Higher-Order Multiple Gestations

Higher-order gestations are multiple pregnancies with three or more fetuses. Once a rarity, these pregnancies are now seen more frequently than in prior decades because of infertility treatments. Data obtained from an annual survey of programs and units providing assisted reproductive technologies (ART) show remarkable consistency in the proportion of multiple births after therapy regardless of the specific method employed.¹ Twins occur in 25%, 27%, 28.6%, and 31.4% of patients treated by in vitro fertilization (IVF), gamete intrafallopian transfer (GIFT), zygote intrafallopian transfer (ZIFT), and donor oocyte use, respectively. Triplets occur in 4.8%, 6.6%, 4.4%, and 5.7%, respectively, of patients treated with the same technologies.¹

These pregnancies have higher morbidity and mortality rates than twin or singleton pregnancies generally because triplets are born at an earlier gestational age and weigh less than twins or singletons. Figure 1 shows the cumulative distribution of live births by gestational categories and plurality for the United States in 1990. Figure 2 shows the cumulative distribution of birth weights of all live births by plurality in the United States for the same year.

This chapter addresses higher-order gestation as an entity separate from twins.

The spontaneous occurrence of triplets and higher-order gestations is rare. The natural incidence of triplets in the United States is 1 in 7925,² whereas quadruplets may be as rare as 1 in 600,000, and quintuplets probably occur only once in 15 to 20 million deliveries.

Data from the vital statistics of the United States help to place these earlier calculations in perspective. In 1994, the frequency of twin births was 1 in 40, whereas it had been 1 in 48 in 1985. In 1994, the frequency of triplets and higher-order multiple births was 1 in 838, whereas it had been 1 in 1912 in 1985.^3,4 In 1994, the proportion of first births to women aged 30 years or older was 21%, whereas it had been only 5% in 1970. At the same time, the proportion of all births to women in this age cohort was 34%, in contrast to 16% in 1970.^4,5 The increase in the rate of triplet births during the decade of the 1980s was 100%.⁶

It is not possible to calculate accurately the spontaneous incidence of sextuplets, septuplets, octuplets, and nonuplets. In the past, the use of Hellin's equation was thought to provide an approximation of such occurrences. Using the known natural incidence of twinning, a geometric progression can be calculated in the form of X^N – ¹, where X equals the incidence of twinning, and N is the particular higher-order gestation (i.e., 80² = 6400, or one triplet pregnancy for every 6400 deliveries). In our opinion, this equation is now of historical interest.

The uncommon occurrence of higher-order gestation precludes clearly identifying the cause of this phenomenon. In specific populations, such as the Yoruba tribe in Nigeria, twin and higher-order gestations occur commonly, and it has been proposed that dietary factors may act as natural ovulation-induction substances, thus generating superovulation.⁷ In other populations with a less pronounced incidence of multiple births, however, it remains unclear what initiates a higher-order gestation. Parity, maternal age, race, family history, nutritional status, pregravid weight, and body mass index all have been identified as risk factors for multiple gestation, but most studies have only compared twins with singletons because of insufficient numbers of natural higher-order gestations.

The possible biologic origin of higher-order gestations may be explained by either superovulation or cleavage of a zygote. In superovulation, three or more ova are released during a single cycle and are fertilized and implanted, thus generating a trizygotic (or higher) gestation (Fig. 3). This kind of placentation is free from vascular anastomosis. In other instances, however, a single fertilized zygote undergoes one or more cleavages, giving rise to monozygotic higher-order gestations. As an example, the Dionne quintuplets were considered monozygotic on the basis of their placental examination.

As Figure 3 also shows, combinations of both mechanisms can coexist. In the event that monozygotic conception gives rise to a monochorionic placentation, the presence of vascular anastomoses may lead to a clinical picture similar to the twin-twin transfusion, and growth discordancy may be seen. Of interest, a 1988 report suggests that up to 25% of all triplets had a monozygotic placentation.⁸ Isolated reports of monochorionic placentation in higher-order multiple gestations have been published anecdotally, but no case series has been presented.

Ovulation-Induction Drugs

The drugs available in the United States for ovulation induction include clomiphene citrate, human menopausal gonadotropin (hMG), and follicle-stimulating hormone (FSH). The administration of these drugs, without the concomitant help of ART, has been known for many years to be associated with the production of multiple gestations. The incidence of twin gestations associated with use of clomiphene ranges from 6% to 8.4%; and the incidence is reported to be 0.5% for triplets, 0.3% for quadruplets, and 0.13% for quintuplets.⁹ With hMG, the reported rates (16% to 66%) of multiple gestations vary widely.⁹ Blankstein and coworkers¹⁰ reported 378 pregnancies achieved with ovulation induction. Of these, 3.5% were triplets, 1% quadruplets, and 0.3% quintuplets.

Assisted Reproduction Technologies

Since 1980, numerous variations of ART have been reported. The early experiences with IVF and GIFT were accompanied by an increasing incidence of higher-order gestations. One review¹¹ documented increasing numbers of higher-order gestations not only in the United States but also in England, Australia, New Zealand, and Japan—all countries in which ART is used with regularity. The reason ART is associated with multiple gestation is the practice of transferring more than one fertilized oocyte (usually in a blastomere stage) with IVF and GIFT. The basis for this is the recognition that if more than one oocyte is transferred back to the mother, the possibility of achieving at least one implantation is higher. This dictum appears valid until four fertilized oocytes are transferred back. After this point, the total pregnancy rate does not increase significantly, but the incidence of higher-order gestations increases.

A report¹² on IVF from Australia also addressed this issue. The transfer of three or more fertilized oocytes generated an incidence of 30% multiple gestations. Craft and colleagues¹³ reported on 726 GIFT procedures. When up to four oocytes were transferred, 7.5% of pregnancies resulted in twin gestation. If the number was increased to five to eight oocytes, 15.3% of pregnancies resulted in multiple gestation. If the number of oocytes was increased to nine or more, the incidence of multiple gestation increased to 17.8%. Additional analysis of Craft's review yielded an incidence of triplets of 6.7%, quadruplets 1.1%, and quintuplets 0.5%.

In an analysis of successful ART conceptions during 1990 to 1991, Wilcox and colleagues¹⁴ documented an increase of 191% in the incidence of triplet and higher-order gestations over that recorded in 1972 to 1974. The two largest components of this increase were ART, accounting for 38%, and childbearing among older women, accounting for 30% of the increase. The total contribution of ART to the total number of triplet and higher-order births during 1990 to 1991 was 22%.

Without doubt, the increasing incidence of multiple births impacts significantly on available medical and economic resources. These economic considerations are illustrated in Table 1, which projects the need for neonatal intensive care unit (NICU) days based on the differences in birth-weight distribution between singletons, twins, and higher-order births. These differences result from the greatly increased rates of preterm delivery in twins, triplets, and higher-order births. In addition, the emotional and physical impact of caring for three or more children at the same stage of development is great. These two considerations have led to calls to curb or modify present regimens of ART, which are associated with high risks of higher-order gestations.¹⁵ At the least, recognition of the costs and social impact of high-order gestations resulting from ART has provided practitioners with a sense of caution that must be balanced with the desire to achieve a pregnancy by these means.

TABLE 1. Need and Cost for NICU Days Per 1000 Live Births

*Calculated at $1000/d.
(Adapted from Keith LG, Papiernik E, Luke B: The costs of multiple pregnancy. Int J Fertil 36:109, 1991)

The perinatal mortality rate (PMR) of higher-order multiple gestations has improved dramatically in the past two decades. When Holcberg and co-workers¹⁶ reviewed 31 triplets delivered between 1960 and 1979 at the Ben Gurion University in Israel, the PMR was 380 in 1000. More recently, Lipitz and co-workers¹⁷ reviewed 78 triplets delivered between 1975 and 1978 at the Tel Aviv University, also in Israel. This series reported a PMR of only 93 in 1000, or a reduction of 75%. Even more recently, Newman and colleagues¹⁸ reported data from 198 triplet gestations managed at several centers in the United States. The overall PMR in this series was an exceptionally low 29 in 1000 births. Although the authors of this latter report believed that their result could be partially explained by all patients being enrolled in an intense antenatal surveillance program, they also acknowledged that advances in the management of very low birth-weight neonates and in the quality of the antenatal obstetric services were at a minimum equally important in reducing the PMR for triplets and, in some instances, higher-order gestations. Similar improvements in outcome have been seen in the United Kingdom. Kingsland and colleagues¹⁹ reported on 43 consecutive sets of triplet deliveries between 1984 and 1987. The PMR was 32.26, and the neonatal death rate was 21.98 per 1000 births.

