- Email: [email protected]
Birth weight discordancy and adverse perinatal outcomes among twin gestations in the United States: The effect of placental abruption
Obstetrics
Birth weight discordancy and adverse perinatal outcomes among twin gestations in the United States: The effect of placental abruption Cande V. Ananth, PhD, MPH,a,c Kitaw Demissie, MD, PhD,c and Maryellen L. Hanley, MD, MPHb New Brunswick and Piscataway, NJ OBJECTIVE: We evaluated whether the relationship between birth weight discordancy of twins and stillbirth, neonatal deaths, and preterm births was modified by the presence of abruption. STUDY DESIGN: We used the 1995 to 1997 matched multiple birth file for United States twin births (n = 269,287). Birth weight discordancy was defined as the ratio of the difference in birth weight of the heavier from the lighter twin to that of the heavier twin and was categorized as <5%, 5% to 9%, 10% to 14%, 15% to 19%, 20% to 29%, 30% to 39%, and 40%. We evaluated the risks of stillbirth (20 weeks of gestation), neonatal deaths (within 28 days after birth), and preterm birth (< 32 weeks) in the presence and absence of abruption. Associations between birth weight discordancy and these perinatal outcomes were expressed as adjusted relative risks and were derived from multivariable logistic regression models, based on the method of generalized estimating equations. Risk of these outcomes were derived for each stratum of birth weight discordancy and abruption status, with the <5% birth weight discordancy, nonabruption status labeled as the reference group. All analyses were performed separately for same and different sex twins. RESULTS: A birth weight discordancy of 20% among same sex (adjusted relative risk, 1.2; 95% CI, 1.1, 1.4), and 40% among different sex twins (relative risk, 2.2; 95% CI, 1.7, 2.8) conferred increased risk for abruption. Among nonabruption births, a birth weight discordancy of 15% among same sex and 30% among different sex twins increased the risk of stillbirths, neonatal deaths, and preterm births. Among abruption births, however, the risks were increased even in the lowest birth weight discordancy category (<5%). The relative risks of stillbirths and neonatal deaths among abruption births were significantly higher for each birth weight discordancy group, both for same and different sex twins, compared with the reference group. The association between birth weight discordancy and preterm birth was not modified by either the presence or absence of abruption. CONCLUSION: Birth weight discordancy of 15% for same sex and 30% for different sex confer greatest risk of adverse perinatal outcomes in the absence of abruption. In the presence of placental abruption, these risks are further compounded. The results underscore the need for careful monitoring of twin pregnancies. (Am J Obstet Gynecol 2003;188:954-60.)
Key words: Placental abruption, stillbirth, neonatal death, twins, birth weight discordancy Abnormal discordance in size at birth between twins within a sibship has been defined variously, from 15% to as much as 40%.1-3 Extreme birth weight discordancy in twins has been associated with adverse perinatal outcomes that include a higher risk of fetal and neonatal From the Section of Epidemiology and Biostatisticsa and the Division of Maternal-Fetal Medicine, Department of Obstetrics, Gynecology, and Reproductive Sciences,b University of Medicine and Dentistry of New Jersey, Robert Wood Johnson Medical School/Saint Peter’s University Hospital, and the Division of Epidemiology, School of Public Health, University of Medicine and Dentistry of New Jersey.c Supported in part by National Institutes of Health grant No. IR01HD038902 (C. V. A.). Received for publication July 22, 2002; revised October 18, 2002; accepted December 3, 2002. Reprints not available from the authors. © 2003, Mosby, Inc. All rights reserved. 0002-9378/2003 $30.00 + 0 doi:10.1067/mob.2003.210
954
deaths4 and preterm birth.3-6 The discordancy in size is often the consequence of fetal growth restriction of the smaller twin.7-12 The causes that are responsible for fetal growth restriction are multifactorial and varied. Birth weight discordancy may be a consequence of abnormal placental disease, such as infarctions that often result in growth restricted babies. We recently demonstrated that, among singleton pregnancies, women with placental abruption are at greater likelihood of delivering a growth restricted infant compared with otherwise normal pregnancies.13 These observations led us to speculate that placental abruption was the end result of an underlying chronic disease process, with origins early in pregnancy, and perhaps even extending to the time of conception.13 These chronic processes and placental abruption can also contribute independently to birth weight discordancy and, when present together, can further elevate
Ananth, Demissie, and Hanley 955
Volume 188, Number 4 Am J Obstet Gynecol
Table I. Relationship between birth weight discordancy and placental abruption: United States twin births 1995 through 1997 Birth weight discordancy
Total twin births
Placental abruption N (%)
Relative risk (95% confidence interval)
Same sex (%) <5 5-9 10-14 15-19 20-29 30-39 ≥40 Total
55,969 43,887 31,183 19,528 18,716 6,073 5,665 181,021
640 (11.4) 539 (12.3) 351 (11.4) 248 (12.7) 258 (13.8) 104 (17.1) 140 (24.7) 2,286 (12.6)
1.0 (Reference) 1.1 (0.9-1.1) 1.0 (0.9-1.1) 1.1 (0.9-1.3) 1.2 (1.1-1.4) 1.5 (1.2-1.9) 2.2 (1.8-2.6) —
Different sex (%) <5 5-9 10-14 15-19 20-29 30-39 ≥40 Total
23,798 20,369 16,657 11,331 10,826 3,036 2,249 88,266
362 (15.2) 314 (15.4) 202 (12.1) 173 (15.3) 145 (13.4) 43 (14.2) 73 (32.5) 1,312 (14.9)
1.0 (Reference) 1.0 (0.9-1.2) 0.8 (0.7-0.9) 1.0 (0.8-1.2) 0.9 (0.7-1.1) 0.9 (0.71.3) 2.2 (1.7-2.8) —
Rate of placental abruption per 1,000 twin births. Relative risks were adjusted for maternal age, primigravida, maternal education, prenatal care, chronic hypertension, and infant size (smaller or larger twin).
