sectio parva), intra-amniotic injection of hypertonic saline or hyperosmolar urea, intra- or extra-amniotic administration of ethacryidine lactate (Rivanol), parenteral, intra-amniotic or extra-amniotic administration of prostaglandin (PG) analogues and i.v. or i.m. administration of oxytocin (WHO, 1997). Rivanol is a dye with antiseptic properties and seems to be less toxic than hypertonic saline. As with hypertonic saline, Rivanol stimulates endogenous PG and thromboxane production, probably because of chemical trauma to the fetal membranes and the decidua, promoting cervical priming and initiating labour. The instillation to delivery interval ranged from 25 to 40 h (Ingemarsson, 1979; Bhathena et al., 1990; Blumenthal et al., 1999) which could be reduced to 15–20 h with concomitant use of oxytocin (Yapar et al., 1996). Among the drawbacks of all these medical methods are the need for puncture of the intra-amniotic space or the introduction of a Foley catheter into the extra-amniotic space, a relatively long induction-to-abortion interval and the need for curettage after the expulsion of the fetus (Bygdeman, 1983). The i.v. infusion of oxytocin was inconvenient to use because of the serious side effects of water intoxication. The risk of disseminated intravascular coagulation with hypertonic saline is 0.8% (Edelman et al., 1976). This is because of rapid intravascular absorption of hypertonic saline from the amniotic cavity or inadvertent i.m. or i.p. injection of saline resulting in hypernatremia and necrosis of the affected tissue. Thus, many of these methods were very cumbersome in respect of their side effects and medical expertise required.

With the introduction of PGs and later PG analogues, the efficacy of medical abortion could be improved, and the risk for complications and side effects was reduced. The method of medically induced abortion could be further improved as mifepristone became available in the 1980s (Bygdeman and Swahn, 1985; Urquhart and Templeton, 1987; Swahn and Bygdeman, 1988; Silvestre et al., 1990; Gottlieb and Bygdeman, 1991). With mifepristone, the induction-to-abortion interval was shortened, and the dose of PG analogues required was reduced. Today, medical abortion is the method of choice in many centres [Royal College of Obstetricians and Gynaecologists (RCOG), 1997].

PGs and PG analogues

PGs play an important role in the regulation of uterine contractility during pregnancy (Mitchell, 1987). The receptors are present throughout the pregnancy; hence, PGs and their analogues are effective for TOP. Naturally occurring PGs, mainly prostaglandin E_l (PGE₁), PGE₂ and PGF₂, are potent stimulants of uterine contractility at any stage of pregnancy and also cause cervical ripening and dilatation. However, because of the rapid metabolism and high incidence of gastrointestinal side effects, they had a limited role in induced abortion and were soon replaced by PG analogues that are more suitable for clinical application (Gillet et al., 1972; Wiqvist et al., 1972; Lauersen et al., 1975). The PG analogues, besides being suitable for administration by non-invasive routes, have a prolonged action as they are relatively resistant to the initial rapid inactivation in the circulation (because of the introduction of a methyl group at C15 or C16). Both PGE and PGF analogues are widely used for TOP, although PGE analogues are preferred because of their selective specificity for the myometrium and fewer gastrointestinal side effects (World Health Organization Task Force on Prostaglandins, 1988). The most extensively studied PG analogues are carboprost, sulprostone, gemeprost and misoprostol.

Carboprost, a 15(S)-15-methyl PGF₂, was the first analogue to be tested clinically on a large scale for the termination of second-trimester pregnancy. It is used either intra-amniotically (viable second-trimester pregnancy) or administered by i.m. injection. It is of limited value as a primary method for abortion because of its association with high rates of gastrointestinal side effects but may be used when other methods have failed (Lauersen and Wilson, 1976; World Health Organization Task Force on Prostaglandins, 1977a). Both carboprost and another PG analogue meteneprost have been tried in the form of vaginal gel in large multicentre trials, and the effects have been similar to the i.m. route (Bygdeman et al., 1975; World Health Organization Task Force on Prostaglandins, 1977b; Green et al., 1988).

Sulprostone, a 16phenoxy-w-17, 18, 19, 20-tetranor PGE₂ methyl sulphonylamide, was used in the 1980s for the termination of second-trimester pregnancy (World Health Organization Task Force on Prostaglandins, 1982, 1988). It was withdrawn from the market because of its association with severe cardiovascular complications, including myocardial infarction attributed to coronary spasm (Ulman et al., 1992; Peyron et al., 1993). No such similar complications have been reported with other PGs.

Gemeprost is a PGE₁ analogue (16, 16-dimethyl-trans-d2-PGE₁ methyl ester) and is used as a vaginal pessary. It has been extensively used as a non-surgical method to dilate the cervix before VA in late-first and early-second-trimester abortion (Smith and Baird, 1980; Welch and Elder, 1982). As a method for TOP in the second trimester, it has been well established that gemeprost is more efficacious when compared with intra-amniotic PGF₂ or extra-amniotic PGE₂ and dinoprostone intracervically (Cameron and Baird, 1984; Andersen et al., 1989; Kjolhede et al., 1994).

Misoprostol is a synthetic PGE₁ analogue (15-deoxy-16-hydroxy-16-methyl PGE₁), initially developed for the prevention and treatment of peptic ulcer and later used off-label as an abortifacient. It has several advantages over other PGs; it is cheap, stable at room temperature and can be stored for a long time. Misoprostol has a limited effect on the bronchi or blood vessels. The oral tablets are effective in different routes of administration. Misoprostol in the required doses has only few (and dose-dependent) side effects, and it is readily available in many countries (http://www.gynuity.org/documents/miso_approval_2004_map_revised_12.05.pdf).