A recent population based analysis of all live births and all fetal deaths in the United States between 1983 and 1988²⁰ documented that the risk of fetal death of triplets (based on 9,523 live births and 204 fetal deaths) was higher than that of twins (based on 352,629 live births and 5,467 fetal deaths) which in turn was higher than that of singletons (based on 19,468,170 live births and 83,263 fetal deaths). In this analysis, the optimal birth weight for triplets (for avoiding fetal death) was 1900- 2200 g and the optimal estimated gestational age was 34–35 weeks. Table 2 shows that these optimal circumstances are still associated with a rate of fetal death five times that of singletons.

TABLE 2. Lowest Fetal Death Rates, USA, 1983-1988

(Luke B: Reducing fetal deaths in multiple births: Optimal birthweights and gestational ages for infants of twin and triplet births. Acta Genet Med Gemellol 45:361, 1996)

The high PMR associated with triplet and higher-order gestations is directly related to the exceptional rates of preterm delivery associated with these pregnancies. The average gestational age at delivery for triplets is about 33 weeks, with up to 95.5% of the deliveries occurring before 37 weeks.¹⁸ Collins and Blyel²¹ reported on 71 quadruplet pregnancies delivered at a variety of institutions. Information was obtained by parental questionnaire. The estimated gestational age (EGA) at delivery for these quadruplets was 31.4 weeks, with 95.8% delivering preterm. The PMR in this series was 67 in 1000. Undoubtedly, this relatively modest figure reflects the intense level of care these infants received after delivery. The authors of this report did not provide long-term follow-up and assessment of late sequelae of extreme prematurity. However, another recent series by Lipitz and colleagues²² on 11 higher-order gestations suggested that adverse neurologic sequelae are a real possibility. Of the 30 infants followed to 2 years of age, 30% had abnormal examinations, ranging from mild motor and mental developmental delay to cerebral palsy and mental retardation. Of interest, Sassoon and co-workers²³ matched 15 triplets to 15 twins in a matched-pair analysis study. The triplets had a fivefold increased risk of NICU admission, with an average hospital stay of 29 days, compared with 8.5 days for twins. Keith and colleagues²⁴ have estimated the risk of neurologic or serious handicaps to increase almost geometrically from singletons to twins to triplets to quadruplets (Table 3). To our knowledge, however, the validity of these estimates has not been tested on a national sample of all higher-order gestations.

TABLE 3. Risk, Relative Risk, and Cost for Handicaps Per 1000 Births

*Calculated at $1000/d.

Other authors also have looked at frequency of admission to NICU, length of stay, and average hospital charge for delivery. In a study reported on by Seoud and colleagues,¹ 23% of twins, 64% of triplets and 75% of quadruplets were admitted to the NICU, where their stays varied, on the average, from 12 days for each twin to 17 days for each triplet and to 58 days for each quadruplet. When the comparative charges for delivery were assessed in singletons, twins, and triplets, the increase in dollars was not proportional to the increase in number of uterine occupants.²⁵ Thus, the average singleton charge was $9800, but the charge for each case of twins was more than four times greater ($38,000), and the cost for triplets was 11 times greater ($110,000).

The rarity of reports of successful delivery of quintuplet through nonuplet gestation makes it difficult to assess the issues of prematurity, PMR, and risk status using twins as a baseline. The scant available data suggest that these risks rise exponentially. A review by Petrikovsky and Vintzileos² calculated a PMR of 257 in 1000 for quintuplets (or four times that of quadruplets) from eight published case reports. The average EGA at delivery was 32.5 weeks.

To our knowledge, there are only four published reports of sextuplet pregnancy delivery,^26,27,28,29 with EGAs at delivery ranging from 30 to 34.5 weeks. Of these 24 fetuses, 14 survived, yielding a PMR of 416 in 1000. Of the published reports of septuplet^30,31 and nonuplet³² pregnancies, none of the fetuses survived. The single report of an octuplet³³ pregnancy reports five surviving fetuses, who were delivered at 33.4 weeks.

Anemia

The increased demands on the iron reserves of the mother from the larger placental mass and for erythropoiesis from three or more fetuses virtually guarantees that higher-order multiple gestational mothers are at risk for anemia. The reported incidence of anemia in triplet pregnancies ranges from 13%¹⁶ to 35%.³⁴ In quadruplet pregnancies, anemia has been reported in 25% of cases.²¹ In a series of 21 triplet and quadruplet pregnancies, 6 patients required blood transfusions as part of their therapy.³⁵

Premature Rupture of the Membranes

Premature rupture of membranes (PROM) occurs frequently in higher-order multiple gestations. The reported incidence in triplets ranges from 6%³⁶ to 26.7%.²³

Mashiach and co-workers³⁷ reported on a triplet pregnancy in a patient with didelphic uterus. Two fetuses were in the right horn, and the third fetus was in the left. Labor ensued at 27 weeks, with delivery of triplet A vaginally and triplet B by cesarean section; both gestated in the right horn. Triplet C was left untouched in the left horn and was delivered 72 days later by a second cesarean section at 37 weeks. Simpson and colleagues³⁸ reported on a triplet pregnancy that was complicated by rupture of membranes at 23.5 weeks with delivery of the first triplet. After 99 days, the patient was delivered of the remaining fetuses by cesarean section. Triplet B was an intrauterine fetal demise (IUFD), but triplet C was delivered alive and well. Banchi³⁹ described a case of a 16-week abortion of triplet A, followed by cerclage and bed rest, with intermittent tocolysis and cesarean section 131 days later at 35 weeks; both infants (triplets B and C) survived. Cardwell and associates⁴⁰ and Schaal and co-workers⁴¹ reported on two additional cases of delayed interval delivery. Cardwell and associates⁴⁰ reported on the delivery of triplets, each one on a different day, a total of 16 days apart. The first delivered at 23 weeks and 6 days, the second at 24 weeks and 3 days, and the third at 25 weeks and 6 days. The patient developed pseudomembranous enterocolitis during the interval between the second and third deliveries; the authors believed that this complication originated from empiric antibiotics administered after a cerclage that followed the second delivery. The last delivery was complicated by chorioamnionitis, but this fetus survived, whereas the others succumbed to prematurity. Schaal and co-workers⁴¹ reported on one set of triplets complicated by the preterm delivery of two fetuses at 26 weeks. After the delivery, a MacDonald cerclage was placed, and prophylactic antibiotics were administered. The patient developed intractable preterm labor and delivered 23 days after the delivery of the first two fetuses.

Daikoku and colleagues⁴² reported on a set of triplets. The first fetus was delivered at 26 weeks and 4 days' gestation; the second and third triplet fetuses were delivered 11 days later. The later born triplets had a less morbid neonatal course. All of the fetuses survived. The authors suggested that delayed birth interval should be considered in the case of a very premature triplet gestation, that is, between 16 and 26 weeks' gestation. Arias⁴³ reported on eight gestations (six sets of twins and two sets of triplet) complicated by premature rupture of membranes. Aggressive intervention with tocolysis, antibiotics, and cervical cerclage was instituted after the delivery of the first fetus. The mean gestational age at the delivery of the first fetus was 19.6 weeks. The mean gestational age of the fetuses with delayed delivery was 26.7 weeks. The mean prolongation of a pregnancy was 48.8 days. The author concluded that his management protocol was a reasonable alternative for some patients with multifetal pregnancies and premature rupture of the membranes. Olatunbosum and colleagues⁴⁴ reported on a case of delayed-interval delivery in quadruplets. The delivery of the first infant occurred at 26 weeks' gestation. The patient was treated with bed rest and tocolysis. The second delivery occurred 8 days later. The remaining two fetuses were delivered after a 36-hour delay, and the indication for delivery was placental abruption. The first infant delivered died of complications caused by prematurity. The remaining three surviving fetuses were developing normally 1 year after birth.

If an attempt is made to prolong a multiple gestation after the preterm delivery of one of the fetuses, the patient should be carefully monitored for possible signs of disseminated intravascular coagulation (DIC) from the retained placenta of the delivered fetus and chorioamnionitis. It is not possible to draw objective conclusions or make to substantive recommendations regarding the empiric use of antibiotics or cerclage to minimize the risk of infection or to maximize gained time for the undelivered fetuses. Nevertheless, if membranes do not prolapse, an attempt to prolong gestation without cerclage should be considered. Watchful expectancy may prevent the additional risk of infection and rupture of the membranes that occasionally accompany cerclage.

