- Email: [email protected]
Outcome of twin gestations with a single anomalous fetus
Outcome of twin gestations with a single anomalous fetus James M. Alexander, MD, Ronald Ramus, MD, Susan M. Cox, MD, and Larry C. Gilstrap III, MD Dallas, Texas OBJECTIVE= Our goal was to determine whether the presence of one anomalous fetus in a twin gestation affects pregnancy outcome when compared with twin pregnancies without fetal anomalies. STUDY DESIGN: Maternal and neonatal data from 970 twin pregnancies delivered from 1988 to 1995 were collected. Three groups of twin gestations were identified: one fetus with a major anomaly (n =18), one fetus with a minor anomaly (n = 38), and both fetuses without anomalies (n = 914). RESULTS= Maternal demographic characteristics (age, race, and antepartum complications) were similar among the groups. There was no difference in neonatal outcome (gestational age at delivery, birth weight, cord pH, sepsis, and death) in the minor anomaly and no anomaly groups. There were significant differences between the major anomaly group and the no anomaly group in gestational age at delivery (32.9 vs 35.6 weeks, p < 0.05), birth weight at delivery (1759 vs 2291 gm, p < 0.05), hospital days (41 vs 13 days, p < 0.05), and perinatal death of the anomalous fetus-(278/1000 vs 10/1000). Except for total days in the hospital, there was no difference in neonatal morbidity or mortality for the normal fetus when compared with the minor group or the no anomaly group. CONCLUSION" The presence of a fetus with a major anomaly in a twin gestation increases the risk of preterm delivery. The neonatal outcome of the nonanomalous fetus does not appear to be affected by the anomalous fetus. (Am J Obstet Gynecol 1997;177:849-52.)
Key words: Anomaly, twins, p r e t e r m delivery
Twin gestations are at increased risk for congenital anomalies. 1-3 S o m e anomalies are u n i q u e to multiple gestations; these involve vascular anastomoses and affect the n e o n a t a l o u t c o m e of b o t h fetuses. However, most anomalies are structural malformations, and the majority affect only one twin. It is unclear how the presence of a congenital anomaly in o n e twin affects its n o r m a l sibling. In singleton gestations the presence of a structural anomaly increases the risk for p r e t e r m delivery. T h e clinician is faced with difficult m a n a g e m e n t decisions r e g a r d i n g the r o u t e of delivery, w h e n to deliver, how aggressively to attempt to p r e v e n t delivery (i.e., tocolysis), and in some cases w h e t h e r to provide antenatal therapy. Twin gestations are already at increased risk for p r e g n a n c y complications, including p r e t e r m delivery, gestational diabetes, and pregnancy-induced hypertension, and the presence of an a n o m a l o u s fetus further complicates the decision-making process. 4-6 If the p r e s e n c e of a structural anomaly in o n e fetus affects the o u t c o m e of the p r e g n a n c y or has an adverse effect on the unaffected twin, antenatal m a n a g e m e n t From the Department of Obstetrics and Gynecology, University of Texas Southwestern Medical Center. Presented at the Seventeenth Annual Meeting of the Society of Perinatal Obstetricians, Anaheim, California, January 20-25, 1997. Reprint requests: James M. Alexander, MD, Department of Obstetrics and Gynecology, Universi~ of Texas Southwestern Medical Center, 5323 Harry Hines Blvd., Dallas, TX 75235-9032. Copyright © 1997 by Mosby-Year Book, Inc. 0002-9378/97 $5.00 + 0 6/6/84152
options such as increased frequency of prenatal visits, fetal surgery, or selective termination may be warranted. Fetal surgery is largely e x p e r i m e n t a l and poses significant risk to the pregnancy. 7 Likewise, selective feticide of the anomalous fetus may pose risk to the unaffected fetus and can result in miscarriage, a T h e objective of our study was to d e t e r m i n e w h e t h e r the presence of one anomalous fetus in a twin gestation affects pregnancy o u t c o m e w h e n c o m p a r e d with twin pregnancies without fetal anomalies. Material and m e t h o d s
All twin pregnancies delivered at Parkland Memorial Hospital between 1988 and 1994 were studied. Those pregnancies with insufficient data, previable gestation, anomalies of twinning (such as twin-twin transfusion syndrome) in which both fetuses were affected, or cases in which both twins were anomalous were excluded. The remaining pregnancies were first divided into two g r o u p s - those in which one twin was anomalous and those in which there were no anomalies. An anomaly was considered major if it had a significant impact on neonatal morbidity and mortality or required significant operative intervention to correct (e.g., gastroschisis). T h e anomalous group was further divided into major and minor anomalies. The 18 twins with major anomalies and the 38 twins with minor anomalies are listed in Table I. Antenatal ultrasonography is n o t routinely p e r f o r m e d at our institution. T h e most f r e q u e n t indications for 849
850
Alexander et al.
