The reasons for termination of pregnancy in the third trimester
Correspondence: Dr M. Dommergues, AP-HP Service de Gynécologie-Obstétrique, Hôpital A. Béclère, 157 Rue de la Porte de Trivaux, 92140 Clamart Cedex, France.
Objective To analyse the process in making decisions leading to termination of pregnancy in the third trimester and to evaluate the maternal morbidity associated with this procedure.
Design Retrospective study.
Setting The Maternité Port Royal University Hospital, Paris, France.
Population A consecutive series of 956 terminations of pregnancy performed for fetal anomalies in singleton pregnancies, 305 of which were in the third trimester and 651 in the second.
Main outcome measures Indications for termination of pregnancy; process leading to late termination of pregnancy; maternal morbidity.
Results One hundred and thirteen (37%) third trimester terminations of pregnancy were associated with false negative resulted from the results of earlier screening tests. In 15 terminations (15%), the decision was postponed, although the poor fetal prognosis was established earlier. In 55 (18%) the diagnosis was not possible earlier than the third trimester, and in 122 (40%) the diagnosis was possible earlier but the poor prognosis for the fetus was not established until the third trimester. Maternal morbidity due to termination of pregnancy was similar in the second and third trimester.
Conclusion One-third of late terminations of pregnancy could have been avoided by more efficient screening in the second trimester. However, because fetal prognosis is not always clear when a malformation is diagnosed, postponing the decision until fetal development allows more thorough evaluation and may avoid unnecessary termination of pregnancy in the second trimester. This could be the main beneficial aspect of not setting a limit to the gestational age for performing termination of pregnancy. To analyse the process in making decisions leading to termination of pregnancy in the third trimester and to evaluate the maternal morbidity associated with this procedure. The Maternit6 Port Royal University Hospital, Paris, France.
Termination of pregnancy in the third trimester is legal in a growing number of European countries, when ‘there is a substantial risk that if the child were born, he would suffer from such mental or physical abnormalities as to be seriously handicapped’1. However, because the aim of fetal medicine is primarily to improve perinatal care for sick children, the legalisation of termination of pregnancy in the third trimester may appear contradictory or even unacceptable, especially since modem neonatal care has reduced the gestational age of viability. The legalisation of late termination of pregnancy does not imply that it is always morally acceptable2 or clinically justifiable, which emphasises the need for accurate appraisal of these cases. We therefore performed a retrospective survey of a consecutive series of 305 terminations of pregnancy in the third trimester and compared these with 651 second trimester terminations of pregnancy performed in the same centre (Maternité Port Royal, Paris). The decisions leading to the terminations of pregnancy were analysed, as well as major maternal complications.
The decision to perform termination of pregnancy in the third trimester may result from the diagnosis of a severe fetal anomaly for the first time in the third trimester, due either to failure of an earlier screening test or to the late onset of the disease itself. Termination of pregnancy may also be performed late in gestation because the prognosis was still uncertain earlier, when the diagnosis of fetal anomaly was made. In such a situation, the possibility of performing a late termination may have a positive effect in preventing a hasty termination of pregnancy in the second trimester, when the prognosis for the infant may be uncertain. The fact that termination will remain an option if ominous prognostic factors are identified later in pregnancy may encourage some woman to accept a conservative policy, and thereby save those infants who have a good prognosis.
In France termination of pregnancy has been legal since 19753 regardless of gestational age, provided ‘there is a high probability that the fetus is affected by a particularly severe condition with no effective therapy available at the time of prenatal diagnosis’. Parental request for a termination of pregnancy needs to be accepted by two experts, one of whom is appointed by the court, the other being affiliated to a fetal medicine centre4,5.
Between 1986 and 1994,1128 women with an abnormal singleton fetus elected to have their pregnancies terminated, according to French law, after being referred to the Maternité Port-Royal, Paris: 172 terminations were performed before 16 weeks, 651 at between 16 and 27 weeks, and 305 after 27 weeks. Mean gestational age at termination was 21.7 weeks in the second trimester and 32 weeks in the third trimester. The institution in which the study was conducted is a tertiary referral centre.
The size of the population it serves cannot be estimated accurately due to the liberal access of health care facilities in France. Based on the Eurocat report6 it can be estimated that 599/1387 (43%) of all terminations of pregnancies performed in Paris following prenatal diagnosis of a fetal anomaly were carried out in our centre during the period 1990–1994.