Malpresentation

Higher-order gestations commonly are associated with fetal malpresentation at the moment of labor and delivery. The number of positional combinations is nine for twins and far more for higher-order gestations. Given these circumstances, in triplet gestations, the incidence of all three fetuses being vertex at the moment of delivery is not greater than 16%.¹⁶ Moreover, in greater than 30% of the cases,⁴⁵ the leading fetus is breech. Table 4 illustrates in detail the combinations of possible triplet presentation reported in the literature. The high incidence of malpresentation greatly increases the potential for birth trauma. In addition, the third triplet may experience increased morbidity and mortality compared with the first. According to Holcberg and colleagues,¹⁶ this increase may approach 40% to 50%. Commonly, the third fetus has low Apgar scores and signs of transient tachypnea and hyaline membrane disease.¹⁷ Such problems apparently also occur in vaginally delivered triplets.

TABLE 4. Possible Triplet Presentation

VTX, vertex; BR, breech; CMP, compound; Trans, transverse; N/A, not available.

Concomitant (Combined) Ectopic Pregnancy

The increasing numbers of higher-order gestations have led to reports of infrequent or rare complications. Among these are combined uterine and ectopic pregnancy. This condition is also known as heterotopic pregnancy. Forbes and Natale⁴⁶ reported on a case of a unilateral tubal triplet pregnancy. Their literature review found three similar cases. Schmitt⁴⁷ reported on a fifth case; Fujii and associates⁴⁸ and Koh and Banks⁴⁹ reported on cases of tubal quadruplet pregnancies; Treub⁵⁰ described the only known case of tubal quintuplets; and Rowland and associates⁵¹ presented a triplet intrauterine pregnancy with an ectopic tubal gestation.

A second instance of heterotopic quadruplet pregnancy was described by Sherer and colleagues⁵² in 1995. In this case, the ruptured interstitial pregnancy was resected by laparoscopy. This intervention was followed by successful delivery of triplets at 34 weeks gestational age by cesarean section.

Reports such as these strongly suggest that careful attention should be given to any complaints of adnexal pain in a higher-order gestation and that appropriate measures should be instituted to rule out combined or heterotopic gestation.

Intrauterine Growth Restriction

Intrauterine growth restriction (IUGR) is diagnosed often in higher-order gestations.⁸ A commonly accepted definition of IUGR is a birth weight that is less than 10% or below two standard deviations from the mean birth weight.

Gonen and co-workers⁵³ reported an IUGR rate of 17% in triplets and of 15% in quadruplets from their data review. Pons and colleagues⁵⁴ reported that 60% of 21 triplets studied were below the 10th percentile for birth weight. Collins and Blyel²¹ reported a 24% incidence of IUGR in quadruplets by 34 weeks, which increased to 62% thereafter. The increased incidence of IUGR may originate from the increased competition for nutrients in cases of higher-order gestation. Although this explanation is speculative, we are unaware of data that would lead to a more scientific explanation.

The competition for nutrients is exemplified by the presence of discordant birth weights. Discordancy in twins has been described as a weight difference between fetuses of 15% to 25% and occurs in 15% to 30% of twin gestations. In a review of 114 normal triplet newborns by Mordel and co-authors,⁵⁵ 58% had a birth weight discordancy of more than 15%, and about 25% had more than 25% discordancy in birth weights. These authors concluded that triplet newborns have a weight discordancy rate double that of twins. They caution that, based on their findings, discordancy alone does not always indicate an underlying complication or poor perinatal outcome.

Intrauterine Fetal Demise

The presence of IUFD of one fetus of a higher-order gestation is clearly a major complication. Because of the infrequent nature of this event, its true incidence is not known. In twins, Enbom⁵⁶ estimates that IUFD of one fetus occurs in 1 of 186 deliveries. An early series published on IUFD in triplets reported five cases. To identify risk factors for IUFD, Gonen and colleagues⁵⁷ matched these five cases to 21 sets of triplets not affected by IUFD in a case-control design. The only statistical risk factor for IUFD was the presence of a monochorionic placentation in four of the five cases. These cases also had a higher incidence of small-for-gestational age (SGA) infants, but this was not statistically significant. The mean EGA at the time of diagnosis was 29.4 weeks, with delivery of the cases occurring about 4 weeks later, at 33.8 weeks, on the average. There were no maternal cases of DIC, and none of the surviving infants had evidence of in utero DIC. The presence of DIC in association with IUFD in multiple gestations has been reported, however. Skelly and associates⁵⁸ described a case of a triplet gestation complicated by one IUFD at 18 weeks. The mother developed hematologic changes compatible with DIC, requiring treatment with plasma and later with heparin. According to these authors, a second fetus died at 35 weeks of placental insufficiency. The third fetus was delivered uneventfully by cesarean section. Thomas⁵⁹ reported on a case of a triplet gestation with one IUFD diagnosed at delivery at 33 weeks. Of the two living neonates, one had intraventricular hemorrhage and clinical DIC, and the other had mild abnormalities in its coagulation profile. Additional evidence of possible thromboembolic complications with an IUFD in a triplet gestation was presented by Markman and colleagues.⁶⁰ They reported a case complicated by a fetus papyraceous. At the time of delivery (38 weeks), one of the surviving infants was affected by aplasia cutis, which was thought to have arisen from the thrombotic material emanating from the dead fetus and placenta.

Further evidence in higher-order multiple gestations that monochorionic placentation is a risk factor for IUFD was presented by Borlum.⁶¹ In a series of 89 sets of triplets, 15 pregnancies (16.9%) were complicated by fetal demise in the third trimester. In only three patients could the continuation of pregnancy be considered after the diagnosis of fetal demise. Fetal demise was associated with monochorionic or dichorionic placentation, and IUGR was a common complication.

From the available data, it is not possible to specify at which point delivery should be attempted in these cases, or if labor should be induced. It is reasonable to consider early delivery if pulmonary maturity is attained in the living fetuses or if signs of fetal compromise become evident. If expectant management is undertaken, weekly maternal coagulation profiles may be reassuring.

Fetofetal Transfusion

The effects of monochorionic placentation are not readily apparent from the literature on triplets. A major concern is the presence of placental anastomoses analogous to those described in monochorionic placentation in twins. The effect is similar in triplet pregnancies. Unless early ultrasound can document the presence of trizygotic triplets by the identification of three separate and distinct placentas, the possibility of two fetuses sharing the same placenta is present (see Fig. 3, line 2).

Rehan and colleagues⁶² authors described a case of fetofetal transfusion syndrome in a triplet gestation. In this case, all the triplets were born alive, but two died within 2 days of delivery.

Postpartum Hemorrhage

The possibility of postpartum hemorrhage presents an additional risk. Its incidence in triplets ranges from 13%¹⁶ to 35%.³⁴ It is commonly believed that this complication originates from the inability of an overdistended uterus to contract effectively after delivery.

Preeclampsia

Preeclampsia is diagnosed commonly in triplets, with an incidence that ranges from 6.7%³⁴ to 46%¹⁶; Keith and coworkers⁶³ reported an overall incidence of 31%. In many instances, the severity of the preeclamptic condition indicates immediate delivery, even at a premature gestational age. In one series of quadruplets, preeclampsia occurred in 32% of cases.¹² Of the nine published reports on quintuplets, only one case was complicated by preeclampsia.

A combined series of 21 triplet and 8 quadruplet pregnancies was published in 1996 by Hardardottir and colleagues.⁶⁴ In this series of 29 patients, 17 developed preeclampsia (14 of the 21 triplet gestations and 3 of the 8 quadruplet gestations). Of the 16 mothers in whom delivery was induced because of preeclampsia, only 8 had hypertension before delivery. Of these, however, 10 presented with epigastric pain, visual disturbances, or headache, 9 had elevated liver enzyme levels, and 7 had low platelet counts. Of particular interest, only 3 patients of this group exhibited proteinuria, and only 6 had edema. On the other hand, 5 women had clinical evidence of the HELLP syndrome (hemolysis, elevated liver enzymes, and low platelets postpartum). All these women were normotensive before delivery. The authors concluded that, on the basis of their series, it is reasonable to contend that the presentation of preeclampsia in higher-order multiple gestations may be different than practitioners expect from their experience with singletons and twins. Stated another way, hypertension is not always the presenting sign, and symptoms consistent with severe preeclampsia and abnormal laboratory values predominate. (Emphasis added by JLZ and LK.)

