Risk of spontaneous abortion in singleton and twin pregnancies - PDF by via28446


									Human Reproduction Vol.18, No.8 pp. 1720±1723, 2003                                                          DOI: 10.1093/humrep/deg308

Risk of spontaneous abortion in singleton and twin
pregnancies after IVF/ICSI

Philippe Tummers, Paul De Sutter and Marc Dhont1
Infertility Centre, University Hospital Ghent, De Pintelaan 185, B-9000 Gent, Belgium
    To whom correspondence should be addressed. E-mail: marc.dhont@UGent.be

BACKGROUND: The risk of spontaneous ®rst trimester abortion is estimated to be between 10 and 20%. Although
it is common knowledge that the incidence of abortion decreases as pregnancy progresses, exact data in relation to
the duration of pregnancy are scarce. METHODS: We reviewed 1597 clinical IVF/ICSI pregnancies with known
outcome and tabulated the number of miscarriages or fetal demise per intervals of 2 weeks. We furthermore com-

                                                                                                                                           Downloaded from humrep.oxfordjournals.org by guest on September 8, 2010
pared the outcome in terms of fetal survival of 1200 singleton pregnancies with that of 397 twin pregnancies.
RESULTS: The overall incidence of non-ongoing singleton pregnancies was 21.7%. Fetal death, after positive heart
activity had been recorded, occurred in 12.2% of singleton pregnancies. The overall incidence of spontaneous abor-
tion in twin pregnancies was 17.1% (12.1% vanishing twins and 5.0% complete miscarriages). The incidence of mis-
carriage in the twin pregnancies, expressed per gestational sac, was 11.1%. Once fetal heart activity was present,
the risk of abortion (per gestational sac) was 7.3%, which is signi®cantly lower than that in singleton pregnancies.
CONCLUSIONS: Our data give an estimate of the probability of miscarriage or fetal demise at any given period of
the ®rst trimester both for singleton and twin pregnancies. Twin pregnancies after IVF have a better potential for
survival than singleton pregnancies.

Key words: abortion/ICSI/IVFI/singleton/twin

Introduction                                                         hybridization (FISH) technique have shown that the incidence
The incidence of ®rst trimester abortion is estimated to be          of aneuploidy in human oocytes increases with age, rising in
between 10 and 20%. The true incidence, however, is not well         women aged >35 years and even more so >40 years (Abdalla
known because many abortions occur before pregnancy is               et al., 1993; Fretts et al., 1995).
clinically recognized. Data from spontaneous pregnancies are            Stoeckel (1945) was the ®rst to suggest that twins are more
dif®cult to ascertain because this would necessitate the             often conceived than born. With the advent of ultrasound, ®rst
systematic follow-up of a large consecutive cohort of women          abdominal and, in the late 1980s, vaginal ultrasound, more
by determination of hCG and by ultrasound. In contrast, after        detailed information regarding early resorption in multiple
infertility treatment and particularly following IVF treatment,      gestations became available. Still the true incidence of
the diagnosis of pregnancy is made at an early stage and most        vanishing twins is dif®cult to assess. Landy and Keith (1998)
women are motivated to have their pregnancy followed up very         reviewed the majority of pertinent studies published since
closely.                                                             1990. Most of these studies describe pregnancies conceived as
   There are no solid data to compare the incidence of               a result of assisted reproductive techniques. Using these data it
miscarriage in spontaneous versus IVF pregnancies but it is          was estimated that ~30% of these twins will ultimately result in
generally accepted that the incidence is slightly higher after       singletons and <10% will end in a complete abortion. When
IVF (Ezra and Schenker, 1995; Simon et al., 1999). The main          monochorionic twins were compared with dichorionic twin
reason for a higher incidence is the age of the patients, which      pregnancies (Sebire et al., 1997), the rate of fetal loss was
on average is 3±5 years higher than that of a fertile population     signi®cantly higher in the former. This makes it dif®cult to
at the time of a ®rst pregnancy. Indeed, studies on the risk of      extrapolate the ®ndings in assisted reproductive treatment twin
spontaneous miscarriage indicate that maternal age is an             pregnancies, where monozygotic twins are rare, to spontaneous
important risk factor (Andersen et al., 2000). There is clear        twin pregnancies.
evidence from oocyte donation programmes that this risk is              In the literature, little information exists on the risk of
associated with the ageing of the oocytes, rather than that of the   miscarriage in relation to gestational age and the presence of
uterus (Navot et al., 1994; Abdalla et al., 1997). Moreover,         fetal heart activity (Hill et al., 1991; Frates et al., 1993;
studies on oocytes and embryos using the ¯uorescent in-situ          Goldstein, 1994). We therefore analysed the outcome of 1200
1720                                                                             ã European Society of Human Reproduction and Embryology
                                                                                            Spontaneous abortion in singleton/twin pregnancies after art