Pharmacokinetics of misoprostol

Misoprostol is rapidly absorbed after oral administration, primarily metabolized in the liver and converted to its pharmacologically active metabolite, misoprostol-free acid. Less than 1% of this metabolite is excreted in urine. The plasma concentration of misoprostol acid after oral administration peaks at 30 min and declines rapidly thereafter with a terminal half-life of 20–40 min, whereas after vaginal administration, the levels increase gradually and reach maximum levels after 70–80 min but remain detectable for a significantly longer time. There is a high variability of plasma levels of misoprostol acid between and within studies (Foote et al., 1995; Zieman et al., 1997). Pharmacokinetic studies show that the systemic bioavailability of vaginally administered misoprostol is three times higher than that of orally administered misoprostol. With vaginal administration, peak plasma levels occurred later and were lower, but elevated plasma levels were sustained for at least up to 4 h (Zieman et al., 1997; Danielsson et al., 1999; Tang et al., 2002). A study on the pharmacokinetics following the sublingual, vaginal and oral routes of misoprostol administration for medical abortion demonstrated that sublingual misoprostol reached peak concentration in the shortest time and had the highest bioavailability (Tang et al., 2002). It was shown that the time to peak concentration after sublingual administration was similar to the oral route, but plasma concentration achieved was higher and sustained. Sublingual misoprostol had a higher serum peak concentration and bioavailability (measured as the area under the curve for the plasma levels) compared with vaginal administration.

Effects of misoprostol on uterine contractility
The ideal agent for the induction of abortion is one that gives rise to uterine contractions subsequent to effective cervical dilatation. PG analogues induce cervical ripening both by a direct effect on the cervix and by a concomitant stimulation of myometrial activity (Uldbjerg and Ulmsten, 1990; El-Refaey et al., 1994). The effect of misoprostol on uterine contractility following different routes (sublingual, vaginal and oral) of administration has been well documented (Aronsson et al., 2004). The time from the start of the treatment to the onset of effect (increase in uterine tonus) and to maximum tonus elevation was shorter with oral and sublingual compared with vaginal administration. The study also showed that after a single oral administration of misoprostol, regular uterine contractions did not develop. Following sublingual and vaginal administration, on the contrary, increased uterine contractility was observed 2 h after administration. The effect on uterine contractions after sublingual administration seemed to be shorter lasting than that seen following vaginal administration. This effect would probably be more significant for misoprostol-only regimens that require more repeated doses of misoprostol to achieve the clinical outcome. For regimens with mifepristone pretreatment, the doses of misoprostol required are usually less, and therefore, sublingual or even oral misoprostol may be enough. Sublingual misoprostol was considered to be more convenient and less uncomfortable but associated with bad taste and a higher frequency of side effects (Tang et al., 2004). The most commonly encountered side effects were fever, shivering, vomiting and diarrhoea. Another advantage with the oral/sublingual route is that the absorption of the drug is not affected if the woman starts to bleed. This is especially true when the repeated administration of misoprostol is required.

Antiprogesterone

Progesterone is a key hormone in maintaining the pregnancy by keeping the uterus in a quiescent state. Progesterone prevents softening and dilatation of the cervix, it reduces PG output from the decidua, and it suppresses uterine contractions by inducing hyperpolarization of the cell membrane, which makes the myocytes less sensitive to electrical stimulation. Furthermore, progesterone induces the inhibition of gap junction formation resulting in the counteraction of co-ordinated uterine contractions. Progesterone antagonists are synthetic steroids that bind to the progesterone receptors and prevent endogenous progesterone from exerting its action (Van Look and Bygdeman, 1989). Mifepristone is the only anti-progestin approved for the induction of abortion. It is a 19-norsteriod substituted at the 11ß position by a p-dimethylamino phenyl group, which binds with high affinity to the progesterone receptor, thus inhibiting the effect of the hormone. Its binding affinity for the progesterone receptor is 2.5–5 times that of progesterone (Lähteenmäki et al., 1987). Blockage of the progesterone receptor results in vascular damage, decidual necrosis and bleeding (Bygdeman and Swahn, 1985; Johansson et al., 1989). Treatment with mifepristone will soften the cervix, increase the sensitivity to PGs and convert the quiet pregnant uterus into an organ of spontaneous activity (Bygdeman and Swahn, 1985; Swahn and Bygdeman, 1988; Norman et al., 1991). The sensitivity of the myometrium is increased by 5 times with maximal effect on uterine contractility and cervical ripening at 36–48 h following treatment (Bygdeman and Swahn, 1985; Rådestad et al., 1988). The benefits of pretreatment with mifepristone in comparison with placebo and laminaria tent are well evidenced (Table I, Panels a and b). Mifepristone and a PG analogue act in synergy, thus a change in the dose/type of one drug or in the route of administration and interval between the mifepristone and PG analogue will have impact on the required dose of the other and the subsequent efficacy and side effects.

View this table: [in this window] [in a new window]   Table I. Methods of mid-trimester induced abortion

 
The pharmacokinetics of mifepristone is linear up to doses of 100–200 mg, and above that, it is non-linear (Swahn et al., 1986; Heikinheimo et al., 1987). The approved dose of mifepristone in medical mid-trimester abortion is 600 mg (similar to early medical abortion), but it has been shown that the abortion rate and induction-to-abortion interval were the same even if the dose was reduced to 200 mg (Webster et al., 1996).

