Human Reproduction vol.13 no.7 pp.2005–2008, 1998
Lower risks of adverse outcome in twins conceived by
artificial reproductive techniques compared with
spontaneously conceived twins
H.Minakami1,3, M.Sayama1, Y.Honma2,
S.Matsubara1, T.Koike1, I.Sato1, A.Uchida2,
Y.Eguchi2, M.Momoi2and S.Araki1
Departments of1Obstetrics and Gynaecology and2Pediatrics, Jichi
Medical School, Minamikawachi-machi, Tochigi, 329-04 Japan
3To whom correspondence should be addressed
The outcomes of twins conceived by 136 women after
medical assistance (MA) such as ovulation induction with
or without assisted reproductive techniques, and twins
conceived spontaneously (SP) by 72 women were compared.
All 208 women were monitored from ?20 weeks gestation;
they all delivered at ?24 weeks gestation. The chorionicity
of the placenta was diagnosed antenatally and confirmed
after delivery. There were 10 perinatal deaths; the physical
and neurological status of the remaining 406 infants was
assessed at 1 year of corrected age. There were no differ-
ences in gestational age at birth, the birth weights of the
larger and smaller twins, the birth weight discordance,
or the incidence of life-threatening major malformations
between groups. Adverse infant outcomes, such as death,
cerebral palsy and mental retardation occurred in nine
(3.3%) of 272 MA twins compared with 12 (8.3%) of 144
SP twins (P ? 0.05). The placenta was monochorionic in
only three (2.2%) of 136 MA twin pregnancies compared
with 41 (57%) of 72 SP twin pregnancies (P ? 0.001). Of the
21 infants with adverse outcomes, nine had monochorionic
placentas. Thus, the risk of an adverse outcome was
2.8-fold higher (95% confidence interval (CI) 1.2–6.4) in
monochorionic twins than in dichorionic twins (10 versus
3.7%; P ? 0.05). There was no difference in the incidence
of adverse infant outcomes between SP (4.8%) and MA
(3.4%) twins with dichorionic placentas. These findings
suggest that ovulation induction in itself was not associated
with an adverse outcome of twin pregnancies. The lower
frequency of monochorionic placentas in MA twins may
have been responsible for the lower risk of an adverse
outcome in MA twins.
Key words: assisted reproductive techniques/cerebral palsy/
mental retardation/twin pregnancy
Infants conceived after ovulation induction, in-vitro fertiliza-
tion (IVF) and gamete intra-Fallopian transfer (GIFT) do not
have an increased risk of major malformations (Cohen et al.,
1988; Beral and Doyle, 1990; Rizk et al., 1991). However,
the risks of obstetric and neonatal complications, such as pre-
© European Society for Human Reproduction and Embryology
eclampsia, intrauterine growth retardation, premature birth,
placenta previa and Caesarean birth, are increased in IVF
pregnancies (Tan et al., 1992; Wang et al., 1994; Tallo et al.,
1995; Tanbo et al., 1995), suggesting that infants born to these
women have an increased risk of an adverse outcome. Twin
pregnancies account for about 20% of the pregnancies after
IVF (Beral and Doyle, 1990; Rizk et al., 1991). It is possible
that the risk of an adverse outcome is higher among twins
conceived after ovulation induction with or without GIFT/IVF
than in naturally occurring twins. However, data on the
outcomes of twins conceived spontaneously (SP) and those
conceived after medical assistance (MA) are limited (Bernasko
et al., 1997).
The twin-to-twin transfusion syndrome (TTTS) occurs in
about 15–30% of twins with monochorionic (MC) placentas
(Rausen et al., 1965; Patten et al., 1989) but is rare in twins
with dichorionic (DC) placentas (Robertson and Neer, 1983)
and TTTS markedly increases the risk of an adverse outcome
(Rausen et al., 1965; Patten et al., 1989). Thus, the chorionicity
of the placenta is an important determinant of the outcome of
twins. Because the frequency of MC placentas is expected to
be lower in twins conceived after MA than in twins conceived
spontaneously (Derom et al., 1987; Wenstrom et al., 1993), it
is possible that twins conceived after MA have a lower risk
of an adverse outcome than twins conceived spontaneously.