October 1997 AmJ Obstet Gynecol
Table I. Categorization of anomalies by system Abnomaly
Major (n = 18) Central nervous system Cardiac Pulmonary Gastrointestinal Musculoskeletal Genitourinary Multiple organ systems Minor (n = 38) Central nervous system Cardiac Genitourinary Musculoskeletal Dermatologic
No.
2 4 2 5 2 2 1 1 4 7 18 8
ultrasonography in our patient population are an uncertain date of last menstrual period and size-date discrepancies. Most of the twins in our study were identified because of a fundal height measurement that was larger than expected by dates; this usually occurred in the late second trimester or early third trimester. Once identified, twin pregnancies are monitored in a high-risk clinic starting at 26 weeks' gestation with weekly digital cervical examinations and ultrasonography every 4 weeks. Women in whom preterm labor developed or who had preterm cervical dilatation were hospitalized until the gestational age reached 34 weeks or the estimated fetal weight reached 2000 gm. Tocolysis other than hydration and sedation was not used, and none of the pregnant women in this study received medication to stop labor. Tile use of antenatal corticosteroids began in May 1994, and women with twin pregnancies with preterm labor or premature rupture of membranes between 24 and 34 weeks' gestation without diabetes or hypertension received treatment. If a structural anomaly was identified antenatally, the case was reviewed and presented in a fetal therapy conference. A plan of management was developed, and support services were arranged with the assistance of neonatal case managers. Neonatal outcomes assessed included birth weight, gestational age, umbilical arterial blood gas values, sepsis, grades III and 1V intraventricular hemorrhage, necrotizing enterocolitis, seizures, death, and total days in the nursery. Statistical analysis was performed with the Student t test, X2 analysis, and Fisher's exact test with KwikStat (TexaSoft, Cedar Hill, Tex.) statistical software. A statistically significant difference was defined as a p value < 0.05. A power analysis was performed with SAS (SAS Institute, Cary, N.C.) statistical software.
Results During the study period there were 116,635 deliveries at Parkland Hospital, 1119 of which were twins. Com-
Table II. Maternal demographics of twin pregnancies with and without an anomalous fetus Major (n = 18)
Maternal age (yr) 26 -+ 4.9 Race Black 4 (22%) White 5 (28%) Hispanic 9 (50%) Other 0 Parity Nulliparous 0* Multiparous 18 (100%)* Antenatal complications Preeclampsia 1 (6%)j" Diabetes 0 Placenta previa 1 (6%)
Minor (n = 38)
None (n = 914)
25 -+ 6.1
24.7 -+ 5.5
14 (37%) 345 (37%) 10 (26%) 172 (19%) 14 (37%) 373 (41%) 0 24 (3%) 12 (32%) 254 (28%) 26 (68%) 660 (72%) 9 (24%) 218 (24%) 0 19 (2%) 0 2 (0.2%)
*p < 0.05, Fisher's exact test. ]-p = 0.06, Fisher's exact test.
plete records were available for 982 of these twin pregnancies. Sixty-eight (6.9%) of the pregnancies were identified with an anomaly. Nine of these were excluded because both fetuses had an anomaly, and three were eliminated because of abnormalities that resulted from vascular anastomoses; therefore 56 pregnancies with one anomalous fetus remained. In 18 of the pregnancies one fetus had a major anomaly and 38 had a minor anomaly (Table I). Fourteen of the 18 twins with a major anomaly underwent antenatal ultrasonography. Seven of the 14 twins had the major anomaly diagnosed by antenatal ultrasonography. The major anomalies not diagnosed with ultrasonography included ventriculoseptal defect (n = 4), tracheoesophageal fistula (n = 2), and Pierre Robin syndrome (n = 1). None of the pregnancies underwent selective feticide or were intentionally delivered prematurely because of the presence of an anomaly. Table II reflects maternal demographics of the three groups. No significant differences were seen in maternal age, race, parity, or antenatal complications (p > 0.05). Although not statistically significant, preeclampsia was seen less in t h e major anomaly group (6%) than in the other two groups (24%) (p = 0.06). There were no nulliparous women in the major anomaly group, and this was different from the other two groups (p < 0.05). Table III presents labor characteristics of tile two groups. The frequency of premature rupture of the membranes, spontaneous labor, and chorioamnionids was similar among the three groups. Indicated deliveries were usually performed because of preeclampsia or premature rupture of membranes, and this was similar among the groups. More than half of the twins were delivered by cesarean section, mainly because of arrested labor or nonvertex position of the presenting twin. There was no significant difference in cesarean section rates among the groups. There was an increased frequency of preterm delivery in
Volume 177, Number 4 Am J Obstet Oynecol
Alexancter et al.