risks of adverse perinatal outcomes to birth weight discordant twins. We evaluated whether twin birth weight discordancy for the risk of fetal and neonatal deaths and preterm birth is modified by the presence or absence of placental abruption. We also evaluated whether the small for gestational age status of one or both twins contributed to the increased risks of these adverse perinatal outcomes. Material and methods We used the “matched multiple birth file” data for all twin births in the United States (1995-1997).14 These data were assembled by the National Center for Health Statistics of the Centers for Disease Control and Prevention and include data from birth and fetal and infant death certificates in twin and higher order pregnancies. The National Center for Health Statistics used a three-stage matching algorithm (that identified >99% of all twin births within a sibship) to match fetuses within the twin and higher order pregnancies, the details of which have been documented elsewhere.14 We restricted our analysis to twin gestations. We abstracted data for each fetus (ie, fetal and infant death or live birth, birth weight and sex of the fetus for each twin pair, gestational age, and other sociodemographic and medical and obstetric complications of pregnancy). Birth weight discordancy was defined as the ratio of the difference in birth weight of the lighter twin from the heavier twin to that of the heavier twin. The discordancy was then categorized as 0% to 4%, 5% to 9%, 10% to 14%, 15% to 19%, 20% to 29%, 30% to 39% and ≥40%. Gestational age assignment was based on the date of last menstrual period predominantly. In approximately 5% of births,
the day and month of the last menstrual period was missing; therefore, it was imputed.15 Inconsistent gestational age was replaced by a clinical estimate in a small fraction of births (<5%).16 The replacement of clinical estimate and the gestational age imputations were performed by the National Center for Health Statistics.15,16 We evaluated the risks of stillbirth (≥20 weeks), neonatal death (within 28 days after birth), and preterm birth (<32 weeks). Analysis pertaining to preterm births was restricted to twin live births. The risk of stillbirth between abruption and nonabruption births by birth weight discordancy was compared when either twin was small for gestational age (one twin being small for gestational age and the other appropriately grown) and when both twins were grown appropriately. Small for gestational age births were defined as birth weight below the 10th percentile for gestational age. This nomogram was based on all twin live births who were delivered in the United States between 1995 and 1997. Relative risks (with associated 95% CIs) for adverse outcomes were calculated for each birth weight discordancy category and abruption status with a birth weight discordancy of <5%; the nonabruption category served as the reference group. Adjusted odds ratios were derived from fitting multivariable logistic regression models and were transformed to adjusted relative risks.17 Confounders were considered for adjustment in the model if they changed the unadjusted relative risk by at least 10%. Potential confounders that were considered for adjustment included maternal age (<20, 20-24, 25-29, 30-34, and ≥35 years), gravidity (primigravida or gravida ≥2), maternal race/ethnicity (white, black, or other race/ethnicity), maternal education (expressed as completed years of school-
956 Ananth, Demissie, and Hanley
April 2003 Am J Obstet Gynecol
Table II. Relationship among birth weight discordancy, stillbirths, and placental abruption: United States twin births 1995 through 1997 Nonabruption births
Placental abruption births
Relative risk (95% CI)
Population attributable fraction (%)
Same sex (%) <5 605 (10.9) 5-9 377 (8.7) 10-14 347 (11.3) 15-19 310 (16.1) 20-29 491 (26.6) 30-39 362 (60.6) ≥40 973 (176.1) Total 3465 (19.4)
1.0 (Reference) 0.8 (0.7-0.9) 1.1 (0.9-1.2) 1.5 (1.3-1.7) 2.6 (2.3-2.9) 6.1 (5.3-7.0) 19.4 (17.4-21.6) —
Different sex (%) <5 118 (5.0) 5-9 102 (5.1) 10-14 84 (5.1) 15-19 64 (5.7) 20-29 66 (6.2) 30-39 64 (21.4) ≥40 282 (129.6) Total 780 (9.2)
1.0 (Reference) 1.0 (0.8-1.3) 1.0 (0.8-1.4) 1.2 (0.9-1.6) 1.3 (0.9-1.7) 4.5 (3.3-6.1) 29.7 (23.8-37.0) —
Birth weight discordancy
N (%)
N (%)
Relative risk (95% CI)
Population attributable fraction (%)
— — 0.1 0.5 1.6 5.1 16.7 —
24 (37.5) 22 (40.8) 15 (42.0) 13 (52.4) 32 (124.0) 11 (105.8) 36 (257.1) 153 (66.9)
3.5 (2.3-5.3) 3.9 (2.5-6.0) 4.1 (2.4-6.9) 5.0 (2.8-8.8) 13.0 (8.8-19.0) 11.0 (5.8-20.8) 32.8 (22.1-48.5) —
2.7 3.0 3.2 4.2 11.4 9.6 24.9 —
— — — 0.1 0.1 1.7 12.2 —
8 (22.1) 8 (25.5) 13 (64.4) 9 (52.0) 6 (41.4) 4 (93.0) 16 (219.2) 64 (48.8)
4.6 (2.2-9.5) 5.5 (2.7-11.4) 13.0 (7.2-23.5) 11.8 (5.9-23.8) 8.7 (3.7-20.1) 21.8 (7.6-66.5) 56.6 (31.3-102.3) —
1.7 2.1 5.9 4.8 3.7 8.8 21.5 —
Stillbirth rates per 1,000 twin births. Relative risks were adjusted for maternal age, primigravida, maternal education, prenatal care, chronic hypertension, and infant size (larger or smaller twin).
ing, <12 or ≥13 years), cigarette smoking and alcohol use during pregnancy, prenatal care (any vs no care), infant size (smaller or larger twin), and chronic hypertension (diagnosed before pregnancy or before 20 weeks of gestation). Smoking and alcohol use were analyzed as indicator variables (yes/no).16 Adjustment for confounding factors are noted in the results. Because twin births violate the statistical assumption of independence of observations, all analyses were based on the method of generalized estimating equations.18 These data do not contain details on chorionicity or zygosity of the twin fetuses within a sibship. Therefore, we performed all analyses separately for same sex and different sex twins and adjusted for twin size (smaller or larger twin) in all the regression models. Finally, we calculated the population attributable fraction,19 which was defined as the proportion of the adverse outcome that can be attributed to birth weight discordancy in the abruption and nonabruption groups. During the period of 1995 to 1997, there were 308,013 twin fetuses delivered in the United States. We excluded 3,602 twins who were delivered before 20 weeks and 35,124 twins with missing information about placental abruption; 269,287 twin births were available for analyses. The study was approved by the institutional review board of the Robert Wood Johnson Medical School. Results Placental abruption was recorded in 3,598 of 269,287 twin births (13.4 per 1000 births). The prevalence rate
was higher among twins of different sex (14.9 per 1000 births) than among same sex twins (12.6 per 1000 births, Table I). Among same sex twins, placental abruption was associated with ≥20% birth weight discordancy, whereas among different sex twins an increased risk was apparent only among the extremely discordant twins (≥40%). The overall stillbirth rate among nonabruption births was >2-fold greater among same sex (19.4 per 1,000 births) compared with different sex twins (9.2 per 1,000 births, Table II). Among same sex twins, increasing birth weight discordancy of >15% was associated with increasing rates of stillbirth in the nonabruption group. Among same sex twins with placental abruption, the risk of stillbirths was high even among twins with the lowest discordancy. Stillbirths were increased among different sex twins only with a birth weight discordancy of ≥30% in the nonabruption group. However, in the placental abruption group, stillbirth rates were increased even in the smallest birth weight discordancy categories. In the nonabruption group, neonatal mortality rates were 24.9 and 19.0 per 1000 live births among same and different sex twins, respectively (Table III). These rates were approximately 3- to 4-fold higher among those infants born to women with placental abruption. Once again, at a birth weight discordancy of ≥15%, the risk of neonatal deaths was increased for same sex twins of a nonabruption pregnancy. At a birth weight discordancy of ≥30%, the risk of neonatal death was increased for different sex twins of nonabruption pregnancies.
Ananth, Demissie, and Hanley 957
Volume 188, Number 4 Am J Obstet Gynecol
Table III. Relationship among birth weight discordancy, neonatal deaths, and placental abruption: United States twin births 1995 to 1997 Nonabruption births
Placental abruption births
N (%)
Relative risk (95% CI)
Population attributable fraction (%)
883 (16.2) 663 (15.5) 483 (15.9) 310 (16.4) 399 (22.4) 227 (40.9) 1255 (279.8) 4220 (24.9)
1.0 (Reference) 1.0 (0.9-1.1) 1.0 (0.9-1.1) 1.0 (0.9-1.2) 1.5 (1.3-1.7) 2.8 (2.4-3.1) 25.0 (22.7-27.4) —
Different sex (%) <5 320 (13.8) 5-9 264 (13.3) 10-14 218 (13.4) 15-19 160 (14.5) 20-29 145 (13.7) 30-39 53 (18.2) ≥40 453 (242.4) Total 1613 (19.0)
1.0 (Reference) 1.0 (0.8-1.2) 1.0 (0.8-1.2) 1.1 (0.9-1.3) 1.0 (0.9-1.3) 1.4 (1.1-1.9) 24.3 (20.8-28.4) —
Birth weight discordancy Same sex (%) <5 5-9 10-14 15-19 20-29 30-39 ≥40 Total
N (%)
Relative risk (95% CI)
Population attributable fraction (%)
— — — 0.1 0.7 2.6 26.8 —
47 (77.0) 31 (60.7) 29 (85.3) 17 (73.0) 15 (67.9) 10 (108.7) 45 (436.9) 194 (91.1)
5.1 (3.7-6.9) 4.0 (2.8-5.9) 6.0 (4.0-8.8) 4.7 (2.9-7.8) 4.6 (2.7-7.7) 7.9 (4.1-15.4) 53.1 (35.4-79.5) —
6.2 4.6 7.1 5.8 5.3 9.5 42.9 —
— — — 0.1 0.1 0.6 23.2 —
23 (66.3) 19 (62.7) 6 (31.9) 14 (86.6) 13 (94.2) 3 (76.9) 23 (410.7) 101 (81.9)
5.3 (3.4-8.2) 5.2 (3.2-8.3) — 7.5 (4.3-13.2) 8.4 (4.7-15.1) — 54.3 (31.1-94.9) —
5.4 5.1 — 7.5 8.3 — 40.3 —
Neonatal death rates per 1,000 twin live births. Relative risks were adjusted for maternal age, primigravida, maternal education, prenatal care, chronic hypertension, and infant size (larger or smaller twin).