Table I. Risk of spontaneous abortion per fetal sac in relation to gestational
age (comparison of twin pregnancies versus singleton pregnancies)
                                             Singleton      Twin       P

Total risk of spontaneous abortion (%)       21.10          11.10      < 0.01
From +ve heart activity onwards (%)          12.20           7.30      < 0.01
After 7 weeks (%)                            11.90           7.30      < 0.01
After 9 weeks (%)                             8.20           4.90      < 0.05
After 11 weeks (%)                            4.20           2.20      NS
After 13 weeks (%)                            2.20           2.00      NS

NS = not signi®cant.

singleton and 397 twin IVF pregnancies, which were all
followed up by transvaginal ultrasound at regular intervals
throughout the ®rst trimester. Although the absolute ®gures
may not be representative for the abortion rate in spontaneous
pregnancies, we may assume that the relative risk of miscar-
riage in relation to the duration of pregnancy and the detection
of fetal heart activity we obtained from our study in IVF

                                                                                                                                                           Downloaded from humrep.oxfordjournals.org by guest on September 8, 2010
patients can be extrapolated to spontaneous pregnancies.                         Figure 1. Risk of spontaneous abortion per fetal sac in relation to
   Also, no large scale studies exist which compare the risk of                  gestational age (comparison of twin pregnancies versus singleton
                                                                                 pregnancies). +FHR = positive fetal heart rate.
spontaneous abortion in singleton pregnancies with that in twin
pregnancies, taking into account the fate of each gestational
sac.                                                                             Results
   We therefore compared the outcome of singleton pregnan-
cies with twin pregnancies for each gestational sac separately.                  Of the 1200 singleton pregnancies, 938 (78.2%) were ongoing,
                                                                                 yielding an overall incidence of spontaneous abortion of 21.8%
                                                                                 (262 spontaneous abortions). After fetal heart activity was
                                                                                 detected on ultrasound, the risk of abortion declined to 12.2%
Materials and methods
                                                                                 (CI: 10.2±14.5%). At 7 weeks this risk decreased to 11.9% (CI:
Patients                                                                         9.8±14.1%), at 9 weeks to 8.2% (CI: 6.7±10.5%), at 11 weeks it
Records of all IVF/ICSI patients treated in our centre between 1993              was 4.2% (CI: 2.9±5.7%) and at 13 weeks the risk of
and 2000 were reviewed for inclusion in the present study. We                    miscarrying had dropped to 2.2% (CI: 1.3±3.4%) (Table I,
included only those patients who were followed in our centre until 12            Figure 1).
weeks of gestation and of whom we had reliable information on the
                                                                                    Of the 397 twin pregnancies, 329 (82.8%) were ongoing, 48
outcome of pregnancy.
   Out of 2778 pregnancies that were obtained in our infertility
                                                                                 (12.1%) ended with a partial miscarriage (vanishing twins) and
department between 1993 and 2000, 1778 (64.0%) were followed up                  20 (5.1%) ended with a complete miscarriage. The risk of
at 2-weekly intervals closely in our centre and had reliable information         waste of at least one gestational sac in twin pregnancies
on the outcome of pregnancy. These pregnancies were analysed; 103                therefore was 17.2%. When the incidence of miscarriage was
(5.8%) were biochemical, 39 (2.2%) were ectopic and 1636 (92.0%)                 calculated for each gestational sac separately, the abortion risk
were clinical. Of the 1636 clinical pregnancies, 1200 were singletons,           was 11.1% per sac. We also calculated the abortion rate (per
397 were twins and 39 were triplets. Only the singletons and twins               gestational sac separately) in relation to the duration of
were further studied.                                                            pregnancy. Once fetal heart activity was positive, the risk of
   The indications for infertility as well as the number of embryos
                                                                                 abortion was 7.3% (CI: 5.2±9.8%). At 9 weeks gestational age
transferred were not signi®cantly different between the group with
singletons and the group with twins.
                                                                                 the abortion risk had declined to 4.9% (CI: 3.2±7.2%), at 11
                                                                                 weeks this risk was 2.2% (CI: 1.1±3.9%) and at 13 weeks it was
De®nitions and analysis                                                          2.0% (CI: 0.9±3.7%) These results are shown in Table I.
A pregnancy was de®ned by the detection of a positive serum hCG                     From Figure 1 it can be seen that there is a signi®cantly
(>0.050 IU/ml) 17 days after oocyte retrieval. A biochemical                     higher risk of abortion expressed per gestational sac in
pregnancy was de®ned as a pregnancy without a intrauterine                       singleton pregnancies compared with twin pregnancies, at
gestational sac that resolved spontaneously. A spontaneous abortion              each interval of the ®rst trimester of pregnancy. This difference
was de®ned as either an empty gestational sac (blighted ovum) or fetal           remains signi®cant until 11 weeks gestational age.
demise. Abortion was registered as having occurred on the day that an
                                                                                    In the group of singleton pregnancies the mean T SD
empty gestational sac or fetal demise was recorded by transvaginal
ultrasound, irrespective of the time of expulsion or evacuation by
                                                                                 maternal age was 31.3 T 0.7 years, while in the twin group the
curettage. An ongoing pregnancy was de®ned as a delivery beyond the              mean maternal age was 30.7 T 0.6 years. To calculate the age-
25th week of pregnancy. Proportions were calculated with their 95%               corrected incidence of spontaneous abortion, we divided the
con®dence intervals (CI). The Z-test was used to compare proportions.            pregnancies into two groups according to maternal age. A ®rst
P < 0.05 was considered statistically signi®cant.                                group contained patients `35 years of age and in the second
P.Tummers, P.De Sutter and M.Dhont