Data from tertiary institutions on the prognosis of twins
may include data on women who were referred because they
were ill or had complications such as premature labour, TTTS
and pre-eclampsia. Thus, the incidence of complications and
adverse infant outcomes may be higher than would be found
in the general population.
We investigated the influence of ovulation induction with
or without GIFT/IVF on the outcome of twins born to women
who had been monitored at our antenatal clinic since ?20
weeks gestation and in whom the chorionicity of the placenta
Materials and methods
Of 242 women with twin pregnancies who delivered at ?24 weeks
of gestation at the Jichi Medical School Hospital between January
1990 and May 1996, 208 women (86%) booked their delivery at our
hospital before 20 weeks gestation and were followed-up at our
antenatal clinic. We reviewed the medical records of these 208
women. Twins were spontaneously conceived in 72 women (SP)
and conceived after ovulation induction with or without assisted
reproductive techniques in 136 medically assisted (MA) women.
Medical assistance consisted of clomiphene only in 15 women (11%),
human menopausal/chorionic gonadotrophin alone in 24 (18%), GIFT
H.Minakami et al.
in 61 (45%), IVF in 34 (25%) and intracytoplasmic sperm injection
in two (1.5%).
In general, ultrasound examinations were performed every 2 weeks
from the time of diagnosis of twins until delivery, unless more
in the first trimester in all women. The gestational age was based on
the date of conception or the chart of the basal body temperature in
MA twins and on the last known menstrual period if it was within 7
days of the first-trimester ultrasound-derived estimate in SP twins. If
the discrepancy was ?7 days, the gestational age was based on the
ultrasound estimate. The chorionicity of placentas was confirmed by
histological examination after delivery; the presumptive antenatal
diagnosis of chorionicity determined by ultrasonography was correct
in 205 (99%) of 208 twin gestations.
There were 10 infant deaths before the age of 1 year; all 10 died
within 28 days of birth. The physical and neurological status of the
remaining 406 infants was assessed by paediatricians on a regular
basis until at least 1 year of corrected age. Birth weights were
recorded, and the intertwin birth weight discordance was calculated
as the percentage of the weight of the larger twin. An adverse infant
outcome was defined as death or disability, including cerebral palsy,
epilepsy, deafness, blindness and mental retardation.
If TTTS was suspected at ?30 weeks, the pregnancy was termin-
were performed. The antenatal diagnosis of TTTS was based on the
following criteria: (i) a gradual increase in the intertwin weight
discordance to ?20%, as estimated by ultrasonography; (ii) a gradual
increase in the intertwin discordance of the amniotic fluid volume,
as estimated by ultrasonography; and/or (iii) the appearance of signs
of fetal cardiac dysfunction, such as an increased cardiothoracic ratio
and/or tricuspid regurgitation, as determined by fetal echocardi-
Pre-eclampsia was defined as a systolic or diastolic blood pressure
?140 or ?90 mmHg, respectively, on two occasions recorded 24 h
apart in association with the onset of proteinuria in a patient who
had been normotensive during the first 20 weeks of gestation.
The data were analysed by Student’s t-test or the chi-square test
with Yates’ correction. Results are presented as the mean ? SD.
Miettinen’s method (1976) was used to determine the 95% confidence
interval (95% CI). A probability level of ? 0.05 was accepted as
The maternal age and the percentage of nulliparous women
were significantly higher in the 136 women with MA twins
than in the 72 women with SP twins (Table I). There were no
differences in the incidence of pre-eclampsia, the gestational
week at delivery, the birth weights of the larger and smaller
twins, or the birth weight discordance between groups.
Adverse outcomes occurred in 12 SP infants (seven larger
and five smaller twins) born to eight mothers and included
four deaths (two larger and two smaller twins) and eight
disabled infants (Table I). Both twins were affected in four
(6%) of 72 women. There were two deaths due to major
malformations; Ebstein’s cardiac anomaly in one infant and
anencephaly in one. In the MA group, adverse outcomes
occurred in nine infants (three larger and six smaller twins)
born to nine women and included six deaths (one larger and
five smaller twins) and three disabled infants. There was no
case of both twins being affected in any of the 136 women.
One death was due to an interrupted aortic arch. Thus, the
incidence of adverse outcomes was significantly higher in SP
twins than in MA twins and both twins were compromised in
a significantly greater proportion of women with SP twins (P
? 0.05). There was no difference between groups in the
incidence of major malformations.