Neonatal Morbidity

As noted previously, infants born of higher-order gestations are at an increased risk of premature delivery. Commonly, intraventricular hemorrhage⁶⁵ and respiratory distress syndrome complicate the neonatal course. Creinin and co-workers⁶⁶ reviewed the neonatal morbidity rates of 13 sets of triplets delivered between 1981 and 1988. A total of 80% of these infants exhibited morbidity, including hyperbilirubinemia (51.3%), hypoglycemia (30.8%), respiratory distress syndrome (28.2%), and intraventricular hemorrhage (15.4%). Of interest, all infants delivered before 34 weeks had one or more complications; however, no infant delivered after 34 weeks was affected by intraventricular hemorrhage or respiratory distress syndrome. Other reports confirm that 34 weeks is a critical point in gestation, after which morbid events decline in incidence.⁶⁷

Long-term morbid outcomes also have been described in triplets and higher-order multiple gestations. A comprehensive report on the prevalence of cerebral palsy in Japanese twins, triplets, and quadruplets was presented by Yokoyama and associates.⁶⁸ These data are of particular interest because, in Japan, at least half of all twins are monochorionic. The study subjects included 705 twin pairs (1410 twins), 296 sets of triplets (287 triplets after exclusion of one infant who died), and 7 sets of quadruplets (27 quadruplets after exclusion of one infant who died). All were born after 1977 and were recruited from the Kinki University Twin and Higher Order Multiple Birth Registry. The prevalence of cerebral palsy was 0.9% among the twins, 3.1% among the triplets, and 11.1% among the quadruplets. The risks of having one child with cerebral palsy per pregnancy increased from 1.5% in twins to 8% in triplets and 42.9% in quadruplets. Using logistic regression to adjust for confounding factors, decreasing gestational age and asphyxia were significantly associated with the risk of cerebral palsy. Infants whose gestational age was less than 32 weeks were 20 times more likely to develop cerebral palsy than infants whose gestational age was equal to or greater than 36 weeks. Figure 4 is a composite of the cumulative percentage of cases by gestational week for twins, triplets, and quadruplets.

The risk of low birth weight and prematurity is accentuated after IVF treatment. Several studies validate this, although specific details vary from study to study.^69,70

Retinopathy of prematurity generally is not described as an aspect of morbidity associated with higher-order multiple pregnancies. Its well-known relation to low birth weight and early gestational age, however, suggests that it be given appropriate attention.⁷¹ A detailed study of the clinical course and systemic correlates of this condition in quintuplets was described by Hall and associates,⁷² who found that severity of eye disease correlated with the duration of mechanical ventilation and parenteral nutrition as well as with frequency of hypoglycemia and hypercapnia.

The long-term growth of children conceived by IVF at the Monash IVF and Royal Women's Hospital reproductive biology unit was documented by Saunders and colleagues.⁷³ A total of 314 children (196 singletons, 47 sets of twins, and 8 sets of triplets) were matched to 150 non-IVF children (113 singletons, 17 sets of twins, and 1 set of triplets). All children were born between January 1, 1989 and July 1, 1991. This study failed to demonstrate an independent IVF effect on the growth and physical outcomes of children at 2 years of age when matched for plurality and gestation. Where noted, poor outcomes were related to the effect of multiple birth. Although the authors questioned the representatives of their data, these findings are of interest and need to be corroborated by further investigation.

Diagnosis

Findings suggestive of multiple gestation include discrepancy between size and date, elevated α-fetoprotein (AFP) levels, and the auscultation of fetal heart tone at two locations. Before ultrasonography was available, a diagnosis of higher-order gestation commonly was made by obtaining a radiograph of the abdomen. As late as 1980,³⁶ diagnosis at the moment of the delivery was not uncommon. Fortunately, such events are now rare because most cases follow treatment for infertility or ART, and the possibility of a higher-order gestation is never in question.

The use of ultrasonography virtually ensures early and accurate diagnosis of all higher-order gestations. Careful documentation of placental location and the number of sacs is essential (Fig. 5). Regardless of the number of sacs, it is necessary to search for evidence of monoamniotic placentation. According to Landy and colleagues,^74,75 a significant number of twin pregnancies spontaneously disappear, giving rise to the so-called vanishing twin syndrome; it also is possible that some naturally occurring higher-order gestations undergo a similar process of reduction, especially in pregnancies that have an odd number (i.e., 3 or 5). Although Allen⁷⁶ has postulated a variety of additive mechanisms by which triplets or quintuplets might occur, his analysis was made in the years before the knowledge of “vanishing,” and he considered only variations of the splitting process.

The phenomenon of a vanishing sac also has been reported in higher-order gestations conceived by IVF.⁷⁷

Higher-order gestations, like twins, have an increased incidence of congenital malformations. It is possible that some of these anomalies originate from the splitting of monovular twins. A detailed ultrasound anatomic survey should be attempted at 18 to 20 weeks' gestation. As a result of technical difficulties involved in such types of scanning, however, the need for this investigation may represent an indication for consultation with a maternal—fetal medicine subspecialist or an expert in obstetric ultrasound. If one of the fetuses is diagnosed with a malformation, the patient should be advised about the availability of selective termination.

Management

DIET.

No doubt remains about the increased nutritional demands of higher-order gestations. A weight gain of 44 pounds has been suggested for twins⁷⁸; in higher-order gestations, the appropriate weight gain remains to be established. Few data are available for the establishment of objective guidelines; reports on triplet gestations describe only the maternal weight gain. Syrop and Varner³⁴ reported on 20 triplet gestations with a mean maternal weight gain of 34.1 lb (range, 18.9 to 69.7 lb). Michlewitz and co-workers⁷⁹ reported on 15 triplet gestations in which maternal weight gain ranged from 6 to 44 lb; average gain was not reported. Both of these reports, however, encompassed several decades of patient delivery. During that time, not only did recommendations provided by physicians regarding weight gain change considerably but also social factors that influenced body image and prepregnant weights changed.

The most recent report on weight gain in triplet gestation was by Luke and co-workers.⁸⁰ This pilot study comprised a cohort of women who brought their triplet children to the Multiple Births Foundation Clinic in London and who had given birth since 1983. The study design was a retrospective, anonymous, pilot telephone survey of demographic, anthropometric, and perinatal data. Telephone interviews were conducted by a research nurse. Specific details about weight gain were obtained from the Cooperation Card retained by the patient during and after her pregnancy, which contained data on prenatal weights at each visit. This study focused not only on total maternal weight gain but also on birth weight. Of interest, better intrauterine growth was achieved when maternal weight gains were equal to or greater than 1.5 pounds per week before 24 weeks' gestation.

Collins and Blyel²¹ reported on the maternal weight gain in 70 quadruplet pregnancies; the mean maternal weight gain was 45.8 lb (range, 11 to 100 lb). Even though these gains are similar to those reported for twins, the periods during which the gain was recorded are dissimilar because quadruplets are delivered up to 4 to 6 weeks earlier than twins, who are delivered, on average, at 36 to 37 weeks. If matched for gestational age, even large gains undoubtedly would be recorded for higher-order gestations. It is important that such gains not be viewed as abnormal. Only the rapid weight gain associated with various stigmata of preeclampsia should be of concern.

Early in pregnancy, consultation with a registered dietitian should be requested to maintain a well-balanced diet and to ensure the adequate and early weight gain that is necessary to assist the early developmental stages of gestation. In addition, prenatal vitamins are recommended. The expected high incidence of anemia associated with higher-order gestations dictates that iron supplementation be started at the initial prenatal visit, even if the hemoglobin level is normal at that time. We suggest 300 to 600 mg/d of oral ferrous sulfate. Red meat is the best and most palatable form of dietary iron supplementation because it provides iron in a bioavailable form without gastrointestinal side effects otherwise associated with iron supplementation. One-half pound of red meat per day should be the minimum for women with higher-order gestations.

ULTRASOUND.

Ultrasonography is essential for the early diagnosis of higher-order gestations; it also is an essential part of antenatal management for these pregnancies. Only with the use of serial ultrasonography is it possible to detect fetal growth abnormalities or fetus-to-fetus transfusion if monochorionic placentation is present.

The use of triplet-specific ultrasound growth tables is not without controversy. There are few reports on this issue⁸¹ because of the difficulty of obtaining data on large numbers of higher-order gestations for the creation of such tables. Weissman and colleagues⁸¹ performed serial ultrasonography in 24 triplet gestations between 1978 and 1987 and documented slowing of the biparietal diameter, head circumference, and abdominal circumference starting at 28 weeks. However, the head circumference/abdominal circumference ratio remained normal, possibly suggesting a pattern of symmetric growth retardation. Nevertheless, femur growth was assessed to be parallel to that seen in singletons. The authors suggested the use of their triplet-specific growth tables for the detection of IUGR. In contrast, Jones and co-workers⁸² reviewed 196 sets of triplets (from the same database as Newman and colleagues¹⁸) and were not able to detect any distinct slowing of the birth-weight growth curves. They plotted a linear curve of growth between 22 and 38 weeks.