Approximately one fifth of all twin births were delivered preterm (<32 weeks of gestation) among nonabruption births, both for same sex and different sex twins (Table IV). However, more than two thirds of pregnancies were delivered before 32 completed weeks of gestation among abruption births. Preterm birth rates were increased slightly for a birth weight discordancy of ≥20% among same sex twins and ≥30% among different sex twins in the nonabruption group. Among placental abruption births, however, the risk of delivering preterm was apparent even for the lowest discordancy groups. On the basis of the analysis of birth weight discordancy and stillbirths (Table II), we evaluated the risk of stillbirths for a weight discordancy of ≥20%, stratified by both twins being small for gestational age, one twin being small for gestational age, or both twins being grown appropriately (Table V). When both twins were appropriately grown and were of the same sex, the stillbirth rate was 11.7 per 1,000 births for a birth weight discordancy of <20% in the nonabruption group. Comparable data when either twin was small for gestational age was 34.4 per 1,000 births. Similarly, the rates among different sex twins were 6.0 and 19.9 per 1,000 births, respectively. For the same sex twins, either a birth weight discordancy of ≥20% or the presence of abruption conferred relative risks for stillbirths of similar magnitude. However, for different sex twins, abruption conferred greater relative risks for stillbirths. Finally, the presence of birth weight discordancy accompanied by abruption conferred the
greatest risk for stillbirths, both for appropriately grown twins and when either twin was growth restricted. Comment This is perhaps the largest study to date to report the effects of birth weight discordancy in twins in regard to the risks of stillbirth, neonatal deaths, and preterm births among women with and without placental abruption. Overall, a birth weight discordancy of ≥15% among same sex and ≥30% among different sex twins confer the greatest risks of stillbirth and neonatal deaths. In the presence of placental abruption, the risks of stillbirth and neonatal deaths were increased even among the smallest weight discordant twins. When both twins were small for gestational age, the risk of stillbirth was 7-fold greater at a birth weight discordancy of ≥20% in the presence of placental abruption. Birth weight discordancy in twins has different underlying causes that are based on placental chorionicity. In the dichorionic twins, the discordancy is thought to arise because of uterine overcrowding, increased placental mass, placental insufficiency,11 and genetic differences in the growth potential. Among monochorionic twins, the discordancy may be the result of hemodynamic imbalance because of vascular anastomoses,10,11 fetal anemia,20,21 or twin-twin transfusion syndrome.20-22 Furthermore, abnormalities of the placenta may also contribute disproportionately to adverse perinatal outcomes in monochorionic versus dichorionic placentae. Re-
958 Ananth, Demissie, and Hanley
April 2003 Am J Obstet Gynecol
Table IV. Relationship among birth weight discordancy, preterm births (<32 weeks of gestation), and placental abruption: United States twin births 1995 through 1997 Nonabruption births
Placental abruption births
N (%)
Relative risk (95% CI)
Population attributable fraction (%)
Same sex (%) <5 5-9 10-14 15-19 20-29 30-39 ≥40 Total
6,043 (18.6) 4,395 (17.5) 3,114 (17.5) 2,045 (18.5) 2,474 (24.3) 1,305 (42.6) 2,756 (74.1) 22,132 (21.4)
1.0 (Reference) 0.9 (0.9-1.0) 0.9 (0.8-1.0) 1.0 (0.9-1.1) 1.3 (1.3-1.4) 2.3 (2.2-2.4) 4.0 (3.9-4.1) —
Different sex (%) <5 5-9 10-14 15-19 20-29 30-39 ≥40 Total
2,331 (16.9) 1,927 (16.2) 1,557 (16.2) 1,018 (15.7) 953 (15.9) 375 (24.6) 935 (67.1) 9,096 (17.9)
1.0 (Reference) 1.0 (0.9-1.0) 1.0 (0.9-1.0) 0.9 (0.8-1.0) 1.0 (0.9-1.0) 1.5 (1.3-1.6) 4.0 (3.8-4.1) —
Birth weight discordancy
N (%)
Relative risk (95% CI)
Population attributable fraction (%)
— — — — 5.3 19.5 35.8 —
265 (67.3) 195 (64.8) 143 (68.1) 104 (81.9) 121 (72.5) 47 (72.3) 106 (94.6) 981 (71.3)
3.6 (3.4-3.9) 3.5 (3.2-3.8) 3.7 (3.4-4.0) 4.4 (4.0-4.7) 3.9 (3.5-4.2) 3.9 (3.3-4.4) 5.1 (4.8-5.2) —
32.6 31.7 33.4 38.7 35.0 35.0 43.3 —
— — — — — 7.8 33.6 —
123 (59.4) 101 (65.2) 62 (54.4) 53 (54.6) 50 (58.1) 14 (45.2) 46 (95.8) 449 (60.8)
3.6 (3.2-4.0) 4.0 (3.5-4.4) 3.2 (2.6-3.7) 3.4 (2.8-4.0) 3.5 (2.8-4.1) 2.8 (1.8-3.8) 5.7 (5.1-5.9) —
30.5 33.6 27.1 27.2 29.7 23.3 44.3 —
Relative risks were adjusted for maternal age, primigravida, maternal education, prenatal care, chronic hypertension, and infant size (larger or smaller twin).
Table V. Rate and relative risk for stillbirths by birth weight discordancy, placental abruption, and small for gestational age status in twin sibships: United States twin births 1995 through 1997 Both twins normal
Stillbirth
N (%)
Relative risk (95% CI)
One twin small for gestational age
Both twins small for gestational age
N (%)
Relative risk (95% CI)
N (%)
Relative risk (95% CI)
Same sex <20%, Nonabruption <20%, Abruption ≥20%, Nonabruption ≥20%, Abruption
1583 (11.7) 70 (43.3) 633 (44.4) 28 (125.6)
1.0 (Reference) 3.5 (3.1-3.9) 1.3 (1.0-1.6) 4.6 (2.8-7.5)
237 (25.1) 8 (70.2) 583 (47.4) 27 (153.4)
1.0 (Reference) 1.5 (0.7-3.5) 2.1 (1.7-2.5) 5.5 (3.4-8.9)
299 (48.9) 9 (78.9) 136 (113.3) 5 (312.5)
1.0 (Reference) 0.8 (0.3-1.6) 3.1 (2.4-4.0) 2.7 (0.9-8.9)
Different sex <20%, Nonabruption <20%, Abruption ≥20%, Nonabruption ≥20%, Abruption
397 (6.0) 35 (37.1) 102 (11.2) 8 (57.1)
1.0 (Reference) 2.7 (2.1-3.5) 2.1 (1.4-3.2) 2.9 (1.2-7.2)
63 (16.1) 4 (56.3) 155 (27.9) 3 (51.7)
1.0 (Reference) 2.1 (1.5-2.9) 2.8 (0.9-8.9) 3.5 (1.1-12.1)
53 (24.7) 4 (60.6) 17 (50.9) 3 (166.7)
1.0 (Reference) 3.5 (1.9-6.4) 2.5 (0.8-7.7) 7.1 (1.7-31.5)
Stillbirth rates per 1,000 twin births. Relative risks were adjusted for maternal age, primigravida, maternal education, prenatal care, chronic hypertension, and infant size (larger or smaller twin).