Pairs in which at least one infant was affected had an earlier
gestational age at delivery, lower birth weights of both the
larger and smaller twins, and a greater birth weight discordance
than unaffected pairs (Table I).
Only three (2.2%) of 136 women with MA twins had an
MC placenta compared with 41 (57%) of 72 women with SP
twins (P ? 0.001). Thus, the relative risk of an MC placenta
was 0.04 (95% CI, 0.01–0.12) in the MA group compared
with the SP group. Of the 21 infants with adverse outcomes,
nine were born to women with MC placentas and the remaining
12 to women with DC placentas. Thus, the risk of an adverse
outcome was 2.8-fold higher (95% CI, 1.2–6.4) in MC twins
(10%) than in DC twins (3.7%) (P ? 0.05). In addition, both
twins were compromised in three (7%) of 44 women with MC
placentas compared with one (0.6%) of 164 women with DC
placentas (P ? 0.05).
TTTS was diagnosed antenatally in 14 (32%) of 44 women
with MC placentas and in none of the 164 women with DC
placentas. Delivery occurred at 32.7 ? 2.8 weeks of gestation
in the 14 women whose pregnancies were complicated by
TTTS. Of the 21 infants with adverse outcomes, eight (38%)
were born to women in whom TTTS was suspected. Thus,
TTTS was associated with an 8.5-fold increase in the risk of
an adverse outcome (95% CI, 3.9–18.9) [8/28 (29%) versus
13/388 (3.4%), P ? 0.001].
When data were analysed only for twins with DC placentas,
there was no difference in the incidence of adverse infant
outcomes between SP and MA twins (Table II). Major mal-
formations occurred in none of the SP twins and in one of the
In the present cohort study, no women had any complications
other than twinning at ?20 weeks gestation and all were
delivered at ?24 weeks gestation at our hospital. Although
only a small number of subjects was studied, the risk of an
adverse outcome was found not to be increased in MA twins
as compared with naturally occurring twins, thus supporting
the results of earlier studies (Bernasko et al., 1997).
In the present study, the risk of an adverse outcome was
2.8-fold higher in MC twins than in DC twins and there was
no difference in the incidence of adverse infant outcomes
between SP and MA twins with DC placentas. These results
suggested that the chorionicity of the placenta was an important
determinant of the outcome of twins, and ovulation induction
in itself was not associated with an adverse outcome.
Twins with MC placentas are monozygous, while those with
DC placentas can be monozygous or dizygous. According to
the results of a study in Belgium (Derom et al., 1987), 445
of 869 (51.2%) naturally occurring twin gestations were
monozygotic. Of these 445, some 284 (63.8%) had MC
Infant outcome: natural versus artificial twins
Table I. Clinical characteristics and infant outcome
SP twins MA twins
(n ? 72)
(n ? 8)
(n ? 64)
(n ? 136)
(n ? 9)
(n ? 127)
Gestational week at delivery
Birth weight (g)
Birth weight discordance (%)
No. of infants with adverse outcome 12/144(8.3)b,f
29.4 ? 4.4a
35.3 ? 2.4
31.3 ? 4.5
33.6 ? 2.9d
29.2 ? 4.3
35.6 ? 2.2
32.0 ? 3.6
35.6 ? 2.4
32.8 ? 1.0
30.9 ? 4.6e
32.0 ? 3.7
2288 ? 470
1998 ? 465
13 ? 11
1917 ? 662d
1396 ? 189c
17 ? 17c
2334 ? 424
2073 ? 400
11 ? 8
2268 ? 424
1984 ? 419
12 ? 10
1610 ? 712e
1144 ? 563e
23 ? 24e
2314 ? 357
2044 ? 338
11 ? 8
aP ? 0.001 versus MA twins.
bP ? 0.05 versus MA twins.
cP ? 0.001 versus subgroup of SP twins with better outcome.
dP ? 0.05 versus subgroup of SP twins with better outcome.
eP ? 0.001 versus subgroup of MA twins with better outcome.
fFour deaths and eight disabled infants.
gSix deaths and three disabled infants.
MA twins, conceived after medical assistance, e.g. ovulation induction; SP twins, spontaneously conceived twins.
Values in parentheses are percentages.