During the study period, three routine sonographic examinations (one per trimester) were generally offered routinely. However, no reliable information is available concerning the training of doctors performing fetal ultrasound.
Maternal serum tests were not used to screen for structural fetal malformations, and were rarely used to screen for aneuploidy during the study period (≥ 10% of the Paris obstetric population underwent maternal serum screening in 1990–1994).
Techniques of termination of pregnancy
During the second and third trimester labour was induced by oxytocin (18 cases) or prostaglandins using vaginal pessaries (250 cases) of gemeprost (Bellon Laboratories, Rhône-Poulenc Rorer, Neuilly-sur-Seine, France), intramuscular (222 cases) or intravenous (335 cases) sulpostrone (Schering Laboratories, Lys les Lannoy, France). A combination of gemeprost and sulprostone was used in 92 cases. Hysterotomy was required in 14 cases (Table 1).
Table 1. Techniques of termination of pregnancy and trimester of pregnancy. Values are given as n.
|Sulprostone and Gemeprost†||65||27||92|
|Dilatation and curettage||16||0||16|
In 490 women cervical ripening was achieved by oral administration of mifepristone (Mifegyne*, Laboratories Roussel-Uclaf, 92800 Puteaux, France) 48 hours before inducing uterine contractions with gemeprost or sulprostone. When the fetus was considered potentially viable, intravascular administration of diazepam and potassium chloride was performed prenatally.
Epidural analgesia in labour was administered to 490 women and general anaesthesia was required at delivery in 578 women. Postmortem examination of the fetus was carried out in all cases.
Analysis of results
The occurrence of major maternal complications was recorded in each case, including uterine rupture, uterine hypotonia with blood loss requiring transfusion, cervical tears requiring surgical repair, and endometritis. Relatively minor side effects related to prostaglandin administration such as nausea, vomiting, rash, diarrhoea and moderate fever without documented infection were not taken into account in the analysis.
The diagnostic process leading to the termination of pregnancy was analysed according to the indication for prenatal diagnosis and the nature of the fetal anomaly.
The clinical circumstances leading to prenatal diagnosis were classified as follows:
- 1High risk of recurrence of a genetic disease in a family with a previously affected child.
- 2Screening for chromosomal anomalies based on maternal age or maternal serum markers.
- 3Prenatal diagnosis resulting from routine sonograms performed in a low risk woman.
- 4Obstetric complications such as extreme intrauterine growth retardation or early premature rupture of membranes.
- 5Maternal serum screening for toxoplasmosis or rubella.
The nature of the fetal abnormality was classified as genetic (inherited errors of metabolism and chromosomal anomalies), single malformations with normal chromosomes (abnormality of the kidneys and genitourinary tract, musculo-skeletal anomalies, congenital heart diseases, neurological anomalies), multiple malformations, and obstetric complications such as extreme intrauterine growth retardation.
The reasons why some terminations were not performed earlier than the third trimester were categorised retrospectively as:
- 1False negative of earlier screening, i.e. cases which were potentially identifiable early in gestation.
- 2Conditions in which the prognosis of the fetal anomaly could not be predicted accurately until the third trimester, even though prenatal diagnosis was feasible earlier.
- 3Anomalies with late onset or anomalies not likely to be amenable to prenatal diagnosis until the third trimester.
- 4Cases in which the decision to perform termination of pregnancy was not taken until the third trimester, although a poor fetal prognosis had been documented earlier during the second trimester.
The differences between groups were tested using the χ2 or Fisher's exact test where appropriate (Stat View 4–02).
In 93% of women undergoing termination of pregnancy in the third trimester the diagnostic process was based on routine ultrasound screening where there was no known genetic risk. In contrast, risk factors for a genetic anomaly had been identified early in gestation in 37% of pregnancies in the second trimester (Table 2).
Table 2. Indications for prenatal diagnosis. Values are given as n (%).
|High genetic risk*||115 (18)||0 (0)|
|Aneuploidy screening based on maternal age or serum markers||123 (19)||0 (0)|
|Routine sonographic screening**||346 (53)||285 (93)|
|Obstetrical complications†||57 (9)||10 (3)|
|Maternal screening for infectious disease‡||10 (2)||10 (3)|
The majority of third trimester terminations were performed because of neurological anomalies, multiple malformations with a normal karyotype, and chromosomal anomalies diagnosed after an abnormal routine ultrasound scan. The main indications for termination of pregnancy in the second trimester were aneuploidies, inherited errors of metabolism, multiple malformations, and neurological anomalies (Tables 3 and 4).