cently, Hanley et al12 noted that almost one half of monochorionic twins with a velamentous umbilical cord insertion were discordant in birth weight (defined as ≥20%). Placental abruption is a chronic process that has its origins early in pregnancy and perhaps extend to the earliest stages of conception.13 In light of the increased rates of placental infarctions and thrombosis in the placentae of abruption pregnancies compared with otherwise normal pregnancies, our finding of increased risks of stillbirth and neonatal deaths among abruption births, regardless of birth weight discordancy, is not surprising. The pres-
ence of disseminated intravascular coagulation as a result of the release of trophoblastic material or infarction because of emboli from the dead twin are thought to result in the death of the co-twin.23 However, these events are more likely to occur in monochorionic than among dichorionic twins. Discordancy in size at birth has been associated frequently with growth restriction of either twin and, less commonly, with both twins within a sibship.24 A birth weight discordancy of ≥20% is associated with one twin being growth restricted in approximately one half of all
Ananth, Demissie, and Hanley 959
Volume 188, Number 4 Am J Obstet Gynecol
twins.7,24 In our study, for a birth weight discordancy of ≥20%, we observed 53% and 62.8% of growth restriction in one or both twins who were of the same sex and different sexes, respectively (Table V). In the presence of placental abruption, these risks were further elevated, which implies that the excess deaths that accompany abruption perhaps result from the chronic disease process that leads to vasculopathy, placental underperfusion, or both. Furthermore, the increased relative risks for stillbirths of same sex twins in the presence of placental abruption may also reflect the well-known twin-twin transfusion syndrome.25 The vital records data files, unfortunately, do not contain information on placental chorionicity. The strengths of our study include its population-based data with a large cohort of twins who were delivered recently in the United States. We were able to examine the role of placental abruption in the cause of birth weight discordancy, with control for important confounding factors. Some limitation of the study warrant attention: the unavailability of placental chorionicity, which is an important determinant of twin birth weight discordancy. Twins who result from monochorionic placentation have greater fetal and neonatal insults, which include stillbirth and neonatal death. Furthermore, preterm birth is also more common among monochorionic than dichorionic twins.11 Because placental chorionicity is not recorded on vital statistics data, we instead stratified all our analysis on whether the twins were of the same or different sex. The possibility of some residual confounding because of unmeasured risk factors (such as cocaine use, chorioamnionitis, and history of placental abruption) for adverse outcomes remains. Unfortunately, the vital statistics databases do not contain information on these risk factors. Another important limitation of our study that chiefly pertains to the analysis of stillbirths relates to the estimation of birth weight. The association between birth weight discordancy and stillbirth (Table II) is perhaps overestimated because birth weight was recorded at the time of birth (delivery) rather than time of (in utero) fetal demise. This overestimation may be more pronounced in cases of stillbirth of one twin as a result of maceration, which thereby allows greater birth weight discordancy. A study of the expectant treatment of twin pregnancies with a single fetal death showed that 69% of these mothers experienced spontaneous labor within a week after the death of a fetus, and 86% of these mothers experienced spontaneous labor within 5 weeks.26 Another study that examined the placental histologic condition of 150 stillborn infants estimated that about 80% of stillbirths were delivered within a week of the fetal demise.27 The analysis that pertains to fetal deaths must therefore be interpreted cautiously. The risk of stillbirth and neonatal deaths among nonabruption births increases with increasing birth weight discordancy, both among twins of same and differ-
ent sex. With evidence of placental abruption, these risks are further elevated. Not only does the frequency of birth weight discordancy increase in the presence of abruption, but the co-occurrence of these two conditions also results in a significant proportion of unfavorable perinatal outcomes. The role of placental abruption in compounding the risks of adverse perinatal outcomes in the presence of birth weight discordancy among twins is important in the understanding of the pathogenesis of discordant size of twin gestations.
REFERENCES
1. Pharaoh POD. Consequences of in-utero death in a twin pregnancy. Lancet 2000;355:1597-602. 2. West CR, Adi Y, Pharoah POD. Fetal and infant death in monoand dizygotic twins in England and Wales 1982-91. Arch Dis Child 1999;80:F217-20. 3. Rydhstroem H. Pregnancy with stillbirth of both twins. Br J Obstet Gynaecol 1996;103:25-32. 4. Demissie K, Ananth CV, Martin J, Hanley ML, MacDorman MF, Rhoads GG. Fetal and neonatal mortality among twin gestations in the United States: the role of intrapair birthweight discordance. Obstet Gynecol 2002;100:474-80. 5. Cooperstock MS, Tummaru R, Bakewell J, Schramm W. Twin birth weight discordance and risk of preterm birth. Am J Obstet Gynecol 2000;183:63-7. 6. Fraser D, Picard R, Ricard E, Leiberman JR. Birth weight discordance and intrauterine growth retardation and perinatal outcomes in twins. J Reprod Med 1994;39:504-8. 7. O’Brien WF, Knuppel RA, Scerbo JC, Rattan PK. Birth weight in twins: an analysis of discordancy and growth retardation. Obstet Gynecol 1986;67:483-6. 9. Blickstein I, Shoham-Schwartz Z, Lancet M, Borenstein R. Characterization of the growth-discordant twin. Obstet Gynecol 197;70:11-5. 9. Gauschino S, Spinillo A, Stola E, Pesando PC. Growth retardation, size at birth and perinatal mortality in twin gestations. Int J Gynaecol Obstet 1987;25:399-403. 10. Eberle AM, Levesque D, Vintzileos AM, Egan JFX, Tsapanos V, Salafia C. Placental pathology in discordant twins. Am J Obstet Gynecol 1993;169:931-5. 11. Victoria A, Mora G, Arias F. Perinatal outcome, placental pathology, and severity of discordance in monochorionic and dichorionic twins. Obstet Gynecol 2001;97:310-5. 12. Hanley ML, Ananth CV, Shen-Schwarz S, Smulian JC, Lai Y-L, Vintzileos AM. Placental cord insertion and birth weight discordancy in twin gestations. Obstet Gynecol 2002;99:477-82. 13. Ananth CV, Wilcox AJ. Placental abruption and perinatal mortality in the United States. Am J Epidemiol 2001;153:332-7. 14. National Center for Health Statistics. 1995-97 Matched multiple birth data set: National Center for Health Statistics CD-ROM series 21, number 12. Hyattsville (MD): Department of Health and Human Services, Centers for Disease Control and Prevention; 2000. 15. Taffel S, Johnson D, Heuse R. A method of imputing length of gestation on birth certificates. Vital Health Stat 2 1982;93:1-11. 16. Taffel SM, Ventura SJ, Gay GA. Revised US certificate of birth— new opportunities for research on birth outcome. Birth 1989;16:188-93. 17. Zhang J, Yu FK. What’s the relative risk? A method of correcting the odds ratio in cohort studies of common outcomes. JAMA 1998;280:1690-1. 18. Zeger SL, Liang K-Y. Longitudinal data analysis for discrete and continuous outcomes. Biometrics 1986;42:121-30. 19. Rothman KJ, Greenland S. Modern epidemiology. 2nd edition. Philadelphia: Lippincott–Williams & Wilkins; 1998. 20. Rizzo G, Arduini D, Romanini C. Cardiac and extracardiac flows in discordant twins. Am J Obstet Gynecol 1994;170:1321-7.
960 Ananth, Demissie, and Hanley
April 2003 Am J Obstet Gynecol
21. Bajoria R, Wee LY, Anwar S, Ward S. Outcome of twin pregnancies complicated by single intrauterine death in relation to vascular anatomy of the monochorionic placenta. Hum Reprod 1999;14:2124-30. 22. Wenstrom KD, Tessen JA, Zlatnik FJ, Sipes SL. Frequency, distribution, and theoretical mechanisms of hematologic and weight discordance in monochorionic twins. Obstet Gynecol 1992; 80:257-61. 23. Moore CM, McAdams AJ, Sutherland J. Intrauterine disseminated intravascular coagulation: a syndrome of multiple pregnancy with a dead twin fetus. J Pediatr 1969;74:523-8.