Although maximal priming effect on the myometrium is achieved 36–48 h after pretreatment with mifepristone, no difference was seen in induction-to-abortion time with mifepristone administered 24, 36 or 48 h before PG administration that probably depends on the PG dose used (Urquhart and Templeton, 1990a). In another study, a shorter interval resulted in a slightly longer induction-to-abortion interval and higher dose of misoprostol used (Heikinheimo et al., 2004).

Around 0.2–0.4% of women abort with mifepristone only (UK Multicenter Study Group, 1997; Tang et al., 2001). The concept of genetic variation of the progesterone receptor has been postulated in first-trimester abortions among women with a continued pregnancy who did not respond to mifepristone–PG combination. There is a need to look more in depth into failed or prolonged second-trimester abortion (Gao et al., 1998).

Medical abortion with mifepristone and a PG analogue
For mid-trimester abortion (13–24 weeks of gestation), medical abortion with mifepristone followed by PG is an appropriate method and has been shown to be safe and effective (RCOG, 2004). It has been well proven that pretreatment with the antiprogesterone mifepristone 36–48 h before PG administration can increase the success rate, shorten the induction-to-abortion interval and reduce the amount of PGs required in second-trimester abortion (Thong and Baird, 1992a; Ho et al., 1995, 1997).

Medical abortion during early pregnancy was first approved in France in 1988 (up to 49 days amenorrhoea) followed by approvals in the UK (1991) and Sweden (1992) (up to 63 days of amenorrhoea in both the countries). A few years later, mifepristone together with a PG analogue was also approved for second-trimester abortion (in Sweden in 1994). Mifepristone has been used in China since 1992. However, it was only in 1999/2000 that both early-first- and -second-trimester medical abortions with mifepristone and PG were approved in several other European countries. Today, mifepristone is available in 29 countries worldwide (http://www.gynuity.org/documents/mife_approval_2005_map.pdf). Since the introduction of the method, extensive research is ongoing focusing on improving efficacy, defining the optimal dose/type and route of the administration of PG analogue.

Mifepristone and gemeprost
The vaginal gemeprost-only regimens achieved an abortion rate of 88–96.5% with longer induction-to-abortion interval. With pretreatment with mifepristone 36–48 h before gemeprost, the induction-to-abortion could be decreased to nearly half (from 15.7 to 6.6 h), and the abortion rate in 24 h was increased from 72 to 95% (Van Look and Bygdeman, 1989). Furthermore, side effects were reduced, and the dose of mifepristone and gemeprost decreased to one-third and one-half respectively, without the loss of clinical efficacy (Thong and Baird, 1993). In a case series report on 197 consecutive cases treated with mifepristone and gemeprost, a significant correlation between pregnancy length and abortion time was shown (Gemzell-Danielsson and Östlund, 2000). The induction-to-abortion interval was also shorter in parous women. Gemeprost was considered as the standard PG analogue in medical abortion and cervical priming until misoprostol emerged and was made available (Bartley et al., 2001). Although shown to be highly effective, gemeprost has several disadvantages compared with misoprostol (i.e. cost, need for refrigeration limits, its usage in developing countries and it is only available as vaginal pessary), which has replaced gemeprost on all indications.

Mifepristone and misoprostol
Misoprostol has been shown to be equally or more effective compared with gemeprost (Ho et al., 1996; Bartley and Baird, 2002) (Table II, Panel C). In a study of 98 women, it was shown that vaginal misoprostol is more effective than oral misoprostol after pretreatment with mifepristone, but more women preferred the oral route (Ho et al., 1997). The induction-to-abortion interval was shorter, and the amount of misoprostol required was lower after vaginally administered misoprostol (Ho et al., 1997; Ngai et al., 2000) (Table II, Panel b). The incidence of diarrhoea was higher with oral misoprostol. However, women preferred oral to vaginal administration considering it more convenient and giving more privacy (Ho et al., 1997; Ngai et al., 2000). Therefore, to improve acceptability, a regimen using a combination of an initial high dose administered vaginally followed by repeated oral doses of misoprostol was developed, which involved the use of 600 µg of vaginal misoprostol as the first dose followed by 400 µg of oral misoprostol for every 3 h. The abortion rate (97%) and the induction-to-abortion interval (6.5 h) were the same as using similar doses of repeated vaginal misoprostol (El-Refaey and Templeton, 1995) (Table II, Panel b). The results were later confirmed in a larger series of patients using a slightly higher initial dose of 800 µg of vaginal misoprostol (Ashok and Templeton, 1999; Ashok et al., 2004) (Table II, Panel b). It was believed that the use of misoprostol vaginally as the first dose could lead to more effective cervical priming, but there was no advantage in the vaginal administration of subsequent doses (Ashok and Templeton, 1999).

View this table: [in this window] [in a new window]   Table II. Methods of mid-trimester induced abortion

 
More recently, it was shown that the combination of mifepristone and sublingual misoprostol provides a safe and an effective regimen for medical abortion (Hamoda et al., 2005a; Tang et al., 2005) (Table II, Panel b). In the first randomized study comparing sublingual versus vaginal misoprostol in the second trimester, no pretreatment with mifepristone was given (Tang et al., 2004) (Table II, Panel d). In this study, the use of vaginal misoprostol resulted in a higher success rate than sublingual misoprostol in 24 h, but the abortion rate was similar at 48 h (Table II, Panel d). The findings might be because of the more prolonged effect of vaginal misoprostol on uterine contractility. When pretreatment with mifepristone was given, all women aborted within 24 h of receiving the first dose of misoprostol, and the induction-to-abortion interval for both study groups (vaginal and sublingual) was less than half of that noted in the former study with misoprostol alone (Hamoda et al., 2005a; Tang et al., 2005) (Table II, Panel b). The authors concluded that this effect could be because of prior treatment with mifepristone and the higher initial dose of misoprostol used (Table II, Panel b).