Table 3. Abnormalities resulting in termination of pregnancy. Values are given as n (%).
|Neurological||92 (14)||98 (32)|
|Multiple malformations||91 (14)||69 (23)|
|Chromosomal||190 (29)||54 (18)|
|Skeletal||29 (4)||27 (9)|
|Kidneys/urinary tract||56 (9)||27 (9)|
|Congenital heart disease||21 (3)||19 (6)|
|Extreme intrauterine growth restriction*||12 (2)||11 (4)|
|Early second trimester prelabour rupture of the membranes**||45 (7)||0 (0)|
|Genetic disease†||115 (18)||0 (0)|
Table 4. Indications for termination of pregnancy in the third trimester.
|Neurological anomaly (n= 98)||Multiple malformations with a normal karyotype (n= 34)|
| ||Myelomeningocele (n= 34)|
| ||Hydrocephalus (n= 25)|
| ||Complex cerebral anomaly (n= 14)|
| ||Agenesis of corpus callosum + hydrocephalus (n= 9)|
| ||Anencephaly (n= 5)|
| ||Holoprosencephaly (n= 4)|
| ||Microcephaly (n= 4)|
| ||Encephalocele (n= 3)|
|Multiple malformations with a normal karyotype (n= 69)||Unexplained multiple malformations (n= 43)|
| ||Recognised syndromes (n= 9)|
| ||Intractable and unexplained hydrops (n= 7)|
| ||Cytomegalovirus (n= 3)|
| ||Toxoplasmosis (n= 6)|
| ||Rubella (n= 1)|
|Chromosomal anomaly (n= 54)*||Trisomy 21 (n= 26)|
| ||Trisomy 18 (n= 10)|
| ||Structural anomaly (n= 11)**|
| ||Trisomy 13 (n= 5)|
| ||Triploidy (n= 1)|
| ||Klinefelter associated with lung malformation (n= 1)|
|Skeletal anomaly (n= 27)||Lethal osteochondrodysplasia (n= 15)|
| ||Other osteochondrodysplasia (n= 7)|
| ||Multiple limb anomalies, phocomelia, limb reduction (n= 5)|
|Kidney/urinary tract (n= 27)||Terminal fetal renal failure diagnosed based on severe oligohydramnios (n= 25)|
| || Bilateral nephropathy (n= 13)†|
| || Bilateral renal agenesis (n= 7)|
| || Bilateral uropathy (n= 5)‡|
| ||Fetal renal insufficiency diagnosed by fetal urine analysis (n= 2)§|
|Congenital heart disease (n= 19)||Complex structural heart defects (n= 9)|
| ||Hypoplastic left heart (n= 7)|
| ||Cardiomyopathy (n= 3)¶|
|Extreme intrauterine growth restriction||n= 11|
The process leading to termination of pregnancy in the third trimester was analysed in more detail (Table 5). In 122 terminations (40%), we estimated that the prognosis of the anomaly could not have been established firmly until the third trimester, although prenatal diagnosis of the anomaly was possible earlier. The majority of these cases comprised cerebral ventriculomegalies, in which the poor prognosis was not established until the third trimester, when associated cerebral anomalies or rapid enlargement of the ventriculomegaly were identified. Similarly, the poor prognosis of some multiple malformation syndromes could be established only after additional anomalies were recognised by ultrasound in the third trimester. In fetuses with urinary anomalies the diagnosis of renal failure was made late in pregnancy because of the delayed onset of severe oligohydramnios.
