16
The fewer the better? Ethical issues in multiple
gestation
Mary B. Mahowald
Department of Obstetrics and Gynecology, University of Chicago School of Medicine, USA
Until the last part of the twentieth century, Hellin’s Law governed the
predictability of multiple births – the natural occurrence of twins in the
general population is 1/100, and the frequency of each higher multiple is
determinable by multiplying the denominator by 100, so that the frequency
of triplets is 1/10 000, the frequency of quadruplets is 1/1 000 000, and so on.
Since the advent of fertility drugs in the 1960s and in vitro fertilization (IVF)
in the 1970s, the incidence of multiple gestations has increased markedly. By
the late 1980s, the rate of multiple births had more than tripled; it appears to
be rising still (Hammon, 1998: p. 338).
With each higher order of multiples, risks to both fetus and pregnant
woman escalate. For women, the risks include anaemia, preterm labour,
hypertension, thrombophlebitis, preterm delivery and haemorrhage.
Tocolytic therapy to avoid preterm delivery introduces further risks. For
fetuses or potential children, the risks include intrauterine growth retarda-
tion, malpresentation, cord accidents and the usual sequelae of preterm
delivery, such as respiratory distress, intracranial haemorrhage and cerebral
palsy (Hammon, 1998: p. 339).
ConXicts between the interests of pregnant women and their fetuses are
not new; attempts to induce abortion and to rescue fetuses have occurred
through most of human history. Although medical advances have consider-
ably reduced the mortality and morbidity risks of childbearing for most
women and their oVspring, that same technology has introduced methods by
which people who would not otherwise reproduce can have biologically
related children. These methods are mixed blessings when the pregnancies
they facilitate exacerbate the risks of gestation for women and their fetuses.
They are also mixed blessings when, while providing a means to desired

dosage.
Obviously, prevention of multiple gestation is desirable and can probably
be accomplished in most cases. As already acknowledged, however, the
possibility of high multiples occurs in nature, albeit rarely, and the mortality
and morbidity of these gestations for women and some of their fetuses can
only eVectively be reduced by terminating other fetuses. In other words, the
criterion on which to base the medical prognosis for women and their
potential children in multiple gestations is ‘the fewer the better’. How, then,
does one reduce many gestating fetuses or embryos to fewer?
An apparent, relatively easy answer occurs in the context of in vitro
fertilization, when higher order multiples can be avoided by declining to
transfer more than three or four embryos after fertilization, storing or
disposing of extra ones in some other way. In fact, this is the usual practice of
reproductive endocrinologists, who tend to consider higher order multiples a
failure rather than a success. The recommendation to transfer only three or
four is thought to strike a balance between the risk of multiples and the risk of
not achieving a pregnancy at all. This approach does not adequately answer
the question raised, however, because multiple gestations are still possible,
248 M.B. Mahowald
regardless of whether fertilization occurs in vitro or in vivo. Moreover, the
disposition of untransferred embryos poses additional questions, which I
have addressed elsewhere (Mahowald, 2000: chapter 12).
Current techniques by which to limit the number of embryos or fetuses in
a multiple pregnancy involve either direct termination or removal of in vivo
embryos. The removal procedures are performed through transcervical suc-
tioning at 8 to 11 weeks’ gestation or through transvaginal aspiration usually
at six to seven weeks’ gestation. Unfortunately, the transcervical technique is
associated with a high (50 per cent) incidence of total pregnancy loss, and the
transvaginal technique precludes rudimentary detection of anomalies such as
nuchal folds (suggestive of Down’s syndrome). Transvaginal aspiration also

The language used to name procedures to reduce the number of developing
fetuses in an established gestation is controversial in its own right. Among the
terms utilized are selective birth, selective abortion, selective reduction, fetal
reduction and multifetal pregnancy reduction (Berkowitz et al., 1996).
Others that could be utilized are partial abortion or partial feticide. The term
‘selective birth’ has been used for cases of multiple gestation in which a
speciWc fetus had been identiWed as anomalous and targeted for termination.
(Targeting could occur for other reasons, such as sex selection.)
Prenatal detection of the anomaly is not possible until weeks, sometimes
months, after detection of the number of gestating fetuses. Ultrasound
guided cardiac injection of the targeted fetus is then the means through
which termination is accomplished. Obviously and perhaps misleadingly, the
term ‘selective birth’ focuses on the fetuses that are not targeted. ‘Selective
abortion’ would more accurately describe the procedure, but only if abortion
is deWned as termination of the fetus rather than termination of pregnancy.
‘Selective reduction’ is accurate if speciWc fetuses are targeted and if the
pregnancy itself is not thought to be ‘reduced’. But women, after all, are
neither more nor less pregnant, regardless of the number of fetuses they are
carrying. What is reduced, therefore, is the number of gestating fetuses. In
situations in which selective reduction of fetuses occurs, the actual procedure
is direct termination of the targeted fetus or fetuses. In these cases, ‘selective
termination’ would be a more accurate representation of what is intended
and done. If abortion is deWned as termination of the fetus rather than
termination of a non-viable pregnancy, ‘selective abortion’ would be accurate
when speciWc fetuses are targeted and ‘partial abortion’ would be accurate in
other cases as well. (Clinical texts usually deWne abortion as termination of a
non-viable pregnancy; popular understandings tend to identify it with termi-
nation of fetuses. Cf. Mahowald, 1982.) If abortion is deWned as termination
of a (non-viable) pregnancy, terminating one fetus while maintaining the
pregnancy through another (or others) is not equivalent to abortion.

strued, as an ethical consideration. Individual liberty may therefore be
subordinated to other goods in order to render the diVerent capabilities of
individuals as equal as possible – for example, in Amartya Sen’s notion of
equality of capability (Sen, 1995). But what are the capabilities to be con-
sidered with regard to FTPP and to whom do they belong?
DiVerent capabilities belong to diVerent individuals whose interests may
be promoted or impeded through FTPP. Although fetuses are not legally
persons, and their personhood is morally debatable, they are in fact living,
human and genetically distinct from the women in whom they develop.
Many human fetuses have the capability of becoming persons both legally
and morally. In high order multiple gestations, however, that capability is so
greatly and unalterably reduced (without intervention) that the scenario is
morally diVerent from, say, a twin gestation, where the capability of both
fetuses becoming legal and moral persons is high. The following cases
illustrate this morally relevant diVerence along with other variables that
inXuence the capabilities of individuals. Consideration of these variables is
crucial to ethical decisions about whether FTPP should be requested or
performed. Case 2a is one in which I was personally involved; case 3a is the
well-publicized case of the McCaughey septuplets. Although the other cases
are Wctitious, all of the features enumerated have occurred in real cases.
∑ Case 1a: Normal twins – during her second prenatal visit, a 36-year-old
mother of Wve children, aged 2 to 12 years, is told that she has a twin
251The fewer the better?
gestation. She tells her doctor that she thinks she can handle a single
newborn but not two at once. ‘I simply don’t have time for twins’, she says.
Having heard about FTPP, she asks whether this is an option for her. The
alternative of adoption is suggested but rejected. The woman has the
Wnancial resources to cover the costs of FTPP.
∑ Case 1b: Same case as 1a except that one fetus has Down’s syndrome.
∑ Case 1c: Same case as 1a except that one fetus has trisomy 13.

to be identiWed because they are sometimes associated with inequality or
unjust discrimination. The variable of ability to pay, for example, expands the
options of some women while restricting those of others; it may thus
exemplify classism. Targeting disabled fetuses or fetuses of a speciWc sex
suggests unequal regard for one individual or group as opposed to another; it
252 M.B. Mahowald