Using the Rossavik growth models, Hata and associates⁸³ compared the fetal growth of 7 triplet, 13 twin, and 20 singleton gestations. Their findings demonstrated growth potential realization in the values for triplets that were similar to those for singletons and twins. These authors concluded that individual assessment of growth in triplets can be performed with the same methods generally used for both singletons and twins. In contrast, Yuval and colleagues⁸⁴ reported on the intrauterine growth of triplets as estimated from live-born birth-weight data. Their study was based on 109 sets of triplets. The estimated intrauterine growth of triplets, in contrast to that of singletons, exhibited neither an acceleration phase during the third trimester nor a flattening of the growth curve during the last few weeks of the third trimester. The mean birth weight of the triplets in this study was slightly below the 10th percentile for singletons at 38 weeks' gestation or later. The authors concluded that the growth of triplets as estimated from live-born birth weights is slower that of singletons. They further recommended that the diagnosis of intrauterine growth in triplet pregnancies should be based on triplet gestation-specific growth curves.

Only one report has been published on the accuracy of the estimation of fetal weight in multiple gestations. Lynch and co-workers⁸⁵ evaluated 1832 singleton, 518 twin, and 51 triplet fetuses. Their analysis concluded that ultrasound estimation of fetal weight is as accurate in twins and triplets as it is in singletons.

Ultrasonographic determination of chorionicity is of great importance because it helps assess the likelihood of morbidity and mortality in multiple gestations. Using transvaginal ultrasonography before 14 weeks' gestation, Monteagudo and colleagues⁸⁶ reported on 43 gestations (40 twin and 3 triplet) with confirmatory placental pathologic evaluation. In all cases, transvaginal ultrasonography correctly predicted the chorionic and amniotic type that subsequently was confirmed by placental pathology. The ultrasonographic findings of interest were the following: if a wedge-shaped junction was seen, it was considered to be the fusion of two chorionic membranes; if a T-shaped junction was seen, it was considered to be the fusion of the two amnions.

All higher-order multiple gestations should be evaluated by targeted ultrasonography at an appropriate age. Pryde and co-workers⁸⁷ stated that the likelihood of having at least one malformed fetus in triplet gestation is about 9%, whereas the incidence of two or three simultaneously but discordantly malformed fetuses in a multizygotic triplet gestation is 0.09% and 0.0027%, respectively; these authors reported the first of such instances.

MATERNAL α-FETOPROTEIN AND GENETICS.

Few data are available on the normal values of maternal serum AFP in higher-order gestations. Wald and co-workers⁸⁸ suggested values in triplets of up to 3.2 times those in singletons at matched gestational ages and values in quadruplets of up to 5 times those in singletons. No normative tables have been published for the interpretation of AFP values in higher-order gestations, and it is not possible to use this screening test for accurate detection of neural tube defects in these pregnancies. If this test is performed, consultation with a perinatologist or a geneticist is advised for the interpretation of results.

The indications for genetic evaluation of a fetus are numerous, the most frequent being advanced maternal age. This state also is an indication for amniocentesis in a higher-order gestations. Family history or a prior infant affected by a chromosomal problem also may be an indication for amniocentesis. In the event that amniocentesis is performed, indigo-carmine should be instilled in each sac to prevent multiple sampling of the same sac.

Prevention of Premature Labor

CERCLAGE.

Elective cerclage has been evaluated in triplet pregnancy^17,89 and appears to be without substantial benefit in the management of these cases. Newman and co-workers¹⁸ reported on several patients treated with cerclage. It is not clear from this report, however, whether these patients had cervical incompetence or whether suture placement was elective. The outcome of operated patients was poorer, with an average EGA of 29.2 weeks at delivery, than the nonoperated EGA of 33.6 weeks at delivery. Understandably, the indications for cerclage placement may have contributed substantially to this unsatisfactory outcome. Similar experiences have been reported by Lipitz and colleagues¹⁷ and by Itzkowic.⁸⁹

A controlled evaluation of 35 women with triplet conceptions hospitalized in the high-risk pregnancy unit at the Hebrew University Hadassah Medical School was described by Mordel and co-workers.⁹⁰ Twelve of the 35 women underwent elective cerclage at 12 to 14 weeks' gestation; 23 served as controls. None of the study subjects had historical or physical findings indicative of cervical incompetence. The gestational age (in weeks) at delivery was 33 ± 5.1 and 34.7 ± 2.8 (mean, ±standard deviation) for the sutured and control groups, respectively. Birth weights were 1833 ± 524.5 and 1884 ± 455.5 for the sutured and control groups, respectively. Despite the retrospective nature of the analysis, the authors believed the rigid criteria for case inclusion (i.e., absence of history or physical signs of cervical incompetence) strengthened their conclusion that triplet gestations did not benefit by late first-trimester or early second-trimester cerclage.

TOCOLYTICS.

Tocolytic therapy often is required to treat preterm labor. The armamentarium of tocolytics includes β-mimetics (e.g., ritodrine, terbutaline, salbutamol), magnesium sulfate, and indomethacin. These agents have been used with varying results.

Prophylactic Tocolytics.

Ron-El⁹¹ compared eight patients who received β-mimetic prophylaxis with six who did not and could not document improvement in EGA at delivery after prophylactic use of these agents. Similarly, Newman and co-worker¹⁸ reported on 115 (58.1%) of 198 sets of triplets who received prophylactic β-mimetics. No improvement was apparent in any outcome variables after therapy. Interpretation of data from these two reports does not support the efficacy of prophylactic β-mimetic tocolysis in triplet gestation. The same caution might also apply to other types of higherorder gestations, but data are limited.

Therapeutic Tocolysis.

Whereas data on prophylactic tocolysis is severely limited, the impact of therapeutic tocolysis on higher-order gestations has been reported more frequently. Itzkowic⁸⁹ reported an incidence of preterm labor of 78% (46 of 59 triplet sets). Of these cases, seven patients received salbutamol in an attempt to stop their labors. Of these seven patients, four delivered within 36 hours, and three delivered at least 3 weeks later. No mention was made of any side effects from the treatment.

Newman and co-workers¹⁸ reported that more than half (66.2%; 131 of 198 triplet sets) of cases in their series was complicated by preterm labor, with a mean EGA at diagnosis of 29.9 weeks. No description of the type of tocolytic agent or the administration regimen was provided. Also, no mention was made of complications from this therapy.

Holcberg and co-workers¹⁶ reported an incidence of preterm delivery of 97% (30 of 31 triplets). Five patients were treated with ritodrine, with an average gain of 2.6 weeks before delivery. Again, no mention was made of side effects.

Keith and associates⁴⁵ reported on 13 sets of triplets, of whom two thirds required tocolysis in the form of ritodrine or magnesium sulfate. No significant complications from this therapy were reported. Yet another study, by Vervliet and colleagues,³⁵ reported on 15 triplets and 6 quadruplets who were treated with intravenous ritodrine. In addition, some patients were treated with rectal indomethacin (Indocin) until 34 weeks' gestation. No mention was made of oligohydramnios as a sequelae of indomethacin therapy.

Elliott and Radin⁹² reported on 12 triplet and 4 quadruplet gestations. These patients received magnesium sulfate for tocolysis. Higher serum levels of magnesium sulfate (7.0 to 7.5 mg/dL) were necessary to achieve adequate cessation of uterine activity. These authors also concluded that rates of infusion of magnesium sulfate of 4 to 5 g/h were necessary in triplets and quadruplets to achieve cessation of uterine activity. These observations, if confirmed by others, have important clinical implications: namely, that the present low levels of success in prolonging gestation in higher-order multiple gestations may be improved with a more aggressive approach toward medication dosage. Because most medications available for tocolytic therapy have substantial side effects, controlled studies are warranted to document the pharmacokinetics of each agent in higher-order multiple gestations.

Although the total number of reported patients who have received therapeutic tocolysis for preterm labor in higher-order gestations is modest, we believe that β-mimetics and probably magnesium sulfate have therapeutic value. The presence of contractions alone is probably an insufficient indication to start treatment, however. A diagnosis of preterm labor should include documentation of cervical change. Utter and co-workers,⁹³ using data from singleton pregnancies, confirmed that the success of tocolytic therapy is not hampered by waiting for cervical change to strengthen the diagnosis of preterm labor. We further recommend that if tocolysis is used, the clinician should be prepared to handle any possible side effects of the medication, including maternal tachycardia, pulmonary edema, myocardial ischemia, hyperglycemia, and electrolyte imbalances. Data from twin gestations suggest a higher incidence of side effects when compared with singletons.⁹⁴ A similar assumption can be made with regard to higher-order multiple gestations.

INDUCTION OF PULMONARY MATURITY.