Table II. Total risk of spontaneous abortion per fetal sac in relation to       duction treatment during the last 10 years mainly is the result of
maternal age                                                                    improving culture conditions. Different embryo characteristics
Age                    Singleton                Twin                   P        (fragmentation of the embryo, presence or absence of any
                                                                                irregular blastomeres, speed of cleavage, etc.) have been
`35 years (%)          19.8 (201/1015)           9.5 (67/704)          <0.001   reported to in¯uence implantation rates signi®cantly (Van
>35 years (%)          28.6 (59/185)            18.8 (17/90)           <0.02
P                      <0.0009                  <0.003                          Royen et al., 1999). Therefore, a better knowledge of embryo
                                                                                characteristics could have a large in¯uence on future embryo
                                                                                transfer policies, and not only improve pregnancy rates but also
                                                                                decrease early pregnancy loss rates.
                                                                                   Following assisted reproduction treatment, the risk of
group the maternal age was >35 years. The risk of abortion was                  miscarriage may seem a bit higher than in spontaneous
calculated for each group (Table II). The difference in                         pregnancies, but this is thought to be due to earlier pregnancy
miscarriage rate per gestational sac between singleton and                      detection on the one hand, and to older maternal age on the
twin pregnancies was signi®cant in both age categories. A                       other. Although the absolute ®gures obtained in our study
signi®cant difference was found when miscarriage of women                       cannot be transposed to spontaneous pregnancies, it may be
`35 years of age was compared with the miscarriage rate of                      assumed that the relative risk of abortion in relation to the
women >35 years. This signi®cant difference was found in both                   detection of fetal heart activity and the duration of gestation
singleton and twin pregnancy groups (Table II).                                 can be extrapolated to spontaneous pregnancies.