Table III. Labor characteristics of twin pregnancies with and without an a n o m a l o u s fetus
Premature rupture of membranes Spontaneous labor Indicated delivery Chorioanmionitis Route of delivery Spontaneous vaginal Cesarean section Delivery <37 wk <34 wk Estimated gestational age at delivery (wk, mean -+ SD)
Major (n = 18) (%)
Minor ( n = 3S)
None (n= 914)
1 (5%)
4 (10%)
42 (5%)
13 (72%) 5 (28%) 0
33 (87%) 636 (70%) 5 (13%) 278 (30%) 4 (10.5%) 27 (3%)
7 (39%) 11 (61%)
10 (26%) 28 (74%)
366 (40%) 548 (60%)
13 (72%)* 19 (50%) 377 (41%) 9 (50%)* 10 (26%) 142 (16%) 32.9 +- 4.6* 34.5 -+ 3.8 35.6 -+ 3
*p < 0.05, X2 analysis, Student t test. the major anomaly group, with a mean gestational age at delivery of 32.9 -+ 4.6 weeks. This was significantly different from the m i n o r anomaly and no anomaly groups, which were delivered at a m e a n of 34.5 -+ 3.8 and 35.6 _+ 3.0 weeks, respectively (p < 0.05). N e o n a t a l o u t c o m e is described in Table IV. T h e m e a n birth weight was also significantly different in the major anomaly g r o u p (1759 + 657 gm) c o m p a r e d with the m i n o r anomaly (2130 _+ 690 gin) and no anomaly groups (2291 + 577 gm). T h e m e a n birth weight of the anomalous twin was 1686 _+ 626 gin, and that o f the n o n a n o m a lous twin was 1907 + 688 gin. T h e r e were no differences in the three groups in the frequency of cord gas values <7.2 or 7.0. T h e r e was a significant increase in neonatal mortality in the twins with major anomalies (28%) comp a r e d with the unaffected sibling (6%), the m i n o r anomaly g r o u p (1%), and the no anomaly g r o u p (1%). T h e r e was no difference in the incidence of sepsis, necrotizing enterocolitis, seizure activity, or intraventricular hemorrhage (grade 3 or 4) in the three groups. T h e g r o u p in which one twin had a major anomaly had an increased n e e d for m e c h a n i c a l ventilation (44%) w h e n c o m p a r e d with the m i n o r anomaly and no anomaly groups (16% and 11%, respectively; p < 0.05). W h e n c o m p a r e d with the n o anomaly group, significantly m o r e of the n o r m a l twins in the major anomaly g r o u p had a n e e d for m e c h a n i c a l ventilation (39%, p < 0.05), but the differe n c e between those twins and the m i n o r anomaly group was not statistically significant (p = 0.06). T h e m e a n n u m b e r of hospital days was greatest in the twins with a major anomaly (42 days) as c o m p a r e d with the unaffected twins (27 days), and this was l o n g e r than the stay in the m i n o r anomaly g r o u p (17 days) and the n o anomaly g r o u p (13 days). T h e difference in length of stay between the m i n o r and no anomaly groups was n o t statistically significant.
851
A larger p e r c e n t a g e of twins in the major anomaly g r o u p (both affected and unaffected) stayed in the nursery >1 m o n t h (38% and 33%, respectively) comp a r e d with the m i n o r and no anomaly groups (15% and 18%, p < 0.05). A power analysis was p e r f o r m e d to d e t e r m i n e the smallest detectable difference between the groups. With e~ = 0.05 and [3 =0.8, this study had the power to detect a difference of 16% between the major and no anomaly groups.
Comment Twin pregnancies have an increased risk of preterm labor, pregnancy-induced hypertension, gestational diabetes, and congenital anomalies, which can lead to increased fetal and neonatal morbidity and mormlityJ -6 Vascular anomalies that result in the twin-twin transfusion syndrome occur in twins along with the structural a n d developmental anomalies also seen in singleton pregnancies. ~-9 Kohl and Casey I identified malformations in 6% of twins (2% major) versus 3.5% of singletons (1% major). Similar results have been found by other authors, and the incidence of anomalies in our study population (7%) agreesY3 When both twins have anomalies or vascular anastomoses are present, the neonatal outcomes of both twins are clearly affectedY 3, 9 It is unclear how the neonatal outcome of a normal twin is affected by an anomalous sibling, and this p r o m p t e d our investigation. In a recent report by Malone et al., 1° 24 twin pregnancies with at least one anomalous fetus were studied. Five were excluded from analysis because both twins had an anomaly, and five w o m e n elected to have selective feticide. The remaining 14 twins were compared with a similar cohort of unaffected twin pregnancies. The adverse effects found on pregnancy outcome included an increased incidence of preterm labor and cesarean delivery. Neonatal outcome was worse in the anomalous twins, with a perinatal mortality of 333 in 1000, but the normal twins had an outcome similar to that of the control group. A potential bias in their study is that all of the twins and their anomalies were diagnosed before 20 weeks; after counseling, selective feticide was performed in 5 (26%) of the pregnancies. It is unknown how their results may have been affected if these pregnancies had continued with both twins. The effect of selective feticide on the outcome of the pregnancy can be adverse,and some authors have reported a 10% to 12% loss rate after the procedure. 8 M t h o u g h there are some differences in study design, our findings are very similar to those of Malone et alJ ° T h e major difference in our population is that the majority of our twins were identified in the late second or third trimester, and n o n e of the pregnancies u n d e r w e n t selective feticide. T h u s the natural course of all of the affected pregnancies can be reported. O u r population has very few m a t e r n a l transports. This limits the selection bias associated with m a t e r n a l transport (i.e., only twins
852
Alexander et al.