Little information exists on this topic with regard to higher-order multiple gestations. Holcberg and colleagues,¹⁶ Ron-El,⁹¹ and Newman and co-workers¹⁸ all described several patients in their respective series who received steroids for the induction of surfactant production by the fetuses. The small numbers of patients treated, however, precludes making a comprehensive evaluation of the effectiveness of this treatment. Pons and co-workers⁵⁴ reported that 10 of their 21 sets of triplets received steroids. None of the newborns of mothers who received steroids developed respiratory distress syndrome. Of the group that did not receive steroids, however, 21% of newborns developed respiratory distress syndrome. No maternal side effects were noted.

The administration of steroids may result in an increase in uterine activity. Elliott and Radin⁹⁵ reported on 10 triplet and 5 quadruplet gestations that received at least one course of steroids for the induction of pulmonary maturity. A significant increase in contractions, a statistically increased likelihood of labor with cervical change, and the need for tocolysis were noted when the basal contraction rate was at least 3.5 contractions per hour. These investigations concluded that there should be fewer than 3.5 contractions per hour to minimize the steroid effect on uterine activity.

If the clinician caring for a higher-order multiple gestation decides to use steroids, careful monitoring is urged, especially if the patient is being treated concomitantly with β-mimetics. An increase in cardiovascular complications, especially pulmonary edema, has been reported in singletons and twins.

BED REST.

During the past several decades, bed rest has been the subject of intense discussion in terms of its value for the prevention of preterm labor and the improvement of birth-weight outcomes in multiple gestations. In the triplet literature, some authors have advocated early hospitalization for bed rest.^16,96,97 Others¹⁸ have suggested decreased physical activity with bed rest at home, limiting hospital admissions to the treatment of premature labor or significant obstetric or medical complications.

One of the few studies that tried to assess the quantity of activity in triplet gestations³⁴ documented a higher incidence of IUGR and preterm delivery among patients with “unrestricted” activity compared with patients with “restricted” activity. The EGAs at delivery were 31.3 weeks' gestation and 34.3 weeks' gestation, respectively. This retrospective series included patients who were prescribed hospital bed rest as well as limitation of activity at home (the group with restricted activity). If no mention was made in the patient's record regarding limitation of physical activity, she was classified as unrestricted. Of the restricted activity group, 1 infant of 39 had IUGR. In contrast, 67% of the infants of the unrestricted activity group had IUGR.

In a retrospective study of 105 sets of triplets delivered between 1975 and 1984 in Zimbabwe, Crowther and Hamilton⁹⁷ reported on the impact of bed rest. Patients managed with bed rest in the hospital had less PMR than the group managed as outpatients.

A more liberal approach was reported by Newman and co-workers,¹⁸ which entailed restricting activity at home and limiting hospital admission to strict indications. A total of 44% of their patients required admission for obstetric complications, and 66.2% had preterm labor requiring admission. The EGA at delivery for this group of 198 sets of triplets was 33.6 weeks, with a perinatal survival of 95%.

The largest published series of quadruplets²¹ was unable to provide an in-depth analysis of the impact of bed rest, but it was almost universally applied.

We recommend that decreased physical activity be applied universally to higher-order multiple gestations after 20 to 24 weeks, with hospital admission indicated only by a complication and not for prophylaxis of preterm labor.

HOME MONITORING OF UTERINE ACTIVITY.

The technique of home monitoring of uterine contractions and the concomitant use of telephone lines to relay the information obtained is being evaluated for its utility in the diagnosis of preterm labor that may not be apparent to the patient or her clinical care givers. Because of the acknowledged risk of preterm labor in higher-order multiple gestations, this technique might logically be considered as having a key role in the early recognition of such a complication.

In the report by Newman and co-workers,¹⁸ all patients were managed with this technology. Outcomes in this study, however, were not different than those of other recent series in which this system was not used universally or was not used at all.¹⁷

We encourage clinicians to evaluate each case separately. Until additional studies help define the possible role of home monitoring in the recognition of preterm labor and the prevention of premature birth, it is too soon to draw conclusions.

Assessment of Fetal Well-Being

In addition to serial ultrasonography and the routine prenatal blood tests that are performed on all gravidas, higher-order multiple gestations require antenatal surveillance in the form of the nonstress test. The use of this test in singletons and its predictiveness for well-being have been well documented.⁹⁸ The lack of similar data from higher-order gestations undergoing surveillance makes it difficult to determine accurate sensitivity, specificity, and negative and positive predictive values in the detection of higher-order gestation fetal compromise. For example, it is not possible to state with certainty when surveillance should be started; between 30 and 32 weeks appears to be reasonable. A nonreactive nonstress test should be pursued in a similar fashion as in singletons, with the exception of the contraction stress test, which probably should not be included as an alternative because of the possibility of triggering premature labor. The biophysical profile is a reasonable alternative to the contraction stress test. The results of an abnormal biophysical profile should be interpreted as for singleton pregnancy. Some clinicians use the biophysical profile as their routine antenatal surveillance test. The nonstress test may be impractical in all instances because of the difficulties of simultaneous assessment in three or more fetuses.

Elliott and Finberg⁹⁹ reported on 18 triplet and 6 quadruplet gestations monitored with biophysical profiles. There were no stillbirths in this series; however, 25% of these gestations were delivered because of nonreassuring biophysical profile testing, all with good outcomes. Four gestations had poor neonatal outcomes in the presence of normal biophysical testing. All of these late pregnancies apparently had acute changes in their clinical status. These investigators concluded that the biophysical profile is a reliable technique for the assessment of fetal well-being in higher-order multiple gestations.

Umbilical Doppler blood flow analysis has received a great deal of attention. The use of umbilical Doppler for surveillance has been reported in three series of triplets. Giles and co-workers¹⁰⁰ reported on 20 sets of triplets. Using continuous-wave Doppler, these authors were able to predict all the cases of SGA as well as all stillborns. Similar findings were reported by Rafla.¹⁰¹

Gaziano and colleagues reported on 94 twin and 7 triplet gestations.¹⁰² Abnormal Doppler findings were associated with delivery 3 to 4 weeks earlier and increased morbidity and mortality when compared with fetuses with normal Doppler results.

Because of the high incidence of SGA in higher-order gestations, this technique may play an important role in the surveillance of these patients.

The delivery of higher-order multiple gestations presents great challenges to obstetricians. Most of these pregnancies are complicated by malpresentation, the need for either abdominal or vaginal operative delivery, and the potential need for immediate specialized care for one or more newborns. Although it is clear that triplets and higher-order gestations can be delivered per vaginum, the published US literature^18,21 shows a preference for abdominal delivery, whereas European clinics^35,67 tend to prefer vaginal delivery, at least for triplets. Newman and co-workers¹⁸ reviewed published series of triplets in which the rate of cesarean section delivery ranged from 14% to 94%.

Pheiffer and Golan⁹⁶ reported on 61 triplets delivered in South Africa between 1967 and 1976. The fact that 60% of cases were diagnosed during labor may have contributed to the low (14%) cesarean section rate. Only two of the patients in this series were scheduled for elective cesarean section (because of prior section), but both presented in active labor and delivered vaginally. The sole cesarean section was performed for obstructed labor. Of the vaginal deliveries, however, manipulation was required in most cases (67.6%). The amount of morbidity or mortality generated by these vaginal operative deliveries cannot be assessed from the data in the report; however, the authors recommended elective cesarean section for all triplets when the leading twin is not vertex.

Lipitz and associates,¹⁷ in their series of 78 triplets delivered between 1975 and 1988, reported that 78% of patients delivered by cesarean section. Their data suggested that, after 26 weeks' gestation, no difference in perinatal mortality between vaginal and abdominal delivery was discernible. When morbidity parameters were evaluated, however, it became apparent that the vaginally born third triplets had a higher incidence of low Apgar scores and respiratory disorders. The final conclusions of this report were based heavily on these latter findings, and the authors stated, “we find little justification to attempt vaginal delivery in these high-risk patients.”¹⁷

Olofsson⁶⁷ reported on a series of 14 triplet and quadruplet pregnancies and recommends vaginal delivery for triplets from 34 weeks onward, even when the leading twin is breech. Using this protocol, 6 of 14 sets of triplets underwent a successful vaginal delivery. Another 3 pregnancies underwent emergency cesarean for fetal distress during a vaginal delivery.

Loucopoulos and Jewelewicz¹⁰³ reported on 27 deliveries of triplets, 7 of quadruplets, and 1 of quintuplets. The cesarean section rate was 42%. When the outcome of triplets was compared, the only improvement provided by cesarean section over vaginal delivery was an improvement of the 1-minute Apgar score for the third-born triplet (6.8 versus 5). Other reports suggest a higher mortality rate for breech triplets delivered vaginally.¹⁰⁴

The largest triplet review to date is by Newman and colleagues¹⁸; this study included 198 sets delivered in 24 regional centers. A total of 186 (93.9%) of patients delivered by cesarean section. Of the 12 attempts at vaginal delivery, 3 patients underwent emergency cesarean section after the vaginal delivery of the first triplet because of intrapartum complications.