                                                                                                                                                     Downloaded from humrep.oxfordjournals.org by guest on September 8, 2010
                                                                                   Many authors (Abdalla et al., 1993; Fretts et al., 1995;
                                                                                Munne et al., 1995; Andersen et al., 2000) have observed an
Discussion                                                                      increased risk of fetal death, and in particular spontaneous
Our data show that the incidence of abortion in singleton IVF                   abortion, with increasing maternal age. This is also con®rmed
pregnancies drops from an overall high of 21.1±12.2% at 6                       by our own data.
weeks gestation, when fetal heart activity is recorded. From 6                     The outcome in terms of survival of twin pregnancies
weeks onwards, the incidence drops further with 2±4% at an                      appears to be better than that of singleton pregnancies.
interval of 2 weeks ending with a residual risk of fetal demise at              Although women with twin pregnancies were somewhat
13 weeks gestation of 2.2%. A similar curve was obtained in                     younger than those with a singleton pregnancy, this difference
twin pregnancies but the incidence of abortion, expressed per                   cannot be explained by differences in age. Indeed, when we
gestational sac, was signi®cantly lower than in singleton                       compare miscarriage rates between singletons and twins in
pregnancies. In both singleton and twin pregnancies, the                        different age groups, the differences between singletons and
incidence of abortion increased with age.                                       twins remained in all age groups. In assisted reproduction
   Human reproduction is a remarkably inef®cient process with                   treatment cycles, it still is a widespread habit to transfer more
a high risk of early fetal wastage. Wilcox et al. (1988)                        than one embryo, in order to maximize the chance of a
investigated the overall incidence of abortion by collecting                    pregnancy. This practice results in a high rate of multiple
daily urine specimens from 221 healthy women who were                           pregnancies. The incidence of multiple pregnancies after IVF
attempting to conceive. Urinary concentrations of hCG were                      does not, however, follow a binomial probability curve.
measured for a total of 707 menstrual cycles. Using a threshold                 Twinning rates are much higher than expected. This means
level of 2.5 IU hCG per litre of urine on 3 consecutive days,                   that embryos do not implant in an independent way, which is
they found that 22% of pregnancies ended before they could be                   also demonstrated by the results obtained with single embryo
detected by ultrasound. The overall rate of pregnancy loss,                     transfer (SET). However, all pregnancies in the present study
including clinically recognized spontaneous abortions, was                      are the result of a strategy of double embryo transfer and no
31% (Wilcox et al., 1988).                                                      elective SET pregnancies were included in our data (we only
   Numerous causes for this low rate of viable conceptuses can                  introduced SET in our centre in 1999 in a systematic way and
be suggested. It is thought that intrinsic abnormalities within                 on a large scale) (Coetsier and Dhont, 1998; De Sutter et al.,
the embryo are the major reason for failed conceptions or early                 2000; Dhont, 2001). The observation that miscarriage rates per
fetal death. The most signi®cant intrinsic factor contributing to               fetal sac are lower in twin than in singleton pregnancies
embryonic loss is aberrations in the ®rst meiotic division                      therefore may suggest that embryos in twin pregnancies have a
resulting in non-disjunction and aneuploidy. Trisomies 13, 15,                  better intrinsic potential than in singleton pregnancies. It would
16, 18 and 21 account for the most common autosomal                             seem that these embryos are part of a better cohort of embryos,
trisomies in spontaneous pregnancy losses (Racowsky, 2002).                     not only possessing a higher implantation potential, but also a
It may be that in the future preimplantation genetic screening                  higher potential for successful further development.
of embryos prior to transfer may reduce early pregnancy                            When the complete pregnancy loss in twins (5.1%) was
wastage resulting from aneuploidy (Handyside et al., 1999;                      compared with the pregnancy loss in singletons (21.1%) the
Munne et al., 1999).                                                            difference is very signi®cant.
   In assisted reproduction technology, inappropriate culture                      It is clear that the chance of complete pregnancy loss is more
conditions may considerably impair human embryonic devel-                       than twice as large in the singleton group when compared with
opment in vitro and implantation potential, and it is thought                   the twin group. This means that the embryological potential for
that the improvement in pregnancy rates after assisted repro-                   successful development is not the same in both groups.
                                                                                             Spontaneous abortion in singleton/twin pregnancies after art