October 1997 Am J Obstet Oynecol
Table IV. Neonatal o u t c o m e s o f twin pregnancies with and without an anomalous fetus
Birth weight (gm, mean -+ SD) Cord, bloo d gas value <7.20 <7.00 Sepsis Culture-proved Necrotizing enterocolitis Intraventricular hemorrhage (grade III or IV) Death Seizure Mechanical ventilation Total time in hospital (days) Nursery stay >1 mo
Major anomaly (n = 18)
Normal twin (n = 18)
Minor anomaly (n = 76)
No anomaly (n = 1828)
1686 -+ 626*
1907 + 6881-
2130 -+ 690
2291 -+ 577
1 (6%) 0
9 (12%) 1 (1%)
154 (8%) 9 (0.4%)
0 0 1 (6%)
0 0 3 (1%)
9 (0.5%) 18 (1%) 19 (1%)
3 (1%) 0 12 (16%) 17 -+ 20 12 (15%)
19 (1%) 6 (0.3%) 206 (11%) 13 -+ 18 322 (18%)
2 (11%) 1 (6%) 0 0 0 5 (28%)* 0 8 (44%)* 41 -+ 54* 8 (38%)*
1 (6%) 1 (5%) 7 (39%)t" 27 -+ 34* 6 (33%)*
*p < 0.05, compared with minor and no anomaly groups. j-Compared with no anomaly group.
with a complication are transported). Finally, our control g r o u p consisted of all unaffected twin pregnancies delivered during the study period. T h e same group of practitioners using the same m a n a g e m e n t schemes took care of all of these pregnancies. T h e most significant finding f r o m o u r study is that the presence of an anomalous fetus in a twin p r e g n a n c y increases the risk o f p r e t e r m delivery. O u r data suggest that a twin pregnancy complicated by a fetus with a major anomaly will be delivered on average 2 weeks earlier than o t h e r twin pregnancies. This does not appear to be intentional delivery because of the anomaly, because rates of spontaneous and indicated delivery were similar between the pregnancies complicated by a major anomaly and those that were not. T h e increase in p r e t e r m delivery results in a lower birth weight, an increased n e e d for mechanical ventilation, and a p r o l o n g e d hospitalization time for the unaffected twin. It is reassuring that o t h e r neonatal outcomes, including mortality of the unaffected twin, were similar to those of the control group. As m i g h t be expected, fetuses with a major anomaly had a m u c h h i g h e r perinatal mortality rate (278/1000) than the o t h e r groups (10/1000 to 5 6 / 1000). O u r data show that the presence Of a m i n o r anomaly does n o t adversely affect maternal or short-term neonatal outcome. Management of a twin pregnancy with an anomalous fetus is complex, and a m a n a g e m e n t scheme that includes surveillance for preterm labor (such as weekly cervical examinations) may be even more important for these pregnancies than the typical twin pregnancy. W h e n a major anomaly is identified in the second trimester, some authors have suggested selective feticide as a m a n a g e m e n t option. This procedure is not without risks and has been associated with miscarriage of the normal twin in 10% to 12% of the
cases. This risk must be weighed against the generally good prognosis we report with expectant management. In summary, this study found that the presence of a major anomaly in one fetus in a twin gestation significantly increased the risk of preterm delivery. Perinatal mortality and neonatal outcomes of the normal twin are similar to those seen in other twin pregnancies. This report provides useful information for Counseling women about their options when faced with this dilemma.
REFERENCES
I. Kohl SG, Casey G. Twin gestation. Mt SinaiJ Med 1975;42: 523-39. 2. Cameron AH, EdwardsJH, Derom R, Theip/M, Boelaert R. The value of twin surveys in the study of malformation. Eur J Obstet Gynecol Reprod Biol 1983;14:347-56. 3. Baldwin VJ. Pathology of multiple pregnancy. In: Wigglesworth JS, Singer J, editors. Textbook of fetal and perinatal pathology. Boston: Blackwell Scientific; 1991. p. 238. 4. Skupski D, Nelson S, Kowalik A, Polaneczky M, Hutson J, Smith-Leviton M, et al. Multiple gestations from in vitro fertilization: incidence of preeclampsia in reduced versus non-reduced triplets. Am J Obstet Gynecol 1996;174:455. 5. Coonrad CV, Hickok DE, Zhu K, Easterling TR, Daling JR. Risk factors for preeclampsia in twin pregnancies: a population-based cohort study. Obstet Gynecol 1995;85:645-50. 6. Santema JG, Bourdrez P, Wallenberg HC. Maternal and perinatal complications in triplets compared with twin pregnancy. EurJ Obstet Gynecol Reprod Biol 1995;60:143-7. 7. Harrison MR, Golbus MS, Filly RA, Nakayama DK, deLorimier AA. Fetal surgical treatment. Pediatr Ann 1982;11: 896-9. 8. Evans MI, GoldbergJD, Dommergues M, Wapner RJ, Lynch L, Dock BS, et aL Efficacy of second-trimester selective termination for fetal abnormalities: international collaborative experience among the world's largest centers. Am J Obstet Gynecol 1994; 17l:90-4. 9. Robertson EG, Neer KJ. Placental injection studies in twin gestation. Am J Obstet Gynecol 1983;147:170-4. 10. Malone FD, Craigo SD, Chelmow D, D'Alton ME. Outcome of twin gestations complicated by a single anomalous fetus. Obstet Gynecol 1996;88:1-5.