O
24. Blickstein I, Lancet M. The growth discordant twin. Obstet Gynecol Surv 1988;43:509-15. 25. Vintzileos AM, Rodis JF. Growth discordance in twins. In: Divon MY, editor. Abnormal fetal growth. New York: Elsevier Science; 1991. 26. Santema JG, Swaak AM, Wallenburg HCS. Expectant management of twin pregnancy with single fetal death. Br J Obstet Gynaecol 1995;102:26-30. 27. Genest DR, Williams MA, Greene MF. Estimating the time of death in stillborn fetuses, I: histologic evaluation of fetal organs; an autopsy study of 150 stillborns. Obstet Gynecol 1992;80:575-84.
N THE MOVE? Send us your new address at least six weeks ahead
Don’t miss a single issue of the journal! To ensure prompt service when you change your address, please photocopy and complete the form below. Please send your change of address notification at least six weeks before your move to ensure continued service. We regret we cannot guarantee replacement of issues missed due to late notification. JOURNAL TITLE: Fill in the title of the journal here.
OLD ADDRESS: Affix the address label from a recent issue of the journal here.
NEW ADDRESS:
Clearly print your new address here. Name Address City/State/ZIP
COPY AND MAIL THIS FORM TO: Mosby Subscription Customer Service 6277 Sea Harbor Dr Orlando, FL 32887
OR FAX TO: 407-363-9661
OR PHONE: 1-(800)654-2452 Outside the U.S., call (407)345-4000
Birth weight discordancy and adverse perinatal outcomes among twin gestations in the United States: The effect of placental abruption Cande V. Ananth, PhD, MPH,a,c Kitaw Demissie, MD, PhD,c and Maryellen L. Hanley, MD, MPHb New Brunswick and Piscataway, NJ OBJECTIVE: We evaluated whether the relationship between birth weight discordancy of twins and stillbirth, neonatal deaths, and preterm births was modified by the presence of abruption. STUDY DESIGN: We used the 1995 to 1997 matched multiple birth file for United States twin births (n = 269,287). Birth weight discordancy was defined as the ratio of the difference in birth weight of the heavier from the lighter twin to that of the heavier twin and was categorized as <5%, 5% to 9%, 10% to 14%, 15% to 19%, 20% to 29%, 30% to 39%, and 40%. We evaluated the risks of stillbirth (20 weeks of gestation), neonatal deaths (within 28 days after birth), and preterm birth (< 32 weeks) in the presence and absence of abruption. Associations between birth weight discordancy and these perinatal outcomes were expressed as adjusted relative risks and were derived from multivariable logistic regression models, based on the method of generalized estimating equations. Risk of these outcomes were derived for each stratum of birth weight discordancy and abruption status, with the <5% birth weight discordancy, nonabruption status labeled as the reference group. All analyses were performed separately for same and different sex twins. RESULTS: A birth weight discordancy of 20% among same sex (adjusted relative risk, 1.2; 95% CI, 1.1, 1.4), and 40% among different sex twins (relative risk, 2.2; 95% CI, 1.7, 2.8) conferred increased risk for abruption. Among nonabruption births, a birth weight discordancy of 15% among same sex and 30% among different sex twins increased the risk of stillbirths, neonatal deaths, and preterm births. Among abruption births, however, the risks were increased even in the lowest birth weight discordancy category (<5%). The relative risks of stillbirths and neonatal deaths among abruption births were significantly higher for each birth weight discordancy group, both for same and different sex twins, compared with the reference group. The association between birth weight discordancy and preterm birth was not modified by either the presence or absence of abruption. CONCLUSION: Birth weight discordancy of 15% for same sex and 30% for different sex confer greatest risk of adverse perinatal outcomes in the absence of abruption. In the presence of placental abruption, these risks are further compounded. The results underscore the need for careful monitoring of twin pregnancies. (Am J Obstet Gynecol 2003;188:954-60.)
Key words: Placental abruption, stillbirth, neonatal death, twins, birth weight discordancy Abnormal discordance in size at birth between twins within a sibship has been defined variously, from 15% to as much as 40%.1-3 Extreme birth weight discordancy in twins has been associated with adverse perinatal outcomes that include a higher risk of fetal and neonatal From the Section of Epidemiology and Biostatisticsa and the Division of Maternal-Fetal Medicine, Department of Obstetrics, Gynecology, and Reproductive Sciences,b University of Medicine and Dentistry of New Jersey, Robert Wood Johnson Medical School/Saint Peter’s University Hospital, and the Division of Epidemiology, School of Public Health, University of Medicine and Dentistry of New Jersey.c Supported in part by National Institutes of Health grant No. IR01HD038902 (C. V. A.). Received for publication July 22, 2002; revised October 18, 2002; accepted December 3, 2002. Reprints not available from the authors. © 2003, Mosby, Inc. All rights reserved. 0002-9378/2003 $30.00 + 0 doi:10.1067/mob.2003.210
954
deaths4 and preterm birth.3-6 The discordancy in size is often the consequence of fetal growth restriction of the smaller twin.7-12 The causes that are responsible for fetal growth restriction are multifactorial and varied. Birth weight discordancy may be a consequence of abnormal placental disease, such as infarctions that often result in growth restricted babies. We recently demonstrated that, among singleton pregnancies, women with placental abruption are at greater likelihood of delivering a growth restricted infant compared with otherwise normal pregnancies.13 These observations led us to speculate that placental abruption was the end result of an underlying chronic disease process, with origins early in pregnancy, and perhaps even extending to the time of conception.13 These chronic processes and placental abruption can also contribute independently to birth weight discordancy and, when present together, can further elevate
Ananth, Demissie, and Hanley 955
Volume 188, Number 4 Am J Obstet Gynecol
Table I. Relationship between birth weight discordancy and placental abruption: United States twin births 1995 through 1997 Birth weight discordancy
Total twin births
Placental abruption N (%)
Relative risk (95% confidence interval)
Same sex (%) <5 5-9 10-14 15-19 20-29 30-39 ≥40 Total
55,969 43,887 31,183 19,528 18,716 6,073 5,665 181,021
640 (11.4) 539 (12.3) 351 (11.4) 248 (12.7) 258 (13.8) 104 (17.1) 140 (24.7) 2,286 (12.6)
1.0 (Reference) 1.1 (0.9-1.1) 1.0 (0.9-1.1) 1.1 (0.9-1.3) 1.2 (1.1-1.4) 1.5 (1.2-1.9) 2.2 (1.8-2.6) —
Different sex (%) <5 5-9 10-14 15-19 20-29 30-39 ≥40 Total
23,798 20,369 16,657 11,331 10,826 3,036 2,249 88,266
362 (15.2) 314 (15.4) 202 (12.1) 173 (15.3) 145 (13.4) 43 (14.2) 73 (32.5) 1,312 (14.9)
1.0 (Reference) 1.0 (0.9-1.2) 0.8 (0.7-0.9) 1.0 (0.8-1.2) 0.9 (0.7-1.1) 0.9 (0.71.3) 2.2 (1.7-2.8) —
Rate of placental abruption per 1,000 twin births. Relative risks were adjusted for maternal age, primigravida, maternal education, prenatal care, chronic hypertension, and infant size (smaller or larger twin).