Acceptability was higher in the sublingual group despite the significantly higher rate of side effects. The sublingual route would be the alternative for women who do not like the vaginal route of administration.

Surgical evacuation of the placenta

Routine surgical evacuation of the uterus is not required following mid-trimester medical abortion. It should only be undertaken if there is clinical evidence that the abortion is incomplete (El-Refaey and Templeton, 1995). In recent large case series of mid-trimester medical abortion, only 8–11% of women needed surgical evacuation following medical abortion (Ashok and Templeton, 1999; Tang et al., 2001; Ashok et al., 2004) (Table II, panels a and b). In the latest report from the Scottish group, the rate of surgical evacuation was as low as 2.5% (Hamoda et al., 2005a). These figures thus negate the need to do a routine curettage unless warranted. A very low incidence of surgical evacuation is also consistent with the previous reports (El-Refaey et al., 1993). Complete abortion is achieved with increasing frequency with advancing gestations with >80% occurring at 20 weeks (UK Multicenter Study Group, 1997). Performing a routine evacuation, however, does not protect against the need for hospital readmission for post-abortion bleeding and uterine curettage (UK Multicenter Study Group, 1997) (Table II, Panel a). A more determined approach to use surgical evacuation only when indicated would probably reduce the length of hospital admission. For this to be achieved, the use of staff experienced in assessing placental completeness after abortion is essential.

Medical abortion with PG alone
Medical abortion with gemeprost or misoprostol alone has been shown to be effective, although higher doses are needed; side effects are more frequent and the induction-to-abortion interval is longer compared with the combined treatment with mifepristone.

Vaginal application of gemeprost gave an abortion rate of about 88–96.5% in 48 h, and the mean induction-to-abortion interval ranged from 14 to 18 h (Cameron et al., 1987; Thong and Baird, 1992b; Thong et al., 1992; Armatage and Luckas, 1996; Nuutila et al., 1997; Wong et al., 1998) (Table II, Panel d). The main side effects reported were vomiting, diarrhoea and fever (Thong et al., 1992; Nuutila et al., 1997; Wong et al., 1998). The most common regimen studied is 1-mg gemeprost every 3 h for 5 doses in 24 h. It is repeated if abortion does not occur within this time. When 3- and 6-h administration of gemeprost were compared, there was no advantage of 3-h intervals (Armatage and Luckas, 1996) (Table II, Panel d). These results suggest that, by lengthening the interval between insertions of pessaries within the first 24 h, the number of pessaries could be reduced without the loss of clinical efficacy.

Misoprostol has been widely studied in different dosages and routes for the second-trimester TOP (Jain and Mishell, 1994; Nuutila et al., 1997; Wong et al., 2000). Various studies have used doses ranging from 200 to 800 µg at intervals ranging from 3 to 12 h (Jain and Mishell, 1994; Ho et al., 1997; Nuutila et al., 1997; Jain et al., 1999; Ngai et al., 2000; Wong et al., 2000; Bebbington et al., 2002; Dickinson and Evans, 2002). In general, one can expect a similar success rate and induction-to-abortion interval as with gemeprost alone (Table II, Panel d). The misoprostol-only regimen would be of use in those countries where mifepristone is not available. Doses of 600 and 800 µg have shown comparable successful abortion rates but are associated with high rates of fever, diarrhoea, nausea and vomiting (Herabutya et al., 2000, 2001; Pongsatha and Tongsong, 2001). It has been seen that 3-h interval is more effective than 6-h interval (Wong et al., 2000) (Table II, Panel d).

Feticide before late abortion
When medical abortion is chosen, in many settings, clinicians are required to ensure that the fetus is dead at the time of abortion. A legal abortion must not be allowed to result in a live birth and terminations after 21 weeks, the method chosen should ensure that the fetus is not born alive (RCOG, 1996). This is a matter of concern especially for late terminations with or without fetal malformations when one according to local guidelines has to resuscitate if the fetus is born alive. Agents used for feticide are hypertonic saline, 1% lidocaine and potassium chloride (Elimian et al., 1999; Bhide et al., 2002; Senat et al., 2003). Feticide with potassium chloride reduced the PG requirement for mid-trimester medical abortion, compared with similar procedures conducted without feticide (Elimian et al., 1999). Up to 20 weeks of pregnancy, the contractions induced by PG make feticide unnecessary.