Table II. Characteristics and outcome of spontaneous (SP) and medically
assisted (MA) dichorionic twins
SP DC twins
(n ? 31)
MA DC twins
(n ? 133)
Gestational week at delivery
Birth weight (g)
Birth weight discordance (%)
No. of infants with adverse outcome
30.8 ? 4.4
36.2 ? 1.3
32.1 ? 3.6
35.7 ? 2.4
2443 ? 354b
2184 ? 323b
10 ? 4
2274 ? 422
1989 ? 420
12 ? 11
aP ? 0.001 versus MA twins.
bP ? 0.05 versus MA twins.
cThree disabled infants.
dSix deaths and three disabled infants.
DC twins, twins with dichorionic placenta; MA twins, conceived after
medical assistance, e.g. ovulation induction; SP twins, spontaneously
Values in parentheses are percentages.
placentas, suggesting that ?30% of the SP twins had MC
placentas. The monozygotic twinning rate per 1000 total
pregnancies is relatively constant worldwide, varying from
about three to five (MacGillivray, 1986), which was consistent
with the 0.45% expected frequency of splitting after spontan-
eous ovulation (Derom et al., 1987). However, the dizygotic
twinning rate varies largely by race, from 1.3/1000 in Japan
to 7–9/1000 in the USA and Europe and 50/1000 in Nigeria
(MacGillivray, 1986). Thus, the relative prevalence of MC
placentas among twins differs between countries and may be
higher in Japan than elsewhere. The relative prevalence of
MC placentas among twins conceived after artificial induction
of ovulation has not been fully investigated. Although a higher
frequency of monozygotic twins is observed in pregnancies
conceived after ovulation induction with or without IVF (10–
13 per 1000 pregnancies) (Edwards et al., 1986; Derom
et al., 1987) than after spontaneous ovulation (3–5 per 1000
pregnancies), monozygotic twins account for only 13% of all
MA twins (Derom et al., 1987). Because some monozygotic
twins have DC placentas, the frequency of MC placentas
among MA twins may be ?13%, consistent with the results
of Wenstrom et al. (1993) and the present study. Thus, an
increase in the number of MA twins decreases the relative
prevalence of MC twins in many countries, although the actual
number of MC twins increases.
compared with twins of different sexes (Fowler et al., 1991;
Rydhstroem, 1996) and the risk of death is higher in twins
with MC placentas than with DC placentas (Fusi and Gordon,
1990; Kilby et al., 1994). TTTS occurs in about 15–30% of
twins with MC placentas (Rausen et al., 1965; Patten et al.,
1989) and markedly increases the risk of an adverse outcome
(Rausen et al., 1965; Patten et al., 1989). As expected, twins
with MC placentas had an increased risk of adverse outcome
compared with those with DC placentas in the present study.
Only 2.2% of women in the MA group had MC placentas
compared with 57% in the SP twin group. There was no
difference in infant outcome between MA and SP twins when
women with MC placentas were excluded from the analysis.
In the present study, twin pairs in which at least one infant
had an adverse outcome had a significantly greater birth weight
discordance than the unaffected pairs in both MA and SP
twins. Nearly all twins with MC placentas shared placental
circulation (Robertson and Neer, 1983). Intertwin blood trans-
fusion may always occur in MC twins. The greater weight
discordance in twins with MC placentas may reflect a greater
volume of transfused blood and/or a longer duration of hypoxic
stress due to anaemia in the donor twin, in whom growth may
be retarded (Snijders et al., 1993; Maier et al., 1995). The
placental function between twins. A greater weight discordance
H.Minakami et al. Download full-text
may indicate that the duration of placental insufficiency and
hypoxic stress are greater in the smaller twin (Snijders et al.,
1993; Maier et al., 1995). Thus, in addition to immaturity
(infants with adverse outcomes had a younger gestational age
at birth than the other twins in both MA and SP twins), the
greater weight discordance was associated with adverse infant
outcomes in both MA and SP twins.
In conclusion, twins conceived after MA such as ovulation
induction with or without GIFT/IVF did not have an increased
risk of an adverse outcome compared with naturally occurring
twins. The frequency of MC placentas was lower in MA twins
than in SP twins, which may have favourably influenced the
outcome of MA twins.