risks of adverse perinatal outcomes to birth weight discordant twins. We evaluated whether twin birth weight discordancy for the risk of fetal and neonatal deaths and preterm birth is modified by the presence or absence of placental abruption. We also evaluated whether the small for gestational age status of one or both twins contributed to the increased risks of these adverse perinatal outcomes. Material and methods We used the “matched multiple birth file” data for all twin births in the United States (1995-1997).14 These data were assembled by the National Center for Health Statistics of the Centers for Disease Control and Prevention and include data from birth and fetal and infant death certificates in twin and higher order pregnancies. The National Center for Health Statistics used a three-stage matching algorithm (that identified >99% of all twin births within a sibship) to match fetuses within the twin and higher order pregnancies, the details of which have been documented elsewhere.14 We restricted our analysis to twin gestations. We abstracted data for each fetus (ie, fetal and infant death or live birth, birth weight and sex of the fetus for each twin pair, gestational age, and other sociodemographic and medical and obstetric complications of pregnancy). Birth weight discordancy was defined as the ratio of the difference in birth weight of the lighter twin from the heavier twin to that of the heavier twin. The discordancy was then categorized as 0% to 4%, 5% to 9%, 10% to 14%, 15% to 19%, 20% to 29%, 30% to 39% and ≥40%. Gestational age assignment was based on the date of last menstrual period predominantly. In approximately 5% of births,
the day and month of the last menstrual period was missing; therefore, it was imputed.15 Inconsistent gestational age was replaced by a clinical estimate in a small fraction of births (<5%).16 The replacement of clinical estimate and the gestational age imputations were performed by the National Center for Health Statistics.15,16 We evaluated the risks of stillbirth (≥20 weeks), neonatal death (within 28 days after birth), and preterm birth (<32 weeks). Analysis pertaining to preterm births was restricted to twin live births. The risk of stillbirth between abruption and nonabruption births by birth weight discordancy was compared when either twin was small for gestational age (one twin being small for gestational age and the other appropriately grown) and when both twins were grown appropriately. Small for gestational age births were defined as birth weight below the 10th percentile for gestational age. This nomogram was based on all twin live births who were delivered in the United States between 1995 and 1997. Relative risks (with associated 95% CIs) for adverse outcomes were calculated for each birth weight discordancy category and abruption status with a birth weight discordancy of <5%; the nonabruption category served as the reference group. Adjusted odds ratios were derived from fitting multivariable logistic regression models and were transformed to adjusted relative risks.17 Confounders were considered for adjustment in the model if they changed the unadjusted relative risk by at least 10%. Potential confounders that were considered for adjustment included maternal age (<20, 20-24, 25-29, 30-34, and ≥35 years), gravidity (primigravida or gravida ≥2), maternal race/ethnicity (white, black, or other race/ethnicity), maternal education (expressed as completed years of school-
956 Ananth, Demissie, and Hanley
April 2003 Am J Obstet Gynecol
Table II. Relationship among birth weight discordancy, stillbirths, and placental abruption: United States twin births 1995 through 1997 Nonabruption births
Placental abruption births
Relative risk (95% CI)
Population attributable fraction (%)
Same sex (%) <5 605 (10.9) 5-9 377 (8.7) 10-14 347 (11.3) 15-19 310 (16.1) 20-29 491 (26.6) 30-39 362 (60.6) ≥40 973 (176.1) Total 3465 (19.4)
1.0 (Reference) 0.8 (0.7-0.9) 1.1 (0.9-1.2) 1.5 (1.3-1.7) 2.6 (2.3-2.9) 6.1 (5.3-7.0) 19.4 (17.4-21.6) —
Different sex (%) <5 118 (5.0) 5-9 102 (5.1) 10-14 84 (5.1) 15-19 64 (5.7) 20-29 66 (6.2) 30-39 64 (21.4) ≥40 282 (129.6) Total 780 (9.2)
1.0 (Reference) 1.0 (0.8-1.3) 1.0 (0.8-1.4) 1.2 (0.9-1.6) 1.3 (0.9-1.7) 4.5 (3.3-6.1) 29.7 (23.8-37.0) —
Birth weight discordancy
N (%)
N (%)
Relative risk (95% CI)
Population attributable fraction (%)
— — 0.1 0.5 1.6 5.1 16.7 —
24 (37.5) 22 (40.8) 15 (42.0) 13 (52.4) 32 (124.0) 11 (105.8) 36 (257.1) 153 (66.9)
3.5 (2.3-5.3) 3.9 (2.5-6.0) 4.1 (2.4-6.9) 5.0 (2.8-8.8) 13.0 (8.8-19.0) 11.0 (5.8-20.8) 32.8 (22.1-48.5) —
2.7 3.0 3.2 4.2 11.4 9.6 24.9 —
— — — 0.1 0.1 1.7 12.2 —
8 (22.1) 8 (25.5) 13 (64.4) 9 (52.0) 6 (41.4) 4 (93.0) 16 (219.2) 64 (48.8)
4.6 (2.2-9.5) 5.5 (2.7-11.4) 13.0 (7.2-23.5) 11.8 (5.9-23.8) 8.7 (3.7-20.1) 21.8 (7.6-66.5) 56.6 (31.3-102.3) —
1.7 2.1 5.9 4.8 3.7 8.8 21.5 —
Stillbirth rates per 1,000 twin births. Relative risks were adjusted for maternal age, primigravida, maternal education, prenatal care, chronic hypertension, and infant size (larger or smaller twin).
ing, <12 or ≥13 years), cigarette smoking and alcohol use during pregnancy, prenatal care (any vs no care), infant size (smaller or larger twin), and chronic hypertension (diagnosed before pregnancy or before 20 weeks of gestation). Smoking and alcohol use were analyzed as indicator variables (yes/no).16 Adjustment for confounding factors are noted in the results. Because twin births violate the statistical assumption of independence of observations, all analyses were based on the method of generalized estimating equations.18 These data do not contain details on chorionicity or zygosity of the twin fetuses within a sibship. Therefore, we performed all analyses separately for same sex and different sex twins and adjusted for twin size (smaller or larger twin) in all the regression models. Finally, we calculated the population attributable fraction,19 which was defined as the proportion of the adverse outcome that can be attributed to birth weight discordancy in the abruption and nonabruption groups. During the period of 1995 to 1997, there were 308,013 twin fetuses delivered in the United States. We excluded 3,602 twins who were delivered before 20 weeks and 35,124 twins with missing information about placental abruption; 269,287 twin births were available for analyses. The study was approved by the institutional review board of the Robert Wood Johnson Medical School. Results Placental abruption was recorded in 3,598 of 269,287 twin births (13.4 per 1000 births). The prevalence rate
was higher among twins of different sex (14.9 per 1000 births) than among same sex twins (12.6 per 1000 births, Table I). Among same sex twins, placental abruption was associated with ≥20% birth weight discordancy, whereas among different sex twins an increased risk was apparent only among the extremely discordant twins (≥40%). The overall stillbirth rate among nonabruption births was >2-fold greater among same sex (19.4 per 1,000 births) compared with different sex twins (9.2 per 1,000 births, Table II). Among same sex twins, increasing birth weight discordancy of >15% was associated with increasing rates of stillbirth in the nonabruption group. Among same sex twins with placental abruption, the risk of stillbirths was high even among twins with the lowest discordancy. Stillbirths were increased among different sex twins only with a birth weight discordancy of ≥30% in the nonabruption group. However, in the placental abruption group, stillbirth rates were increased even in the smallest birth weight discordancy categories. In the nonabruption group, neonatal mortality rates were 24.9 and 19.0 per 1000 live births among same and different sex twins, respectively (Table III). These rates were approximately 3- to 4-fold higher among those infants born to women with placental abruption. Once again, at a birth weight discordancy of ≥15%, the risk of neonatal deaths was increased for same sex twins of a nonabruption pregnancy. At a birth weight discordancy of ≥30%, the risk of neonatal death was increased for different sex twins of nonabruption pregnancies.