Mid-trimester surgical abortion

Vacuum aspiration
VA is the surgical method of choice for first-trimester pregnancy termination. During the procedure, the uterus should be emptied by suction curette and by blunt forceps (if required). This procedure can also be used during the early second trimester. It is generally agreed that the risk of complications increases with gestational age (Brenner and Edelman, 1974). After 8 weeks of gestation, the risk of major complications appears to rise by 15–30% for each week of delay (Cates et al., 1979). The method of choice at gestations 12–15 weeks depends on the skill and experience of the concerned clinicians. It is well evident from a cohort study that abortion procedures, particularly those at 12 weeks, should not be allotted to the most junior staff member (Child et al., 2001). Surgical abortion by conventional suction termination, without the need for specialized instruments, can be undertaken up to 15 weeks of gestation if clinicians have gained experience with this method (RCOG, 2004). The complications could be reduced by preoperative cervical dilatation (Schulz et al., 1983; Grimes et al., 1984). Cervical trauma, the rate of damage to the external cervical os, at the time of surgical abortion is reported to be 1% [Jacot et al., 1993; Peterson et al., 1983; Schulz et al., 1983; Royal College of General Practioners (RCGP), RCOG, 1985]. The rate is lower when abortions are performed by experienced clinicians and early in pregnancy. The World Health Organization’s Technical and Policy Guidance on Safe Abortion and RCOG recommends ‘cervical preparation before surgical abortion for durations of pregnancy over 9 weeks for nulliparous women, for women younger than 18 years of age, and for all women with durations of pregnancy more than 10 weeks’. Cervical dilatation can be achieved through mechanical dilatation, laminaria tents, with PG analogues and mifepristone. A large number of studies have proven the safety of PG analogues for cervical dilatation when compared with mechanical dilatation and laminaria tents (Gold et al., 1979; Christensen et al., 1983). The degree of cervical dilatation is related to the duration of treatment. The use of misoprostol/mifepristone has been proven effective for cervical priming before surgical abortion in the first trimester (Urquhart and Templeton, 1990a; Urquhart and Templeton, 1990b; Ngai et al., 1999). In a study by Gottilieb et al. (1991), constituting 127 women at 13–14th week of pregnancy, no case of cervical injury or uterine perforation or recurretage was reported, and a post-abortion infection rate of 1.6% was noted. The mean amount of blood loss was 49 ml (range 0–400 ml), and only six patients had blood loss >100 ml. These results were comparable to the same procedure when used in the first trimester without an increased risk, provided the patients were pretreated with PGs.

Dilatation and evacuation

Dilatation and evacuation (D&E) is the standard method at gestations above 13 weeks in many parts of the world. The conventional suction termination would be an appropriate method for gestations between 12 and 15 weeks, whereas D&E would be a safe and an effective option for gestations above 15 weeks when undertaken by specialist practitioners with a sufficient workload to maintain their skills according to RCOG (2004). Although the safety and efficacy of D&E for the termination of mid-trimester pregnancy by experienced hands is reassured by the evidence provided (Grimes et al., 1977; Schneider et al., 1996; Autry et al., 2002), some practitioners feel it very distressing to perform this procedure at an advanced gestation. A report on the confidential inquires into the maternal deaths in the UK questioned the appropriateness of D&E as a method of terminating second-trimester pregnancy when safe and effective medical alternatives exist (Report of Confidential Enquires into Maternal Deaths in the United Kingdom, 1994–1996). When D&E was compared to primary PGs, PGF₂, D&E was found to be faster, safer and more acceptable up to about 18 weeks of gestation (Grimes et al., 1980). Cervical injury is more frequent with D&E in the second trimester, and hence, preoperative cervical priming reduces the complications (Schulz et al., 1983; Grimes et al., 1984).

Surgical versus medical abortion
Medical abortion at gestations 9–13 weeks has been shown as a safe, acceptable and effective alternative to surgical abortion (Ashok et al., 1998, 2002; Hamoda et al., 2005b). A randomized controlled trial comparing medical abortion with VA at gestations up to 9 weeks showed that, although both the methods were highly acceptable to women, medical abortion was more painful and less effective with advancing gestation (Henshaw et al., 1993, 1994). Since then, the developments of improved regimens including the use of misoprostol has had lot of advantages such as increased efficacy and reduced side effects of medical abortion. Presently, medical abortion is as effective as VA in the late-first and early-second trimester (Ashok et al., 2005a,b).

No randomized study comparing mifepristone and a PG analogue and D&E for mid-trimester abortion has been published. An attempt has been performed, but the trial was stopped after 1 year because of slow enrolment (Grimes et al., 2004). The women declined the method of abortion in each group, and sample size was too small to draw conclusive evidence except for hypothesis generation. In a retrospective cohort study of 297 women, the complication rate of D&E and medical methods for mid-trimester abortion was compared (Autry et al., 2002). The method used for medical abortion was not specified, but in most cases vaginal misoprostol alone was used. The combined treatment with mifepristone was not available. Besides a higher frequency of failed treatment (7 versus 0%) and incomplete abortion (21 versus 0.7%), there were no differences in complication rates. One patient with a history of a previous Caesarean delivery who received misoprostol 200 µg vaginally every 4 h (total dose not reported) had a uterine rupture. There is an increased risk of perforation of the uterus during D&E at advancing gestations. The use of real-time ultrasonography during D&E was reported to reduce the perforation rate. The rate of uterine perforation was 0.2% in the scanned group compared to 1.4% in the unscanned group (Darney and Sweet, 1989). Historically, it has been considered that D&E is a risk factor for subsequent adverse pregnancy outcomes, including cervical incompetence, pregnancy loss and preterm birth. A retrospective case series of 600 women with mid-trimester D&E concluded that ‘second trimester D&E is not a risk factor for subsequent mid-trimester pregnancy loss or spontaneous preterm birth’ (Kalish et al., 2002). In this study, interpretation of the findings was difficult as no reference cohort of women who had not undergone D&E was described. But still, the rates of pregnancy outcomes appeared similar to those of an unselected population. Thus, D&E can safely be undertaken by gynaecologists who have been trained in the procedure, have the necessary instruments and have a caseload sufficient to maintain their skills. For those lacking the necessary expertise and caseload and for the betterment of their patient’s mid-trimester medical abortion, using mifepristone with PG is appropriate. Side effects including nausea, vomiting and diarrhoea are characteristics of PG administration and are because of PG’s stimulatory effect on the gastrointestinal tract. Diarrhoea is more common in women using gemeprost, whereas fever is more common with misoprostol (Wong et al., 1996). Serious complications, such as uterine rupture, major haemorrhage and cervical tear, are rare (UK Multicenter Study group, 1997; Gemzell-Danielsson and Östlund, 2000). Uterine rupture cases are reported to occur with both gemeprost and misoprostol, with or without priming by mifepristone (Wiener and Evans, 1990; Norman, 1995; Chen et al., 1999). The incidence of uterine rupture in women without previous scar is estimated to be 0.1–0.2% in the second trimester of pregnancy using mifepristone and gemeprost (Atienza et al., 1980; RCOG, 2004). Major bleeding is usually associated with prolonged retention of the placenta. The drawbacks of medical abortion with older methods such as long duration of labour, hospitalization for several days, need for curettage and invasive administration have been reduced or eliminated when mifepristone and misoprostol (or gemeprost) is used.