Key words: Anomaly, twins, p r e t e r m delivery
Twin gestations are at increased risk for congenital anomalies. 1-3 S o m e anomalies are u n i q u e to multiple gestations; these involve vascular anastomoses and affect the n e o n a t a l o u t c o m e of b o t h fetuses. However, most anomalies are structural malformations, and the majority affect only one twin. It is unclear how the presence of a congenital anomaly in o n e twin affects its n o r m a l sibling. In singleton gestations the presence of a structural anomaly increases the risk for p r e t e r m delivery. T h e clinician is faced with difficult m a n a g e m e n t decisions r e g a r d i n g the r o u t e of delivery, w h e n to deliver, how aggressively to attempt to p r e v e n t delivery (i.e., tocolysis), and in some cases w h e t h e r to provide antenatal therapy. Twin gestations are already at increased risk for p r e g n a n c y complications, including p r e t e r m delivery, gestational diabetes, and pregnancy-induced hypertension, and the presence of an a n o m a l o u s fetus further complicates the decision-making process. 4-6 If the p r e s e n c e of a structural anomaly in o n e fetus affects the o u t c o m e of the p r e g n a n c y or has an adverse effect on the unaffected twin, antenatal m a n a g e m e n t From the Department of Obstetrics and Gynecology, University of Texas Southwestern Medical Center. Presented at the Seventeenth Annual Meeting of the Society of Perinatal Obstetricians, Anaheim, California, January 20-25, 1997. Reprint requests: James M. Alexander, MD, Department of Obstetrics and Gynecology, Universi~ of Texas Southwestern Medical Center, 5323 Harry Hines Blvd., Dallas, TX 75235-9032. Copyright © 1997 by Mosby-Year Book, Inc. 0002-9378/97 $5.00 + 0 6/6/84152
options such as increased frequency of prenatal visits, fetal surgery, or selective termination may be warranted. Fetal surgery is largely e x p e r i m e n t a l and poses significant risk to the pregnancy. 7 Likewise, selective feticide of the anomalous fetus may pose risk to the unaffected fetus and can result in miscarriage, a T h e objective of our study was to d e t e r m i n e w h e t h e r the presence of one anomalous fetus in a twin gestation affects pregnancy o u t c o m e w h e n c o m p a r e d with twin pregnancies without fetal anomalies. Material and m e t h o d s
All twin pregnancies delivered at Parkland Memorial Hospital between 1988 and 1994 were studied. Those pregnancies with insufficient data, previable gestation, anomalies of twinning (such as twin-twin transfusion syndrome) in which both fetuses were affected, or cases in which both twins were anomalous were excluded. The remaining pregnancies were first divided into two g r o u p s - those in which one twin was anomalous and those in which there were no anomalies. An anomaly was considered major if it had a significant impact on neonatal morbidity and mortality or required significant operative intervention to correct (e.g., gastroschisis). T h e anomalous group was further divided into major and minor anomalies. The 18 twins with major anomalies and the 38 twins with minor anomalies are listed in Table I. Antenatal ultrasonography is n o t routinely p e r f o r m e d at our institution. T h e most f r e q u e n t indications for 849
850
Alexander et al.
October 1997 AmJ Obstet Gynecol
Table I. Categorization of anomalies by system Abnomaly
Major (n = 18) Central nervous system Cardiac Pulmonary Gastrointestinal Musculoskeletal Genitourinary Multiple organ systems Minor (n = 38) Central nervous system Cardiac Genitourinary Musculoskeletal Dermatologic
No.