Because of the controversy about the ideal method for the delivery of triplets, Feingold and associates¹⁰⁵ designed a historical cohort study to evaluate the merits of both possibilities. The study group (delivered by cesarean section) had significantly higher Apgar scores at 1 and 5 minutes. In addition, when combined mortality and morbidity were compared among first-, second-, and third-born triplets, no differences were found within the study group; in contrast, in the control group, the combined morbidity and mortality rate increased from 21% in the first-born triplet to 31% in the second to 43% in the third. Because of these findings, a more liberal approach toward cesarean section was advocated. Confounding aspects of this study were selection bias and the possibility that different levels of medical care had been afforded the two groups. This latter consideration was real because the control group was delivered during an earlier period than the study group (1954 to 1976 versus 1977 to 1986, respectively), and the concepts of prenatal and intrapartum care had changed greatly in this 20-year interval.

A case control approach was described by Dommergues and colleagues¹⁰⁶ at the Maternite Port Royal in Paris. A vaginal delivery was attempted in 23 otherwise uncomplicated triplet pregnancies that formed the study group. These patients were compared with 23 controls undergoing elective cesarean section for delivery and matched for gestational age at birth. Neonatal mortality did not differ between the groups (1 in 69 versus 0 in 69). Infants delivered vaginally had significantly higher Apgar scores (9.5 versus 8.4) and significantly shorter number of days (overall) in the NICU (6 versus 18). Although differences in length of stay for the three triplets did not achieve statistical significance, the raw numeric differences were impressive.

An additional assessment of vaginal delivery was described by Clarke and Roman¹⁰⁷ in New Zealand. In this study, vaginal delivery was conducted for the first triplet in five cases. After delivery of the first child, general anesthesia was induced, and the respective second and third infants were born by breech extraction or internal version followed by breech extraction. Although the 15 infants survived, the authors failed to commit themselves on heart rate monitoring after the delivery of the first infant. Such an approach may not be looked on with favor by most clinicians practicing in North America, where the possibility of litigation remains high. Furthermore, the induction of “crash” general anesthesia is not without its own hazards, and it is possible that the same outcome might have been achieved using regional anesthesia from the outset. With either type of anesthesia, the possibility of abruption after or associated with intrauterine manipulations is real, and this risk provides additional support for the use of continuous fetal monitoring.

Another approach to monitoring has been described by van Eyck and Arabin.¹⁰⁸ These investigators used an advanced model of the MT430 Actocardiotocograph (Toitu Company Limited, Tokyo) to obtain optimal monitoring of the three fetal heart tracings simultaneously during labor and delivery.

Two groups of investigators evaluated the biochemical parameters of triplets at term. Antoine and associates¹⁰⁹ reported on six sets of triplets delivered by cesarean section. They compared Apgar scores, umbilical artery and venous pH, PO₂ and PCO₂ levels, lactate levels, and base deficits in the third-born triplet with those measures in the first- and second-born triplets. No differences were found. Creinin and co-workers¹¹⁰ performed a similar analysis on 11 sets of triplets, only one of which was delivered vaginally. Even though the 1-minute Apgar score was lower in the group of third-born triplets, no difference was seen in the 5-minute Apgar score, nor in any of the umbilical artery or venous cord gas analyses. Nevertheless, the only acidotic pH (7.08) occurred in the third-born triplet of the only set born vaginally.

For quadruplets and higher-order gestations, vaginal deliveries have been reported,² but the problems associated with them suggest that elective cesarean section may be the ideal route of delivery.

Table 4 reviews data from four studies and lists the possible presentations for triplets. The vertex-breech-breech combination was most common (18.1%), followed by vertex-vertex-breech (16.4%).

Aside from malpresentation, a broad spectrum of obstetric complications faces the obstetrician attempting vaginal delivery of a higher-order gestation. Cord prolapse,¹¹¹ fetal heart rate deceleration, and placental abruption with hemorrhage are all possible. An additional and crucial factor in favor of delivering all higher-order gestations by elective cesarean section is the nature of the neonatal team that must be assembled. Usually, such teams are staffed by one pediatrician-neonatologist, two nurses, and one respiratory technician for each infant. Additional support may be required if the delivery is preterm. An elective, scheduled cesarean section permits sufficient time to assemble such a team to allow for optimal conditions at the moment of delivery, including the provision of three beds in the NICU or nursery.

The cesarean section should be executed by a team of experienced obstetricians capable of performing a hysterectomy in case of an intractable postpartum hemorrhage. Moreover, an experienced anesthesiologist with sufficient support staff should be available. Epidural or general anesthesia is appropriate for this operation. Blood should be readily available in case of severe hemorrhage. Uterotonics, such as ergonovine derivatives or injectable prostaglandins, also should be available in case of hemorrhage.²³

The advantages and disadvantages of transverse lower segment incision compared with longitudinal uterine incision were described by Mordel and associates.¹¹² These authors concluded that the transverse incision may be preferable to a longitudinal incision, especially in women who desire future pregnancies in which a vaginal birth would be considered. No differences were observed in the outcome parameters that were studied.

If, after considering all factors, a decision is made to attempt vaginal delivery, an equally prepared neonatal team should be assembled for each child. The delivery should not be conducted by a single obstetrician. At least two scrubbed obstetricians should be present, and both should be proficient in performing intrauterine manipulations on a malpresenting fetus. A third obstetrician should be present to perform real-time ultrasonography during the second stage of labor for planning the delivery and to assist with manipulation if required.

As prior sections have indicated, higher-order gestations are at an extremely high risk for complications and morbidity. The single most important risk factor is the virtual inevitability of preterm labor. The specter of very early preterm delivery followed by prolonged admission to a NICU and all the possible sequelae thereof have led some clinicians to offer their patients with higher-order gestations the possibility of “selective reductions,” even in the absence of known genetic disease or anatomic deformity demonstrated by ultrasound. As noted previously, the PMR of triplets is about 50 in 1000.^17,18 This figure compares favorably with that of twins.¹¹³ Thus, an argument could logically be made not to reduce triplets, and this position has been taken publicly, both in the United States and Europe. It should be remembered, however, that the PMR of 50 in 1000 associated with triplets is five times the PMR reported for singletons.

The term selective reduction means the selective termination of one or more embryos, using a transabdominal or transvaginal approach, under careful real-time ultrasonographic monitoring. The operation usually is performed late in the first trimester or early in the second trimester. Commonly, no more than two embryos are terminated at one session, and occasionally a second intervention is indicated.

Review of the Technique

This procedure ideally is performed in an outpatient setting. Some authors advocate antibiotic prophylaxis (either ampicillin or a cephalosporin) given as a single intravenous dose before the procedure. Using real-time ultrasonography, all fetuses are identified. The fetuses closest to the abdominal wall usually are selected for termination. If any monoamniotic sets of fetuses are identified, they are preferentially terminated because of the inherent risks of such chorionicity. Similarly, if discordancy is detected, the smaller fetus is terminated. Obviously, any fetus with ultrasonographic findings suggestive of a malformation should be considered for termination.

After cleaning the skin with a disinfectant solution, the area is infiltrated with local anesthesia using a sterile technique. A 22-gauge spinal needle is inserted through the maternal abdominal wall and into the sac of the embryo closest to it. The needle is guided into the fetal chest as close as possible to the heart.

Although a variety of techniques to achieve cardiac standstill have been reported, the injection of KCl into the heart is preferred. The amount required ranges between 2 and 5 mEq of KCl, using a concentration of 2 mEq KCl/mL. Once cardiac standstill is evident, the needle is left in place for at least 3 minutes. This is a safeguard against the possibility of the reappearance of cardiac activity requiring another pass through the abdomen to reinject the KCl solution. Also, waiting a few minutes after the standstill is first seen allows the ultrasonographer to visualize the heart from different planes and, in that manner, to confirm the cardiac standstill. If more than two embryos are to be terminated, or if after the procedure one of the supposedly terminated embryos resumes cardiac activity, a second reductive attempt can be performed 1 week after the first.

After the procedure is completed, the patient is observed for complications, including bleeding, spontaneous abortion, rupture of membranes. If no complications are detected, she can be discharged home the same day of the procedure.

Background

The initial reports on selective termination were published in 1981.¹¹⁴ These and those that followed described the reduction of twin to singleton pregnancies because one twin was determined to be affected by a hereditary condition, such as thalassemia major, Tay-Sachs disease, or trisomy 21. To our knowledge, the first reported case of a truly selective termination of a higher-order pregnancy was published in 1986, on a quintuplet pregnancy.¹¹⁵ Additional case reports were published in 1987^116,117 and thereafter.