    This is congruent with data comparing the outcome of IVF                    Dhont, M. (2001) Single embryo transfer. Semin. Reprod. Med., 19,
singleton and twin pregnancies with that of spontaneous                         Dhont, M., De Sutter, P., Ruyssinck, G., Martens, G. and Bekaert, A. (1999)
singleton and twin pregnancies (Dhont et al., 1999).                              Perinatal outcome of pregnancies after assisted reproduction: a case±control
    Because the decision whether a pregnancy was a singleton or                   study. Am. J. Obstet. Gynecol., 181, 688±695.
a twin could only be made at the moment of the ®rst ultrasound,                 ESHRE Capri Workshop Group (2000) Multiple gestation pregnancy. Hum.
                                                                                  Reprod., 15, 1856±1864.
it can be assumed that a fraction of these singletons were in fact              Ezra, Y. and Schenker, J. (1995) Abortion rate in assisted reproductionÐtrue
early vanishing twins. This early pregnancy loss might have                       increase? Early Pregnancy, 1, 171±175.
contributed to further pregnancy loss later on (after the ®rst                  Frates, M.C., Benson, C.B. and Doubilet, P.M. (1993) Pregnancy outcome
                                                                                  after a ®rst trimester sonogram demonstrating fetal cardiac activity. J.
ultrasound).                                                                      Ultrasound Med., 12, 383±386.
    Keeping in mind that multiple pregnancies are associated                    Fretts, R.C., Schmittdiel, J.M.A, McLean, F.H., Usher, R.H. and Goldman,
with a signi®cantly higher risk of complications for both                         M.B. (1995) Increased maternal age and the risk of fetal death. New Engl. J.
mother and child than singleton pregnancies, and given the                        Med., 333, 953±957.
                                                                                Gerris, J. and Van Royen, E. (2000) Avoiding multiple pregnancies in ART. A
economic consequences of twin pregnancies for society                             plea for single embryo transfer. Hum. Reprod., 15, 1884±1888.
(ESHRE Capri Workshop Group, 2000; Olivennes, 2000),                            Goldstein, S.R. (1994) Embryonic death in early pregnancy: a new look at the
the results of this study should not be interpreted as a plea for                 ®rst trimester. Obstet. Gynecol., 84, 294±297.
                                                                                Handyside, A.H. and Olgivie, C.M. (1999) Screening oocytes and
twin pregnancies and double embryo transfer. On the contrary,
                                                                                  preimplantation embryos for aneuploidy. Curr. Opin. Obstet. Gynecol.,
this study provides additional proof that SET is a logical                        11, 301±305.
strategy to consider (Vilska et al., 1999; Gerris and Van Royen,                Hill, L.M., Guzick, D., Fries, J. and Hixson, J. (1991) Fetal loss rate after
                                                                                  ultrasonically documented cardiac activity between 6 and 14 weeks