2 4 2 5 2 2 1 1 4 7 18 8
ultrasonography in our patient population are an uncertain date of last menstrual period and size-date discrepancies. Most of the twins in our study were identified because of a fundal height measurement that was larger than expected by dates; this usually occurred in the late second trimester or early third trimester. Once identified, twin pregnancies are monitored in a high-risk clinic starting at 26 weeks' gestation with weekly digital cervical examinations and ultrasonography every 4 weeks. Women in whom preterm labor developed or who had preterm cervical dilatation were hospitalized until the gestational age reached 34 weeks or the estimated fetal weight reached 2000 gm. Tocolysis other than hydration and sedation was not used, and none of the pregnant women in this study received medication to stop labor. Tile use of antenatal corticosteroids began in May 1994, and women with twin pregnancies with preterm labor or premature rupture of membranes between 24 and 34 weeks' gestation without diabetes or hypertension received treatment. If a structural anomaly was identified antenatally, the case was reviewed and presented in a fetal therapy conference. A plan of management was developed, and support services were arranged with the assistance of neonatal case managers. Neonatal outcomes assessed included birth weight, gestational age, umbilical arterial blood gas values, sepsis, grades III and 1V intraventricular hemorrhage, necrotizing enterocolitis, seizures, death, and total days in the nursery. Statistical analysis was performed with the Student t test, X2 analysis, and Fisher's exact test with KwikStat (TexaSoft, Cedar Hill, Tex.) statistical software. A statistically significant difference was defined as a p value < 0.05. A power analysis was performed with SAS (SAS Institute, Cary, N.C.) statistical software.
Results During the study period there were 116,635 deliveries at Parkland Hospital, 1119 of which were twins. Com-
Table II. Maternal demographics of twin pregnancies with and without an anomalous fetus Major (n = 18)
Maternal age (yr) 26 -+ 4.9 Race Black 4 (22%) White 5 (28%) Hispanic 9 (50%) Other 0 Parity Nulliparous 0* Multiparous 18 (100%)* Antenatal complications Preeclampsia 1 (6%)j" Diabetes 0 Placenta previa 1 (6%)
Minor (n = 38)
None (n = 914)
25 -+ 6.1
24.7 -+ 5.5
14 (37%) 345 (37%) 10 (26%) 172 (19%) 14 (37%) 373 (41%) 0 24 (3%) 12 (32%) 254 (28%) 26 (68%) 660 (72%) 9 (24%) 218 (24%) 0 19 (2%) 0 2 (0.2%)
*p < 0.05, Fisher's exact test. ]-p = 0.06, Fisher's exact test.
plete records were available for 982 of these twin pregnancies. Sixty-eight (6.9%) of the pregnancies were identified with an anomaly. Nine of these were excluded because both fetuses had an anomaly, and three were eliminated because of abnormalities that resulted from vascular anastomoses; therefore 56 pregnancies with one anomalous fetus remained. In 18 of the pregnancies one fetus had a major anomaly and 38 had a minor anomaly (Table I). Fourteen of the 18 twins with a major anomaly underwent antenatal ultrasonography. Seven of the 14 twins had the major anomaly diagnosed by antenatal ultrasonography. The major anomalies not diagnosed with ultrasonography included ventriculoseptal defect (n = 4), tracheoesophageal fistula (n = 2), and Pierre Robin syndrome (n = 1). None of the pregnancies underwent selective feticide or were intentionally delivered prematurely because of the presence of an anomaly. Table II reflects maternal demographics of the three groups. No significant differences were seen in maternal age, race, parity, or antenatal complications (p > 0.05). Although not statistically significant, preeclampsia was seen less in t h e major anomaly group (6%) than in the other two groups (24%) (p = 0.06). There were no nulliparous women in the major anomaly group, and this was different from the other two groups (p < 0.05). Table III presents labor characteristics of tile two groups. The frequency of premature rupture of the membranes, spontaneous labor, and chorioamnionids was similar among the three groups. Indicated deliveries were usually performed because of preeclampsia or premature rupture of membranes, and this was similar among the groups. More than half of the twins were delivered by cesarean section, mainly because of arrested labor or nonvertex position of the presenting twin. There was no significant difference in cesarean section rates among the groups. There was an increased frequency of preterm delivery in
Volume 177, Number 4 Am J Obstet Oynecol
Alexancter et al.