Table 5. Diagnostic process leading to third trimester termination.
|Fetal prognosis could not be evaluated until the third trimester, although second trimester diagnosis possible (n= 122)|
| Congenital heart defect (n= 3)|
| Persistent large cystic hygroma (n= 1)|
| Toxoplasmosis: cerebral lesions occurring despite fetal therapy (n= 6)|
| Lethal osteochondrodysplasia (n= 6)|
| Campomelic dwarfism (n= 1)|
| Osteogenesis imperfecta (n= 5)|
| Severe intrauterine growth restriction (n= 11)|
| Multiple malformation syndromes in which additional anomalies were found by ultrasound scan at follow-up (n= 28)|
| Nephropathy or uropathy with fetal renal failure (n= 20)|
| Cerebral ventriculomegaly, either major or associated with complex cerebral anomalies (n= 43)|
| Microcephaly (n= 4)|
|Second trimester diagnosis probably not feasible due to the apparently late onset of the condition (n= 55)|
| None immune hydrops (n= 8)|
| Aneuploidy diagnosed because of abnormal third trimester ultrasound such as intrauterine growth restriction, chylothorax, bowel dilatation or associations of more subtle anomalies difficult to diagnose earlier short femur, pyelectasis, facial dysmorphology (n= 30)|
| Trisomy 21 (n= 20)|
| Structural anomalies (n= 9)|
| Triploidy (n= 1)|
| Congenital heart anomalies (n= 5)|
| 3 cardiomyopathies, and 2 Ebstein's anomalies leading to major myocardial dysfunction and cardiac insufficiency in the third trimester. These cases had a normal second trimester ultrasound|
| Nonlethal osteochondrodysplasia (n= 7)|
| Cytomegalovirus infections revealed by third trimester intrauterine growth restriction and structural anomalies (n= 3)|
| Cystic fibrosis revealed by third trimester hyperechoic bowel (n= 1)|
| Rapidly growing cerebral tumour with normal second trimester ultrasound (n= 1)|
|Diagnosis and prognostic evaluation completed during second trimester, but decision to terminate the pregnancy taken later in gestation (n= 15)|
| Chromosomal anomaly (n= 9)|
| Trisomy 21 (n= 6)|
| Structural anomalies (n= 2)|
| Klinefelter syndrome associated with congenital adenomatoid malformation of the lung (n= 1)|
| Multiple malformations with normal karyotype (n= 1)|
| Myelomeningocele (n= 2)|
| Other cerebral anomalies (n= 2)|
| Hypoplastic left heart (n= 1)|
|Third trimester prenatal diagnosis of conditions potentially identifiable earlier (n= 113)|
| Chromosomal anomalies with gross malformations|
| Trisomy 13 (n= 5)|
| Trisomy 18 (n= 10)|
| Gross congenital heart defects (n= 10)|
| Gross musculoskeletal anomalies (n= 14)|
| Gross multiple malformations with normal karyotype (n= 21)|
| Bilateral renal agenesis (n= 7)|
| Myelomeningocele (n= 32)|
| Anencephaly (n= 5)|
| Other gross cerebral anomalies (n= 9)|
In 55 cases (18%), we believe the anomaly was not likely to have been diagnosed earlier than the third trimester ultrasound examination which is routinely performed in France. This group includes cases of nonimmune hydrops occurring in the third trimester; Down's syndrome revealed by sonographic anomalies such as fetal chylothorax or bowel obstruction, or the association of minor signs such as pyelectasis, short femur and facial dysmorphology; structural chromosomal anomalies revealed by intrauterine growth restriction associated with minor anatomical anomalies; non-lethal osteochondrodysplasias such as achondroplasia; structural abnormalities caused by cytomegalovirus infection; and other malformations (Table 5) which we believe could not be diagnosed by ultrasound performed in the second trimester.
In 15 (5%) the poor prognosis of the fetus was established in the second trimester but the termination of pregnancy was postponed, either by the physician in charge or because the couple required a longer time to come to a decision concerning termination.
In 113 (37%) the condition for which termination of pregnancy was performed in the third trimester could have been identified earlier in pregnancy. These cases are due to failure of screening by ultrasound in the second trimester. Myelomeningocele, trisomy 18 and 13, and lethal dwarfism accounted for the majority of these potentially avoidable third trimester terminations.
As for maternal morbidity, there were no significant differences between the women undergoing termination of pregnancy in the second and third trimesters (Table 6). The incidence of scarred uterus was similar in the second trimester group (n= 75, 12%), and in the third trimester group (n= 28, 9%). The incidence of placenta praevia was slightly, but not significantly, higher in the second trimester group (n= 6, 0.9%) than in the third trimester group (n= 1, 0.3%). The overall incidence of uterine rupture was 3/853 (0.35%) in women with an unscarred uterus compared with 3/103 (3%) in women with a scarred uterus (χ2 test with Yate's correction: 5.99; P= 0.01), suggesting that a uterine scar was the major risk factor for subsequent uterine rupture.