There are only few studies reporting regimens for women who do not abort within 24 h. According to some protocols, if abortion does not occur, mifepristone is given followed by repeated vaginal misoprostol (Ashok et al., 2004). Any patient who fails to abort during the second day will get a third dose of mifepristone followed by gemeprost 1 mg every 3 h. There is still insufficient consensus to set a guideline for the failed abortion group of patients, but it could be argued that for women going on to a second or third day, D&E would be a more appropriate approach (Ashok et al., 2004). However, in a later series of 1002 cases with a combination of mifepristone and misoprostol, seven women who failed to abort after the third course eventually aborted using mifepristone and gemeprost (Ashok et al., 2004).

There is a gradual increase in second-trimester abortion because of wide scale introduction of prenatal screening programs detecting women whose pregnancies are complicated by serious fetal abnormalities such as cardiovascular and skeletal malformation. In these cases, examination of the fetus could provide valuable information especially after medical abortion to confirm the congenital anomaly and further evaluate the subsequent recurrence risk and provide information to help in counselling of these patients (Boyd et al., 2004).

Day-care abortion
Day-care abortion is recognized as a cost-effective management of service provision. In the era of using older methods for inducing abortion in the second trimester, the majority of women had to be inpatients for a couple of days to achieve abortion. The availability of abortion as a day-care procedure can minimize disruption to the lives of women and their families. The introduction of treatment with mifepristone before mid-trimester abortion with PGs has reduced induction-to-abortion intervals to an extent such that many women undergoing these procedures can be managed as day cases. In a multicenter study performed in 20 hospitals including 267 women, almost 60% of the women aborted within 8 h of PG being commenced (UK Multicenter Study Group, 1997). Of the 956 women, 68.6% aborted within 10 h (Tang et al., 2001). In a study of 100 women receiving 200 mg of mifepristone followed 36–48 h later by either gemeprost or misoprostol, >80% of the women aborted the fetus within 12 h in both the groups, confirming that in the majority it is possible to arrange the abortion as a day-care procedure (Bartley and Baird, 2002). In a case series report of 1000 women undergoing mid-trimester abortion with mifepristone followed by misoprostol, over two-thirds were managed as day-cases (Ashok et al., 2004). From these studies, it is suggested that up to 10% of women undergoing induced medical abortion will require inpatient care for medical, social or geographical reasons (or a combination of these).

Pain management
Abdominal pain is one of the most common adverse effects of medical abortion (Spitz et al., 1998; Honkanen et al., 2004). Pain is subjective and difficult to quantify. The amount of pain varies from woman to woman, and it is important to discover what are the factors causing severe pain and what helps to alleviate the pain. Services should make a range of oral and parenteral analgesics available to meet women’s needs (RCOG, 2004). In routine clinical practice, analgesia is offered to women following surgical abortion and both during and after medical abortion. However, the analgesia requirements and regimens for medical abortion reported in the literature vary widely (Wiebe, 2001), and there is scant research evidence to select an analgesic regimen and to show the predictors of analgesia (Smith et al., 1979; Westhoff et al., 2000). Analgesic requirement and the perception of pain were significantly higher in women of younger age, longer gestation, those with longer induction-to-abortion interval and with increased number of misoprostol doses (Smith et al., 1979; Belanger et al., 1989; Borgatta and Nickinovich, 1997; Westhoff et al., 2000; Hamoda et al., 2004), whereas the use of analgesia was less in older, parous women and those at shorter gestations (Hamoda et al., 2004). The pain associated with the mid-trimester TOP is probably because of the larger fetus passing through the cervical canal inducing more pain. In a study population of 2049 subjects undergoing first-trimester medical abortion, the variable ‘most strongly’ (P < 0.0001) associated with narcotic use was study center (Westhoff et al., 2000). Further research is needed to evaluate the role of advance or prophylactic analgesia and its effectiveness as well as women’s satisfaction and acceptance.

Non-steroidal anti-inflammatory drugs (NSAIDs) have been shown to be very effective for treatment of dysmenorrhoea, which is related to intensive uterine contractility (Smith, 1987). They inhibit the production of endogenous PGs, which are important messengers responsible for uterine contractions, cramps and pain sensation. Because treatment with mifepristone increases the endogenous production of PGs that might be of importance for the induced uterine activity (Norman et al., 1991), there has been concern that the use of NSAIDs might attenuate uterine contractility if used for pain treatment during medical abortion. Therefore, such drugs are frequently avoided or recommended against in protocols for medical abortion. However, this recommendation does not seem valid because several studies have shown that the use of NSAIDs did not interfere with the action of misoprostol and/or mifepristone on inducing cervical ripening, uterine contractility (Norman et al., 1991; Creinin and Shulman, 1997; Li et al., 2003) or the time to abortion and expulsion of the products of conception (Fiala et al., 2005).