Beral, V. and Doyle, P. (1990) Report of the MRC Working Party on Children
Conceived by In Vitro Fertilisation. Births in Great Britain resulting from
assisted conception, 1978–87. Br. Med. J., 300, 1229–1233.
Bernasko, J., Lynch, L., Lapinski, R. and Berkowitz, R.L. (1997) Twin
pregnancies conceived by assisted reproductive techniques: maternal and
neonatal outcome. Obstet. Gynecol., 89, 368–372.
Cohen, J., Mayaux, M.J. and Guihard-Moscato, L. (1988) Pregnancy outcomes
after in vitro fertilization. A collaborative study on 2342 pregnancies. Ann.
N.Y. Acad. Sci., 541, 1–6.
Derom, C., Vlietinck, R., Derom, R. et al. (1987) Increased monozygotic
twinning rate after ovulation induction. Lancet, i, 1236–1238.
Edwards, R., Mettler, L. and Walters, D. (1986) Identical twins and in vitro
fertilization. J. In Vitro Fertil. Embryo Transfer, 3, 114–117.
Fowler, M.G., Kleinman, J.C., Kiely, J.L. and Kessel, S.S. (1991) Double
jeopardy: twin infant mortality in the United States, 1983 and 1984. Am.
J. Obstet. Gynecol., 165, 15–22.
Fusi, L. and Gordon, H. (1990) Twin pregnancy complicated by single
intrauterine death. Problems and outcome with conservative management.
Br. J. Obstet. Gynaecol., 97, 511–516.
Kilby, M.D., Govind, A. and O’Brien, S. (1994) Outcome of twin pregnancies
complicated by a single intrauterine death: a comparison with viable twin
pregnancies. Obstet. Gynecol., 84, 107–109.
MacGillivray, I. (1986) Epidemiology of twin pregnancy. Semin. Perinatol.,
Maier, R.F., Bialobrzeski, B., Gross, A. et al. (1995) Acute and chronic fetal
hypoxia in monochorionic and dichorionic twins. Obstet. Gynecol., 86,
Miettinen, O.S. (1976) Estimability and estimation in case-referent studies.
Am. J. Epidemiol., 103, 226–235.
Patten, R.M., Mack, L.A., Harvey, D. and Cyr, D.R. (1989) Disparity of
amniotic fluid volume and fetal size: problem of the stuck twin – US
studies. Radiology, 172, 153–157.
Rausen, A.R., Seki, M. and Strauss, L. (1965) Twin transfusion syndrome.
J. Pediatr., 66, 613–628.
Rizk, B., Doyle, P., Tan, S.L. et al. (1991) Perinatal outcome and congenital
malformations in in-vitro fertilization babies from the Bourn-Hallam group.
Hum. Reprod., 6, 1259–1264.
Robertson, E.G. and Neer, K.J. (1983) Placental injection studies in twin
gestation. Am. J. Obstet. Gynecol., 147, 170–174.
Rydhstroem, H. (1996) Pregnancy with stillbirth of both twins. Br. J. Obstet.
Gynaecol., 103, 25–32.
Snijders, R.J.M., Abbas, A., Melby, O. et al. (1993) Fetal plasma erythropoietin
concentration in severe growth retardation. Am. J. Obstet. Gynecol., 168,
Tallo, C.P., Vohr, B., Oh, W. et al. (1995) Maternal and neonatal morbidity
associated with in vitro fertilization. J. Pediatr., 127, 794–800.
Tan, S.L., Doyle, P., Campbell, S. et al. (1992) Obstetric outcome of in vitro
fertilization pregnancies compared with normally conceived pregnancies.
Am. J. Obstet. Gynecol., 167, 778–784.
Tanbo, T., Dale, P.O., Lunde, O. et al. (1995) Obstetric outcome in singleton
pregnancies after assisted reproduction. Obstet. Gynecol., 86, 188–192.
Wang, J.X., Clark, A.M., Kirby, C.A. et al. (1994) The obstetric outcome of
singleton pregnancies following in-vitro fertilization/gamete intra-Fallopian
transfer. Hum. Reprod., 9, 141–146.
Wenstrom, K.D., Syrop, C.H., Hammitt, D.G. and Van Voorhis, B.J. (1993)
Increased risk of monochorionic twinning associated with assisted
reproduction. Fertil. Steril., 60, 510–514.
Received on December 31, 1997; accepted on March 31, 1998