Ananth, Demissie, and Hanley 957
Volume 188, Number 4 Am J Obstet Gynecol
Table III. Relationship among birth weight discordancy, neonatal deaths, and placental abruption: United States twin births 1995 to 1997 Nonabruption births
Placental abruption births
N (%)
Relative risk (95% CI)
Population attributable fraction (%)
883 (16.2) 663 (15.5) 483 (15.9) 310 (16.4) 399 (22.4) 227 (40.9) 1255 (279.8) 4220 (24.9)
1.0 (Reference) 1.0 (0.9-1.1) 1.0 (0.9-1.1) 1.0 (0.9-1.2) 1.5 (1.3-1.7) 2.8 (2.4-3.1) 25.0 (22.7-27.4) —
Different sex (%) <5 320 (13.8) 5-9 264 (13.3) 10-14 218 (13.4) 15-19 160 (14.5) 20-29 145 (13.7) 30-39 53 (18.2) ≥40 453 (242.4) Total 1613 (19.0)
1.0 (Reference) 1.0 (0.8-1.2) 1.0 (0.8-1.2) 1.1 (0.9-1.3) 1.0 (0.9-1.3) 1.4 (1.1-1.9) 24.3 (20.8-28.4) —
Birth weight discordancy Same sex (%) <5 5-9 10-14 15-19 20-29 30-39 ≥40 Total
N (%)
Relative risk (95% CI)
Population attributable fraction (%)
— — — 0.1 0.7 2.6 26.8 —
47 (77.0) 31 (60.7) 29 (85.3) 17 (73.0) 15 (67.9) 10 (108.7) 45 (436.9) 194 (91.1)
5.1 (3.7-6.9) 4.0 (2.8-5.9) 6.0 (4.0-8.8) 4.7 (2.9-7.8) 4.6 (2.7-7.7) 7.9 (4.1-15.4) 53.1 (35.4-79.5) —
6.2 4.6 7.1 5.8 5.3 9.5 42.9 —
— — — 0.1 0.1 0.6 23.2 —
23 (66.3) 19 (62.7) 6 (31.9) 14 (86.6) 13 (94.2) 3 (76.9) 23 (410.7) 101 (81.9)
5.3 (3.4-8.2) 5.2 (3.2-8.3) — 7.5 (4.3-13.2) 8.4 (4.7-15.1) — 54.3 (31.1-94.9) —
5.4 5.1 — 7.5 8.3 — 40.3 —
Neonatal death rates per 1,000 twin live births. Relative risks were adjusted for maternal age, primigravida, maternal education, prenatal care, chronic hypertension, and infant size (larger or smaller twin).
Approximately one fifth of all twin births were delivered preterm (<32 weeks of gestation) among nonabruption births, both for same sex and different sex twins (Table IV). However, more than two thirds of pregnancies were delivered before 32 completed weeks of gestation among abruption births. Preterm birth rates were increased slightly for a birth weight discordancy of ≥20% among same sex twins and ≥30% among different sex twins in the nonabruption group. Among placental abruption births, however, the risk of delivering preterm was apparent even for the lowest discordancy groups. On the basis of the analysis of birth weight discordancy and stillbirths (Table II), we evaluated the risk of stillbirths for a weight discordancy of ≥20%, stratified by both twins being small for gestational age, one twin being small for gestational age, or both twins being grown appropriately (Table V). When both twins were appropriately grown and were of the same sex, the stillbirth rate was 11.7 per 1,000 births for a birth weight discordancy of <20% in the nonabruption group. Comparable data when either twin was small for gestational age was 34.4 per 1,000 births. Similarly, the rates among different sex twins were 6.0 and 19.9 per 1,000 births, respectively. For the same sex twins, either a birth weight discordancy of ≥20% or the presence of abruption conferred relative risks for stillbirths of similar magnitude. However, for different sex twins, abruption conferred greater relative risks for stillbirths. Finally, the presence of birth weight discordancy accompanied by abruption conferred the
greatest risk for stillbirths, both for appropriately grown twins and when either twin was growth restricted. Comment This is perhaps the largest study to date to report the effects of birth weight discordancy in twins in regard to the risks of stillbirth, neonatal deaths, and preterm births among women with and without placental abruption. Overall, a birth weight discordancy of ≥15% among same sex and ≥30% among different sex twins confer the greatest risks of stillbirth and neonatal deaths. In the presence of placental abruption, the risks of stillbirth and neonatal deaths were increased even among the smallest weight discordant twins. When both twins were small for gestational age, the risk of stillbirth was 7-fold greater at a birth weight discordancy of ≥20% in the presence of placental abruption. Birth weight discordancy in twins has different underlying causes that are based on placental chorionicity. In the dichorionic twins, the discordancy is thought to arise because of uterine overcrowding, increased placental mass, placental insufficiency,11 and genetic differences in the growth potential. Among monochorionic twins, the discordancy may be the result of hemodynamic imbalance because of vascular anastomoses,10,11 fetal anemia,20,21 or twin-twin transfusion syndrome.20-22 Furthermore, abnormalities of the placenta may also contribute disproportionately to adverse perinatal outcomes in monochorionic versus dichorionic placentae. Re-
958 Ananth, Demissie, and Hanley
April 2003 Am J Obstet Gynecol
Table IV. Relationship among birth weight discordancy, preterm births (<32 weeks of gestation), and placental abruption: United States twin births 1995 through 1997 Nonabruption births
Placental abruption births
N (%)
Relative risk (95% CI)
Population attributable fraction (%)
Same sex (%) <5 5-9 10-14 15-19 20-29 30-39 ≥40 Total
6,043 (18.6) 4,395 (17.5) 3,114 (17.5) 2,045 (18.5) 2,474 (24.3) 1,305 (42.6) 2,756 (74.1) 22,132 (21.4)
1.0 (Reference) 0.9 (0.9-1.0) 0.9 (0.8-1.0) 1.0 (0.9-1.1) 1.3 (1.3-1.4) 2.3 (2.2-2.4) 4.0 (3.9-4.1) —
Different sex (%) <5 5-9 10-14 15-19 20-29 30-39 ≥40 Total
2,331 (16.9) 1,927 (16.2) 1,557 (16.2) 1,018 (15.7) 953 (15.9) 375 (24.6) 935 (67.1) 9,096 (17.9)
1.0 (Reference) 1.0 (0.9-1.0) 1.0 (0.9-1.0) 0.9 (0.8-1.0) 1.0 (0.9-1.0) 1.5 (1.3-1.6) 4.0 (3.8-4.1) —
Birth weight discordancy
N (%)
Relative risk (95% CI)
Population attributable fraction (%)
— — — — 5.3 19.5 35.8 —
265 (67.3) 195 (64.8) 143 (68.1) 104 (81.9) 121 (72.5) 47 (72.3) 106 (94.6) 981 (71.3)
3.6 (3.4-3.9) 3.5 (3.2-3.8) 3.7 (3.4-4.0) 4.4 (4.0-4.7) 3.9 (3.5-4.2) 3.9 (3.3-4.4) 5.1 (4.8-5.2) —
32.6 31.7 33.4 38.7 35.0 35.0 43.3 —
— — — — — 7.8 33.6 —
123 (59.4) 101 (65.2) 62 (54.4) 53 (54.6) 50 (58.1) 14 (45.2) 46 (95.8) 449 (60.8)
3.6 (3.2-4.0) 4.0 (3.5-4.4) 3.2 (2.6-3.7) 3.4 (2.8-4.0) 3.5 (2.8-4.1) 2.8 (1.8-3.8) 5.7 (5.1-5.9) —
30.5 33.6 27.1 27.2 29.7 23.3 44.3 —
Relative risks were adjusted for maternal age, primigravida, maternal education, prenatal care, chronic hypertension, and infant size (larger or smaller twin).