During suction evacuation or D&E procedure, a majority of the patients experience severe pain. There are usually strong uterine contractions at the end of the suction evacuation procedure, which is when the women are more likely to experience severe cramps and pain. These contractions may last for a few hours, which may lead to post-operative pain. The evacuation pain sensation is a combination of pain because of cervical dilatation and because of uterine contractions. Being simple and easy to administer, a paracervical block (PCB) is accepted as a standard method for local anaesthesia during the procedure. A few randomized control trials have shown that PCB can reduce the pain (Wiebe, 1992; Glantz and Shomento, 2001). The cervix and the lower uterine segment are innervated by parasympathetic fibres from S2 to S4, which form ganglia lateral to the cervix and enter along with the blood vessels. The fundus is innervated by sympathetic fibres from T10 to L1 via the inferior hypogastric nerve, which enters the uterus at the uterosacral ligaments, as well as via the ovarian plexuses that enter at the corna (Paul et al., 1999). By anaesthetizing the nerve plexuses that lie adjacent to the cervix, PCB reduces pain induced by cervical manipulation and dilatation. However, it has less effect on the cramping pain from the fundus of the uterus. Therefore, PCB alone is not effective for pain relief after cervical priming with misoprostol and use of intravenous sedation is needed regardless of whether the local anaesthetic was injected into the cervix or vaginal vault (Kan et al., 2004). Research into other methods of pain relief and conscious sedation may be helpful to further enhance patient satisfaction.

Midtrimestrer induced abortion and prior CS

There have been increasing CS rates over the past few years, more so because of the increasing use of prenatal diagnosis, one child norm, HIV and improved fetal survivability. The present CS delivery rate is 24.4% in USA, and it is generally increasing worldwide. CS delivery is now the most frequent major surgical procedure performed in the field of obstetrics and gynaecology. Although many studies have demonstrated the small risk of complications for vaginal birth at term after a prior CS (Miller et al., 1994) and the safety of early medical abortion (Gao and Wang, 1999; Xu et al., 2001; Gautam and Agrawal, 2003), the experience of a mid-trimester pregnancy termination in women with prior uterine scar is more limited (Rosen et al., 1991) (Table III). Uterine rupture, haemorrhage and hysterotomy/hysterectomy remain uncommon and inevitable complications of any termination method used in second-trimester pregnancy. As per the literature review, it seems that uterine rupture is associated with the use of intravenous high-dose oxytocin (Atienza et al., 1980). Three small case series of 87 women with at least one CS undergoing second-trimester pregnancy termination with misoprostol reported no case of uterine rupture (Herabutya et al., 2003; Pongsatha and Tongsong, 2003; Rouzi, 2003) (Table III). The retrospective case series of 606 cases using PGE₂, concentrated oxytocin and dilute oxytocin reported an increased incidence of uterine rupture and need for blood transfusion in women with prior CS delivery (Chapman et al., 1996) (Table III). There was no increased risk of complications when 101 patients with a prior CS were compared with 619 women with an unscarred uterus (Dickinson, 2005). In a cohort study, similar result of blood loss >500 ml was seen as reported in the Chapman study (Herabutya et al., 2003). Among 23 women with a history of CS treated with the combination of mifepristone and gemeprost, one case of asymptomatic uterine rupture was reported (Boulot et al., 1993). In addition, there are case reports of uterine rupture with the use of misoprostol in the scarred uterus (Chen et al., 1999; Berghahn et al., 2001). It could be speculated that, with the combination of mifepristone and misoprostol, the incidence of uterine rupture should be lower because of the advantage of cervical dilatation facilitated by mifepristone even before the uterus begins to contract to expel its contents. More studies are needed to evaluate the optimal mifepristone and misoprostol combination in women with prior CS.

View this table: [in this window] [in a new window]   Table III. Mid-trimester induced abortion and prior Caesarean section

 

	Conclusions

TOP Abstract Introduction Materials and methods Conclusions References

 
Treatment with mifepristone–misoprostol is a safe and an effective method for mid-trimester medical abortion. In most patients, routine surgical evacuation of the placenta is not necessary. More studies are required to compare the medical methods to surgical ones. Patients with prior CS undergoing medical abortion need to be managed more carefully for the early detection of impending complications. Further studies are needed to confirm the safety of medical regimens in these patients.

Women should be offered analgesics whenever required. Studies should focus on improving pain management. Medical abortion is advantageous with regard to evaluation of the fetus and placenta in cases of fetal malformation. This would help in future research practice for betterment of complicated pregnancies. Further studies are also needed on the treatment of women with failed medical abortion after 24 h.

Current recommendations

1. Cervical priming is mandatory before mid-trimester surgical abortion.
   
2. Misoprostol 400 µg (2x200 µg tablets) vaginally 3 h before surgery.
   
3. Gemeprost 1 mg vaginally 3 h before surgery.
   
4. Mifepristone 200 mg orally 12–48 h before surgery.
   
5. The treatment may need to be repeated or the interval may be increased in late second-trimester abortions.
   

Mid-trimester surgical abortion

1. VA can be carried out up to 15 weeks gestation preceded by cervical priming.
   
2. D&E can be used by trained and skilled providers with sufficient experience.
   

Mid-trimester medical abortion

1. Mifepristone 200 mg orally, followed 24–48 h later by misoprostol 800 µg vaginally and thereafter by repeated doses of misoprostol 400 µg orally, every 3 h, to a maximum of four oral doses.
   
2. Mifepristone 200 mg orally, followed 24–48 h later by gemeprost 1 mg vaginally every 6 h to a maximum of five pessaries.
   