Table 6. Maternal morbidity. Values are given as n (%).
|Uterine rupture||3 (0.5)||3 (1)|
|Cervical tear||7 (1)||3 (1)|
|Endometritis||5 (0.8)||3 (1)|
|Uterine hypotonia requiring blood transfusion||6 (0.9)||2 (0.6)|
|None of the above||630 (96.8)||294 (96.4)|
The aim of fetal medicine is to assist the treatment of babies affected by disorders diagnosed in pregnancy. However the prognosis of some conditions is so poor that termination of pregnancy may be considered in countries where it is legally available. We are aware that termination of a pregnancy in which the fetus is severely compromised raises many ethical problems especially in the third trimester7–12.
Answering these moral questions from a philosophical point of view is beyond the scope of this paper. However, since third trimester termination of pregnancy is becoming legal in a growing number of European countries1,2,13, we felt it could be useful to analyse the processes leading to the decision to perform termination of pregnancy in the third trimester and the morbidity associated with this procedure. Most terminations were carried out on women at low risk for fetal abnormalites. The anomalies for which termination of pregnancy was indicated in the third trimester did not differ fundamentally from those of the second trimester. However, chromosomal anomalies diagnosed because of maternal age or serum markers, genetic disorders secondary to the family history, and early prelabour rupture of the membranes, were indications specific to second trimester terminations of pregnancy. The majority of anomalies involved in third trimester terminations of pregnancy were neurological defects, multiple malformation syndromes, and chromosomal anomalies suggested by an abnormal ultrasound scan.
Why were so many terminations performed so late in pregnancy? Our classification of the process leading to termination of pregnancy may appear arbitrary; nevertheless, in most cases, we could easily agree on the reason why terminations were performed in the third trimester. A large number of terminations of pregnancy in the third trimester were due to failure of ultrasound screening in the second trimester (large myelomeningoceles, multiple malformations with several gross structural defects, major heart defects with a severely abnormal four chamber view, renal agenesis). Improving the efficiency of ultrasound screening in the second trimester may reduce the number of terminations carried out later in pregnancy. A greater number of women accepting serum screening for aneuploidy may also contribute to a reduction in late terminations.
However, in more than half of the cases in our study, the decision to terminate the pregnancy was not feasible before the third trimester since diagnosis of the anomaly was not possible, or it was not possible to judge the prognosis for the fetus, before the third trimester. In some cases prenatal diagnosis cannot be done earlier, due to the natural history of the disease itself these include hydrops fœtalis, chromosomal anomalies revealed by intestinal obstruction, severe embryopathy due to cytomegalovirus infection, cardiomyopathy, and rapidly growing hydrocephalus with a normal ultrasound examination in the second trimester.
In other cases, diagnosis is possible in the second trimester, but third trimester follow up is needed to establish the prognosis. Moderate dilatation of the cerebral ventricles in the second trimester may be a benign disorder, or it may be secondary to complex cerebral malformations which will not be diagnosed until the third trimester. The hopeless prognosis of some cases of obstructive uropathy and severe intrauterine growth restriction may not be apparent until the third trimester. Toxoplasmosis is a relatively common condition in France. Anti-parasitic therapy is usually effective in preventing cerebral damage to the fetus, but failure of treatment and damage to the fetus can occur, and can only be diagnosed in the third trimester. In all these situations, the possibility of termination of pregnancy in the third trimester is beneficial, since it allows a more accurate assessment of prognosis to be made, and may prevent unnecessary terminations of pregnancy in the second trimester.
A few terminations of pregnancy are performed in the third trimester even though poor fetal prognosis has been established earlier. This may be due to failure to refer the woman early enough to a fetal medicine centre, or to inform the parents of the possibility of having a termination performed, or to women who need a long time before coming to a decision regarding termination of pregnancy.
In conclusion, we found that in one third of the women undergoing termination of pregnancy in the third trimester, the diagnosis of severe fetal anomaly could have been made earlier in pregnancy allowing a termination of pregnancy in the second trimester. In over half the cases the fetal anomaly could only be diagnosed in the third trimester, or an accurate prognosis for the fetus could only be made in the third trimester. Termination of pregnancy in the third trimester was not associated with a greater risk of serious maternal complications than termination of pregnancy in the second trimester.
The authors would like to thank Mrs G. Patin whose help was crucial for adequately handling the records of all patients.