Table III. Labor characteristics of twin pregnancies with and without an a n o m a l o u s fetus
Premature rupture of membranes Spontaneous labor Indicated delivery Chorioanmionitis Route of delivery Spontaneous vaginal Cesarean section Delivery <37 wk <34 wk Estimated gestational age at delivery (wk, mean -+ SD)
Major (n = 18) (%)
Minor ( n = 3S)
None (n= 914)
1 (5%)
4 (10%)
42 (5%)
13 (72%) 5 (28%) 0
33 (87%) 636 (70%) 5 (13%) 278 (30%) 4 (10.5%) 27 (3%)
7 (39%) 11 (61%)
10 (26%) 28 (74%)
366 (40%) 548 (60%)
13 (72%)* 19 (50%) 377 (41%) 9 (50%)* 10 (26%) 142 (16%) 32.9 +- 4.6* 34.5 -+ 3.8 35.6 -+ 3
*p < 0.05, X2 analysis, Student t test. the major anomaly group, with a mean gestational age at delivery of 32.9 -+ 4.6 weeks. This was significantly different from the m i n o r anomaly and no anomaly groups, which were delivered at a m e a n of 34.5 -+ 3.8 and 35.6 _+ 3.0 weeks, respectively (p < 0.05). N e o n a t a l o u t c o m e is described in Table IV. T h e m e a n birth weight was also significantly different in the major anomaly g r o u p (1759 + 657 gm) c o m p a r e d with the m i n o r anomaly (2130 _+ 690 gin) and no anomaly groups (2291 + 577 gm). T h e m e a n birth weight of the anomalous twin was 1686 _+ 626 gin, and that o f the n o n a n o m a lous twin was 1907 + 688 gin. T h e r e were no differences in the three groups in the frequency of cord gas values <7.2 or 7.0. T h e r e was a significant increase in neonatal mortality in the twins with major anomalies (28%) comp a r e d with the unaffected sibling (6%), the m i n o r anomaly g r o u p (1%), and the no anomaly g r o u p (1%). T h e r e was no difference in the incidence of sepsis, necrotizing enterocolitis, seizure activity, or intraventricular hemorrhage (grade 3 or 4) in the three groups. T h e g r o u p in which one twin had a major anomaly had an increased n e e d for m e c h a n i c a l ventilation (44%) w h e n c o m p a r e d with the m i n o r anomaly and no anomaly groups (16% and 11%, respectively; p < 0.05). W h e n c o m p a r e d with the n o anomaly group, significantly m o r e of the n o r m a l twins in the major anomaly g r o u p had a n e e d for m e c h a n i c a l ventilation (39%, p < 0.05), but the differe n c e between those twins and the m i n o r anomaly group was not statistically significant (p = 0.06). T h e m e a n n u m b e r of hospital days was greatest in the twins with a major anomaly (42 days) as c o m p a r e d with the unaffected twins (27 days), and this was l o n g e r than the stay in the m i n o r anomaly g r o u p (17 days) and the n o anomaly g r o u p (13 days). T h e difference in length of stay between the m i n o r and no anomaly groups was n o t statistically significant.
851
A larger p e r c e n t a g e of twins in the major anomaly g r o u p (both affected and unaffected) stayed in the nursery >1 m o n t h (38% and 33%, respectively) comp a r e d with the m i n o r and no anomaly groups (15% and 18%, p < 0.05). A power analysis was p e r f o r m e d to d e t e r m i n e the smallest detectable difference between the groups. With e~ = 0.05 and [3 =0.8, this study had the power to detect a difference of 16% between the major and no anomaly groups.
Comment Twin pregnancies have an increased risk of preterm labor, pregnancy-induced hypertension, gestational diabetes, and congenital anomalies, which can lead to increased fetal and neonatal morbidity and mormlityJ -6 Vascular anomalies that result in the twin-twin transfusion syndrome occur in twins along with the structural a n d developmental anomalies also seen in singleton pregnancies. ~-9 Kohl and Casey I identified malformations in 6% of twins (2% major) versus 3.5% of singletons (1% major). Similar results have been found by other authors, and the incidence of anomalies in our study population (7%) agreesY3 When both twins have anomalies or vascular anastomoses are present, the neonatal outcomes of both twins are clearly affectedY 3, 9 It is unclear how the neonatal outcome of a normal twin is affected by an anomalous sibling, and this p r o m p t e d our investigation. In a recent report by Malone et al., 1° 24 twin pregnancies with at least one anomalous fetus were studied. Five were excluded from analysis because both twins had an anomaly, and five w o m e n elected to have selective feticide. The remaining 14 twins were compared with a similar cohort of unaffected twin pregnancies. The adverse effects found on pregnancy outcome included an increased incidence of preterm labor and cesarean delivery. Neonatal outcome was worse in the anomalous twins, with a perinatal mortality of 333 in 1000, but the normal twins had an outcome similar to that of the control group. A potential bias in their study is that all of the twins and their anomalies were diagnosed before 20 weeks; after counseling, selective feticide was performed in 5 (26%) of the pregnancies. It is unknown how their results may have been affected if these pregnancies had continued with both twins. The effect of selective feticide on the outcome of the pregnancy can be adverse,and some authors have reported a 10% to 12% loss rate after the procedure. 8 M t h o u g h there are some differences in study design, our findings are very similar to those of Malone et alJ ° T h e major difference in our population is that the majority of our twins were identified in the late second or third trimester, and n o n e of the pregnancies u n d e r w e n t selective feticide. T h u s the natural course of all of the affected pregnancies can be reported. O u r population has very few m a t e r n a l transports. This limits the selection bias associated with m a t e r n a l transport (i.e., only twins
852
Alexander et al.