3. Misoprostol only regimens (in countries where mifepristone is not available); vaginal misoprostol 800 µg every 12 h.
   
4. Mid-trimester TOP in prior CS patients should be carried out with caution.
   
5. Routine surgical evacuation of the uterus following medical abortion is not required.
   
6. Analgesics should be offered to all women when required.
   
7. Vaginal misoprostol or gemeprost can be administered either by the woman or by a clinician according to the preference of the woman.
   

	Acknowledgements

TOP Abstract Introduction Materials and methods Conclusions References

 
We are grateful to Anna Hildenbrand for valuable help with language checking. The studies from the Karolinska University Hospital referred to in the article were supported by UNDP/UNFPA/WHO/World Bank Special Programme of Research Development and Research Training in Human Reproduction, World Health Organization, Geneva, Switzerland and the Swedish Medical Research Council (number 05696, 05170, 0855, 2003-3869).

	References

TOP Abstract Introduction Materials and methods Conclusions References

 

Anonymous (1995) Abortion in law, history and religion. Childbirth by Choice. Toronto, ON.

Alan Guttmacher Institute (1999) Sharing Responsibility: Women, Society and Abortion Worldwide. The Alan Guttmacher Institute, New York and Washington DC.

Andersen LF, Poulsen HK, Sorensen SS, Christensen BM, Sponland G and Skjeldestad FE (1989) Termination of second trimester pregnancy with gemeprost vaginal pessaries and intra-amniotic PGF2 alpha. A comparative study. Eur J Obstet Gynecol Reprod Biol 31,1–7.[CrossRef][ISI][Medline]

Armatage RJ and Luckas MJ (1996) A randomized trial of 2 regimens for the administration of vaginal prostaglandins (gemeprost) for the induction of midtrimester abortion. Aust N Z J Obstet Gynaecol 36,296–299.[ISI][Medline]

Aronsson A, Bygdeman M and Gemzell-Danielsson K (2004) Effects of misoprostol on uterine contractility following different routes of administration. Hum Reprod 19,81–84.[Abstract/Free Full Text]

Ashok PW and Templeton AA (1999) Non Surgical mid-trimester termination of pregnancy: a review of 500 consecutive cases. Br J Obstet Gynaecol 106,706–710.[ISI][Medline]

Ashok PW, Flett GM and Templeton A (1998) Termination of pregnancy at 9–13 weeks’ amenorrhoea with mifepristone and misoprostol. Lancet 15,542–543.[CrossRef]

Ashok PW, Kidd A, Flett GM, Fitzmaurice A, Graham W and Templeton A (2002) A randomized comparison of medical abortion and surgical vacuum aspiration at 10–13 weeks gestation. Hum Reprod 17,92–98.[Abstract/Free Full Text]

Ashok PW, Templeton A, Wagaarachchi PT and Flett GM (2004) Midtrimester medical termination of pregnancy: a review of 1002 consecutive cases. Contraception 69,51–58.[CrossRef][ISI][Medline]

Ashok PW, Hamoda H, Flett GM, Kidd A, Fitzmaurice A and Templeton A (2005a) Patient preference in a randomized study comparing medical and surgical abortion at 10–13 weeks gestation. Contraception 71,143–148.[CrossRef][ISI][Medline]

Ashok PW, Hamoda H, Flett GM, Kidd A, Fitzmaurice A and Templeton A (2005b) Psychological sequelae of medical and surgical abortion at 10–13 weeks gestation. Acta Obstet Gynecol Scand 84,761–766.[CrossRef][ISI][Medline]

Atienza MF, Burkman RT and King TM (1980) Midtrimester abortion induced by hyperosmolar urea and prostaglandin F2 alpha in patients with previous cesarean section: clinical course and potential for uterine rupture. Am J Obstet Gynecol 138,55–59.[ISI][Medline]

Autry AM, Hayes EC, Jacobson GF and Kirby RS (2002) A comparison of medical induction and dilation and evacuation for second-trimester abortion. Am J Obstet Gynecol 187,393–397.[CrossRef][ISI][Medline]

Bartley J and Baird DT (2002) A randomised study of misoprostol and gemeprost in combination with mifepristone for induction of abortion in the second trimester of pregnancy. BJOG 109,1290–1294.[CrossRef][ISI][Medline]

Bartley J, Brown A, Elton R and Baird DT (2001) Double-blind randomized trial of mifepristone in combination with vaginal gemeprost or misoprostol for induction of abortion up to 63 days gestation. Hum Reprod 16,2098–2102.[Abstract/Free Full Text]

Bebbington MW, Kent N, Lim K, Gagnon A, Delisle MF, Tessier F and Wilson RD (2002) A randomized controlled trial comparing two protocols for the use of misoprostol in midtrimester pregnancy termination. Am J Obstet Gynecol 187,853–857.[CrossRef][ISI][Medline]

Belanger E, Melzack R and Lauzon P (1989) Pain of first-trimester abortion: a study of psychosocial and medical predictors. Pain 36,339–350.[CrossRef][ISI][Medline]

Berer M (2004) National laws and unsafe abortion: the parameters of change. Reprod Health Matters 12,1–8.[CrossRef][ISI][Medline]

Berghahn L, Christensen D and Droste S (2001) Uterine rupture during second-trimester abortion associated with misoprostol. Obstet Gynecol 98,976–977.[Abstract/Free Full Text]

Bhathena RK, Sheriar NK, Walvekar VR and Guillebaud J (1990) Second trimester pregnancy termination using extra-amniotic ethacridine lactate. Br J Obstet Gynaecol 97,1026–1029.[ISI][Medline]

Bhide A, Sairam S, Hollis B and Thilaganathan B (2002) Comparison of feticide carried out by cordocentesis versus cardiac puncture. Ultrasound Obstet Gynecol 20,230–232.[CrossRef]