                                                                                                                                                                   Downloaded from humrep.oxfordjournals.org by guest on September 8, 2010
2000; Martikainen et al., 2001; De Sutter et al., 2002). Indeed,
                                                                                  menstrual age. J. Clin. Ultrasound, 19, 221±223.
in pregnancies occurring after SET, the embryo which was                        Landy, H.J. and Keith, L.G. (1998) The vanishing twin: a review. Hum.
selected for transfer would have implanted if a double embryo                     Reprod., 4, 177±183.
transfer had been performed. This implies that this single                                                               Â
                                                                                Martikainen, H., Tiitinen, A., Tomas, C., Tapanaien, J.S., Orava M.,
                                                                                  Tuomivaara, L., Vilska, S., Hyden-Granskog, C. and Hovatta, O. (2001)
embryo leading to pregnancy after SET has the same devel-
                                                                                  One versus two embryo transfer after IVF and ICSI: a randomized study.
opmental potential as when it would have been part of a twin                      Hum. Reprod., 16, 1900±1903.
pregnancy following a double transfer. It will be very                                 Â
                                                                                Munne, S., Alikani, M. and Tomkin, G. (1995) Embryo morphology,
interesting to extend this study in the future to all SET                         developmental rates and maternal age are correlated with chromosome
                                                                                  abnormalities. Fertil. Steril., 64, 382±391.
pregnancies, and it is to be expected that the miscarriage rates                       Â
                                                                                Munne, S., Magli, C., Cohen, J., Morton, P., Sadowy, S., Gianaroli, L., Tucker
in SET pregnancies will be as low as those reported for the                       M., Marquez, C., Sable, D., Ferrarretti, A.P. et al. (1999) Positive outcome
twins in the present study.                                                       after preimplantation diagnosis of aneuploidy in human embryos. Hum.
                                                                                  Reprod., 14, 2191±2199.
    In conclusion, after fetal heart activity is established, the risk
                                                                                Navot, D., Bergh, P.A. and Williams, M.A. (1994) Age related decline in
of abortion in IVF pregnancies is halved. The potential for                       female fertility is not due to diminished capacity of the uterus to sustain
survival is signi®cantly higher in twin pregnancies at all stages                 embryo implantation. Fertil. Steril., 61, 97±101.
of the ®rst trimester, pointing to a cohort phenomenon.                         Olivennes, F. (2000) Avoiding multiple pregnancies in ART double trouble:
                                                                                  yes a twin pregnancy is an adverse outcome. Hum. Reprod., 15, 1663±1665.
                                                                                Racowsky, C. (2002) High rates of embryonic loss, yet high incidence of
References                                                                        multiple births in human ART: is this paradoxal? Theriogenology, 57,
Abdalla, H.I., Burton, G., Kirkland, A., Johnson, M.R., Leonard, T., Brooks,    Sebire, N.J., Snijders, R.J, Hughes, K., Sepulveda, W. and Nicolaides, K.
  A.A. and Studd, J.W. (1993) Age, pregnancy and miscarriage: uterine             (1997) The hidden mortality of monochorionic twin pregnancies. Br. J.
  versus ovarian factors. Hum. Reprod., 8, 1512±1517.                             Obstet. Gynecol., 104, 1203±1207.
Abdalla, H.I., Wren, M.E., Thomas, A. and Korea, L. (1997) Age of the uterus    Simon, C., Landeras, J., Zuzuarregui, J., Martin, J., Remohi, J. and Pellice, A.
  does not affect pregnancy or implantation rates; a study of egg donation in     (1999) Early pregnancy losses in in vitro fertilization and oocyte donation.
  women of different ages sharing oocytes from the same donor. Hum.               Fertil. Steril., 72, 1061±1065.
  Reprod., 12, 827±829.                                                         Stoeckel, W. (1945) Lehbuch der Geburtshilfe. Gustav Fischer, Jena. (Quoted
Andersen, N., Wohlfahrt, J., Christens, P., Olsen, J. and Melbye, M. (2000)       in Landy and Keith, 1998.)
  Maternal age and fetal loss: population based register linkage study. Br.     VanRoyen, E., Mangelschots, K., De Neubourg, D.,Valkenburg, M., Van de
  Med. J., 320, 1708±1712.                                                        Meerssche, M., Ryckaert, G., Eestermans, W. and Gerris, J. (1999)
Coetsier, T. and Dhont, M. (1998) Avoiding multiple pregnancies in in-vitro       Characterization of a top quality embryo, a step forwards single embryo
  fertilization: who's afraid of single embryo transfer? Hum. Reprod., 13,        transfer. Hum. Reprod., 14, 2345±2349.
  2663±2664.                                                                                                 Â
                                                                                Vilska, S., Titinen, A. Hyden-Granskog, C. and Hovatta, O. (1999) Elective
DeSutter, P., Coetsier, T., Van der Elst, J. and Dhont, M. (2000) Elective        transfer of one embryo results in an acceptable pregnancy rate and
  single embryo transfer in IVF/ICSI: an analysis of 126 cases. 16th annual       eliminates the risk of multiple birth. Hum. Reprod., 14, 2392±2395.
  meeting of the European Society of Human Reproduction and Embryology.         Wilcox A.J., Weinberg C.R., O'Conner J.F., Baird, D., Schlatterer, J.,
  Hum. Reprod., 14, 2207±2215.                                                    Can®eld, R., Armstrong, G. and Nisula, B. (1988) Incidence of early loss of
DeSutter, P., Gerris, J. and Dhont, M. (2002) A health-economic decision-         pregnancy. New Engl. J. Med., 319, 189±194.
  analytic model comparing double with single embryo transfer in IVF/ICSI.
  Hum. Reprod., 17, 2891±2896.                                                  Submitted on January 13, 2003; accepted on April 16, 2003


To top