October 1997 Am J Obstet Oynecol
Table IV. Neonatal o u t c o m e s o f twin pregnancies with and without an anomalous fetus
Birth weight (gm, mean -+ SD) Cord, bloo d gas value <7.20 <7.00 Sepsis Culture-proved Necrotizing enterocolitis Intraventricular hemorrhage (grade III or IV) Death Seizure Mechanical ventilation Total time in hospital (days) Nursery stay >1 mo
Major anomaly (n = 18)
Normal twin (n = 18)
Minor anomaly (n = 76)
No anomaly (n = 1828)
1686 -+ 626*
1907 + 6881-
2130 -+ 690
2291 -+ 577
1 (6%) 0
9 (12%) 1 (1%)
154 (8%) 9 (0.4%)
0 0 1 (6%)
0 0 3 (1%)
9 (0.5%) 18 (1%) 19 (1%)
3 (1%) 0 12 (16%) 17 -+ 20 12 (15%)
19 (1%) 6 (0.3%) 206 (11%) 13 -+ 18 322 (18%)
2 (11%) 1 (6%) 0 0 0 5 (28%)* 0 8 (44%)* 41 -+ 54* 8 (38%)*
1 (6%) 1 (5%) 7 (39%)t" 27 -+ 34* 6 (33%)*
*p < 0.05, compared with minor and no anomaly groups. j-Compared with no anomaly group.
with a complication are transported). Finally, our control g r o u p consisted of all unaffected twin pregnancies delivered during the study period. T h e same group of practitioners using the same m a n a g e m e n t schemes took care of all of these pregnancies. T h e most significant finding f r o m o u r study is that the presence of an anomalous fetus in a twin p r e g n a n c y increases the risk o f p r e t e r m delivery. O u r data suggest that a twin pregnancy complicated by a fetus with a major anomaly will be delivered on average 2 weeks earlier than o t h e r twin pregnancies. This does not appear to be intentional delivery because of the anomaly, because rates of spontaneous and indicated delivery were similar between the pregnancies complicated by a major anomaly and those that were not. T h e increase in p r e t e r m delivery results in a lower birth weight, an increased n e e d for mechanical ventilation, and a p r o l o n g e d hospitalization time for the unaffected twin. It is reassuring that o t h e r neonatal outcomes, including mortality of the unaffected twin, were similar to those of the control group. As m i g h t be expected, fetuses with a major anomaly had a m u c h h i g h e r perinatal mortality rate (278/1000) than the o t h e r groups (10/1000 to 5 6 / 1000). O u r data show that the presence Of a m i n o r anomaly does n o t adversely affect maternal or short-term neonatal outcome. Management of a twin pregnancy with an anomalous fetus is complex, and a m a n a g e m e n t scheme that includes surveillance for preterm labor (such as weekly cervical examinations) may be even more important for these pregnancies than the typical twin pregnancy. W h e n a major anomaly is identified in the second trimester, some authors have suggested selective feticide as a m a n a g e m e n t option. This procedure is not without risks and has been associated with miscarriage of the normal twin in 10% to 12% of the
cases. This risk must be weighed against the generally good prognosis we report with expectant management. In summary, this study found that the presence of a major anomaly in one fetus in a twin gestation significantly increased the risk of preterm delivery. Perinatal mortality and neonatal outcomes of the normal twin are similar to those seen in other twin pregnancies. This report provides useful information for Counseling women about their options when faced with this dilemma.
REFERENCES
I. Kohl SG, Casey G. Twin gestation. Mt SinaiJ Med 1975;42: 523-39. 2. Cameron AH, EdwardsJH, Derom R, Theip/M, Boelaert R. The value of twin surveys in the study of malformation. Eur J Obstet Gynecol Reprod Biol 1983;14:347-56. 3. Baldwin VJ. Pathology of multiple pregnancy. In: Wigglesworth JS, Singer J, editors. Textbook of fetal and perinatal pathology. Boston: Blackwell Scientific; 1991. p. 238. 4. Skupski D, Nelson S, Kowalik A, Polaneczky M, Hutson J, Smith-Leviton M, et al. Multiple gestations from in vitro fertilization: incidence of preeclampsia in reduced versus non-reduced triplets. Am J Obstet Gynecol 1996;174:455. 5. Coonrad CV, Hickok DE, Zhu K, Easterling TR, Daling JR. Risk factors for preeclampsia in twin pregnancies: a population-based cohort study. Obstet Gynecol 1995;85:645-50. 6. Santema JG, Bourdrez P, Wallenberg HC. Maternal and perinatal complications in triplets compared with twin pregnancy. EurJ Obstet Gynecol Reprod Biol 1995;60:143-7. 7. Harrison MR, Golbus MS, Filly RA, Nakayama DK, deLorimier AA. Fetal surgical treatment. Pediatr Ann 1982;11: 896-9. 8. Evans MI, GoldbergJD, Dommergues M, Wapner RJ, Lynch L, Dock BS, et aL Efficacy of second-trimester selective termination for fetal abnormalities: international collaborative experience among the world's largest centers. Am J Obstet Gynecol 1994; 17l:90-4. 9. Robertson EG, Neer KJ. Placental injection studies in twin gestation. Am J Obstet Gynecol 1983;147:170-4. 10. Malone FD, Craigo SD, Chelmow D, D'Alton ME. Outcome of twin gestations complicated by a single anomalous fetus. Obstet Gynecol 1996;88:1-5.