Several publications have reviewed the experience of the world's largest centers performing multifetal pregnancy reduction. Evans and co-workers¹¹⁸ reported on 463 completed pregnancies having undergone multifetal pregnancy reduction. There was a 100% technical success, with no failed procedures. Of these patients, 83.8% had a potentially viable delivery, and 83.5% delivered at 33 weeks or later. The risk of fetal loss was 3.9% at 2 weeks after the reduction or earlier. This comprehensive analysis suggests that multifetal pregnancy reduction is an effective and safe technique that can improve perinatal outcome in higher-order multiple gestations by eliminating the high risk of preterm delivery in untreated patients. In this series, the improvement was unquestionable for quadruplets or higher-order pregnancies but was arguable for triplets. No maternal complications followed the procedure. The major risk was that of fetal loss. Surviving fetuses were not damaged by the reduction. The authors stressed the importance of parental autonomy in the process of making an informed decision.

In a subsequent analysis by Evans and co-workers¹¹⁹ of 183 completed cases of selective termination from nine centers in four countries, 83.8% of viable deliveries occurred after 33 weeks' gestation, and only 4.3% occurred between 25 and 28 weeks' gestation. Once again, no maternal morbidity was reported, nor was there evidence of coagulopathy or ischemic damage in the surviving fetuses. The gestational age at the moment of performing the reduction correlated positively with loss rate and inversely with gestational age at delivery. The authors emphasized the need for early diagnosis in multifetal pregnancies so that adequate counseling on multifetal pregnancy reduction can be performed.

In a large series of 400 patients reported on by Berkowitz and colleagues,¹²⁰ pregnancy loss rates after multifetal pregnancy reduction were 7.3% for triplets, 8.4% for quadruplets, 6.1% for quintuplets, and 17.6% for sextuplets or higher-order gestations. In a prior study by Berkowitz and associates,¹²¹ 200 gestations that underwent multifetal reductions were reported. The analysis of the data suggested that the incidence of IUGR in the reduced gestations was not increased over that expected in nonreduced twin gestations.

Four studies have been done on the multifetal reduction of triplets to twins and on the impact of this procedure on perinatal morbidity and mortality. Porreco and associates¹²² reported on 13 triplet gestations that underwent multifetal reduction and 11 triplet gestations in which the parents declined or were not offered multifetal reduction. In their analysis, multifetal pregnancy reduction for triplet gestations did not improve pregnancy outcome. Macones and co-workers¹²³ compared triplet gestations that were reduced to twins to the perinatal outcome of nonreduced triplets and to the outcome of nonreduced twins. The mean gestational age at delivery for triplets that were reduced to twins was 35.6 weeks, compared with 31.2 weeks in the group of triplet gestations that were not reduced. The PMR was 30 per 1000 births for the triplet gestations reduced to twins and 210 per 1000 births for the triplet gestations that did not undergo reduction. These findings were statistically significant. There were no statistically significant differences between the reduced and nonreduced twins.

Lipitz and colleagues¹²⁴ reported on 140 triplet gestations, of which 34 underwent multifetal pregnancy reduction to twins. These authors had similar findings of improved pregnancy outcome in the triplets that were reduced to twins compared with the triplet gestations that did not undergo the procedure. The mean gestational age at delivery for the triplets that were reduced to twins was 36.7 weeks, compared with 33.5 weeks for the nonreduced triplets. No neonatal mortality rates were reported for the triplets that underwent the multifetal reduction, whereas the neonatal mortality rate in the nonreduced triplets was 4.6%. This later comparison did not achieve statistical significance, but it must be considered clinically significant. In a more recent study by Lipitz and co-workers,¹²⁵ 43 triplet gestations that were reduced to twins were compared with 134 twin gestations. Analysis of these data suggested that the outcome of triplet gestations reduced to twins did not differ significantly from that of nonreduced twin gestations. After reviewing the data of the four reports cited, one can conclude that a reduction from triplets to twins appears to decrease perinatal morbidity and mortality. Some authors have concluded that it is justifiable to continue to offer this technique to patients with triplet gestations until more definite data on long-term outcomes is available for analysis.¹²⁰

Depp and associates¹²⁶ evaluated the fetal growth after multifetal pregnancy reduction of higher-order gestations to that of twins. These investigators reported on a total of 236 gestations (113 triplets, 89 quadruplets, and 34 quintuplets or higher-order gestations). This group was compared to a control group of nonreduced twins. The incidence of IUGR was 19.4% in the nonreduced twins, 36.3% in the pregnancies reduced from triplets, 41.6% in the pregnancies reduced from quadruplets, and 50% in the pregnancies reduced from quintuplets or higher-order gestations. There was a significant trend toward an increased frequency of IUGR with increasing starting fetal number. The authors concluded that the reduced higher-order gestations had a higher incidence of IUGR than that encountered in the nonreduced twins. This finding should justify a more frequent use of ultrasonography in the evaluation of fetal growth in reduced higher-order gestations.

Clinicians should realize that there are significant emotional challenges for the couples undergoing multifetal pregnancy reduction. Many of these couples are long-standing infertility patients who find this procedure completely opposite to their reproductive desires. Furthermore, many couples experience significant moral and ethical dilemmas when considering this technique. It is advisable that adequate psychological evaluation and counseling be initiated before the procedure. In addition, the clinician should discuss thoroughly with the patient the technical aspects of the procedure. Patient autonomy is of utmost relevance when considering a multifetal pregnancy reduction.

Porreco and co-workers¹²⁷ routinely offer a questionnaire to patients undergoing multifetal pregnancy reduction that was designed to evaluate the various psychological issues related to this intervention. In a report of 31 patients who underwent multifetal pregnancy reduction and 13 who declined it, these investigators found significant financial concerns to be an important factor in the decision to undergo this procedure. Nonetheless, almost one third of the patients who underwent multifetal pregnancy reduction considered themselves to be uncomfortable or very uncomfortable with their decision during the course of the pregnancy. None of the patients who declined multifetal reduction reported such emotions. Furthermore, patients who underwent multifetal pregnancy reduction reported persistent anxiety regarding their decision and the fate of the terminated fetuses. These emotions appear to be similar to the normal grieving process after a perinatal loss, with one important exception. There is an exceptionally strong potential for “anniversary” depression for the simple reason that each birthday of surviving infants is tangible evidence of the deaths of the others. In singleton perinatal loss, no such tangible circumstance exists, and grieving may be silent and unobserved.

In a series of 91 of the first 100 multifetal pregnancy reductions performed at the Mount Sinai Medical Center, Schreiner-Engel and co-workers¹²⁸ reported more than a 65% incidence of acute feelings of emotional pain, stress, and fear during the procedure. Seventy percent of mothers mourned for the lost fetuses, with a mean grieving period of 3.2 months. An anniversary grief reaction tied to the date of the reduction was experienced by 37% of these patients. The grief and emotional reaction to the procedure were related to the frequency of the patients viewing the fetuses on ultrasound. Nonetheless, 93% of the patients stated that they would again make the decision to undergo a multifetal pregnancy reduction.

As these two articles suggest, it is of great importance that careful attention be given to the emotional status of the couple undergoing the procedure of multifetal pregnancy reduction.

Summary

It appears that selective termination of pregnancy is a procedure that carries a low potential for morbidity, with most of the reduced pregnancies going on to deliver at, or near, term twin gestations. In our opinion, the parents of all higher-order multiple gestations should be counseled about all the possible complications associated with the gestation and should be provided information about the availability of selective terminations. It should be emphasized that this procedure is a safe technique in experienced hands but that there will always be a small risk for losing the nonreduced fetuses.

During the past several decades, there has been a continuous decrease in the mortality and morbidity associated with higher-order multiple gestations. Nonetheless, these gestations remain at 5 to 10 times higher risk of complications than do singletons. Once a higher-order gestation has been diagnosed and the patient has been counseled regarding the possibility of undergoing a selective termination, if the patient wishes to continue with the higher-order gestation, decreased physical activity should be instituted, with home bed rest to start at 20 to 24 weeks. Frequent prenatal visits should be the rule. Serial ultrasound examinations (every 3 to 4 weeks) must be done to detect any growth abnormalities. If diagnosed, additional antenatal surveillance is indicated. Premature labor, once diagnosed, should be treated vigorously because prematurity is the highest cause of morbidity and mortality in these cases. At the time of delivery, a primary cesarean section is indicated. The only possible exception is a triplet gestation with all fetuses in a vertex presentation and meeting the additional requirements delineated in the text.