Table V. Rate and relative risk for stillbirths by birth weight discordancy, placental abruption, and small for gestational age status in twin sibships: United States twin births 1995 through 1997 Both twins normal
Stillbirth
N (%)
Relative risk (95% CI)
One twin small for gestational age
Both twins small for gestational age
N (%)
Relative risk (95% CI)
N (%)
Relative risk (95% CI)
Same sex <20%, Nonabruption <20%, Abruption ≥20%, Nonabruption ≥20%, Abruption
1583 (11.7) 70 (43.3) 633 (44.4) 28 (125.6)
1.0 (Reference) 3.5 (3.1-3.9) 1.3 (1.0-1.6) 4.6 (2.8-7.5)
237 (25.1) 8 (70.2) 583 (47.4) 27 (153.4)
1.0 (Reference) 1.5 (0.7-3.5) 2.1 (1.7-2.5) 5.5 (3.4-8.9)
299 (48.9) 9 (78.9) 136 (113.3) 5 (312.5)
1.0 (Reference) 0.8 (0.3-1.6) 3.1 (2.4-4.0) 2.7 (0.9-8.9)
Different sex <20%, Nonabruption <20%, Abruption ≥20%, Nonabruption ≥20%, Abruption
397 (6.0) 35 (37.1) 102 (11.2) 8 (57.1)
1.0 (Reference) 2.7 (2.1-3.5) 2.1 (1.4-3.2) 2.9 (1.2-7.2)
63 (16.1) 4 (56.3) 155 (27.9) 3 (51.7)
1.0 (Reference) 2.1 (1.5-2.9) 2.8 (0.9-8.9) 3.5 (1.1-12.1)
53 (24.7) 4 (60.6) 17 (50.9) 3 (166.7)
1.0 (Reference) 3.5 (1.9-6.4) 2.5 (0.8-7.7) 7.1 (1.7-31.5)
Stillbirth rates per 1,000 twin births. Relative risks were adjusted for maternal age, primigravida, maternal education, prenatal care, chronic hypertension, and infant size (larger or smaller twin).
cently, Hanley et al12 noted that almost one half of monochorionic twins with a velamentous umbilical cord insertion were discordant in birth weight (defined as ≥20%). Placental abruption is a chronic process that has its origins early in pregnancy and perhaps extend to the earliest stages of conception.13 In light of the increased rates of placental infarctions and thrombosis in the placentae of abruption pregnancies compared with otherwise normal pregnancies, our finding of increased risks of stillbirth and neonatal deaths among abruption births, regardless of birth weight discordancy, is not surprising. The pres-
ence of disseminated intravascular coagulation as a result of the release of trophoblastic material or infarction because of emboli from the dead twin are thought to result in the death of the co-twin.23 However, these events are more likely to occur in monochorionic than among dichorionic twins. Discordancy in size at birth has been associated frequently with growth restriction of either twin and, less commonly, with both twins within a sibship.24 A birth weight discordancy of ≥20% is associated with one twin being growth restricted in approximately one half of all
Ananth, Demissie, and Hanley 959
Volume 188, Number 4 Am J Obstet Gynecol
twins.7,24 In our study, for a birth weight discordancy of ≥20%, we observed 53% and 62.8% of growth restriction in one or both twins who were of the same sex and different sexes, respectively (Table V). In the presence of placental abruption, these risks were further elevated, which implies that the excess deaths that accompany abruption perhaps result from the chronic disease process that leads to vasculopathy, placental underperfusion, or both. Furthermore, the increased relative risks for stillbirths of same sex twins in the presence of placental abruption may also reflect the well-known twin-twin transfusion syndrome.25 The vital records data files, unfortunately, do not contain information on placental chorionicity. The strengths of our study include its population-based data with a large cohort of twins who were delivered recently in the United States. We were able to examine the role of placental abruption in the cause of birth weight discordancy, with control for important confounding factors. Some limitation of the study warrant attention: the unavailability of placental chorionicity, which is an important determinant of twin birth weight discordancy. Twins who result from monochorionic placentation have greater fetal and neonatal insults, which include stillbirth and neonatal death. Furthermore, preterm birth is also more common among monochorionic than dichorionic twins.11 Because placental chorionicity is not recorded on vital statistics data, we instead stratified all our analysis on whether the twins were of the same or different sex. The possibility of some residual confounding because of unmeasured risk factors (such as cocaine use, chorioamnionitis, and history of placental abruption) for adverse outcomes remains. Unfortunately, the vital statistics databases do not contain information on these risk factors. Another important limitation of our study that chiefly pertains to the analysis of stillbirths relates to the estimation of birth weight. The association between birth weight discordancy and stillbirth (Table II) is perhaps overestimated because birth weight was recorded at the time of birth (delivery) rather than time of (in utero) fetal demise. This overestimation may be more pronounced in cases of stillbirth of one twin as a result of maceration, which thereby allows greater birth weight discordancy. A study of the expectant treatment of twin pregnancies with a single fetal death showed that 69% of these mothers experienced spontaneous labor within a week after the death of a fetus, and 86% of these mothers experienced spontaneous labor within 5 weeks.26 Another study that examined the placental histologic condition of 150 stillborn infants estimated that about 80% of stillbirths were delivered within a week of the fetal demise.27 The analysis that pertains to fetal deaths must therefore be interpreted cautiously. The risk of stillbirth and neonatal deaths among nonabruption births increases with increasing birth weight discordancy, both among twins of same and differ-
ent sex. With evidence of placental abruption, these risks are further elevated. Not only does the frequency of birth weight discordancy increase in the presence of abruption, but the co-occurrence of these two conditions also results in a significant proportion of unfavorable perinatal outcomes. The role of placental abruption in compounding the risks of adverse perinatal outcomes in the presence of birth weight discordancy among twins is important in the understanding of the pathogenesis of discordant size of twin gestations.
REFERENCES
1. Pharaoh POD. Consequences of in-utero death in a twin pregnancy. Lancet 2000;355:1597-602. 2. West CR, Adi Y, Pharoah POD. Fetal and infant death in monoand dizygotic twins in England and Wales 1982-91. Arch Dis Child 1999;80:F217-20. 3. Rydhstroem H. Pregnancy with stillbirth of both twins. Br J Obstet Gynaecol 1996;103:25-32. 4. Demissie K, Ananth CV, Martin J, Hanley ML, MacDorman MF, Rhoads GG. Fetal and neonatal mortality among twin gestations in the United States: the role of intrapair birthweight discordance. Obstet Gynecol 2002;100:474-80. 5. Cooperstock MS, Tummaru R, Bakewell J, Schramm W. Twin birth weight discordance and risk of preterm birth. Am J Obstet Gynecol 2000;183:63-7. 6. Fraser D, Picard R, Ricard E, Leiberman JR. Birth weight discordance and intrauterine growth retardation and perinatal outcomes in twins. J Reprod Med 1994;39:504-8. 7. O’Brien WF, Knuppel RA, Scerbo JC, Rattan PK. Birth weight in twins: an analysis of discordancy and growth retardation. Obstet Gynecol 1986;67:483-6. 9. Blickstein I, Shoham-Schwartz Z, Lancet M, Borenstein R. Characterization of the growth-discordant twin. Obstet Gynecol 197;70:11-5. 9. Gauschino S, Spinillo A, Stola E, Pesando PC. Growth retardation, size at birth and perinatal mortality in twin gestations. Int J Gynaecol Obstet 1987;25:399-403. 10. Eberle AM, Levesque D, Vintzileos AM, Egan JFX, Tsapanos V, Salafia C. Placental pathology in discordant twins. Am J Obstet Gynecol 1993;169:931-5. 11. Victoria A, Mora G, Arias F. Perinatal outcome, placental pathology, and severity of discordance in monochorionic and dichorionic twins. Obstet Gynecol 2001;97:310-5. 12. Hanley ML, Ananth CV, Shen-Schwarz S, Smulian JC, Lai Y-L, Vintzileos AM. Placental cord insertion and birth weight discordancy in twin gestations. Obstet Gynecol 2002;99:477-82. 13. Ananth CV, Wilcox AJ. Placental abruption and perinatal mortality in the United States. Am J Epidemiol 2001;153:332-7. 14. National Center for Health Statistics. 1995-97 Matched multiple birth data set: National Center for Health Statistics CD-ROM series 21, number 12. Hyattsville (MD): Department of Health and Human Services, Centers for Disease Control and Prevention; 2000. 15. Taffel S, Johnson D, Heuse R. A method of imputing length of gestation on birth certificates. Vital Health Stat 2 1982;93:1-11. 16. Taffel SM, Ventura SJ, Gay GA. Revised US certificate of birth— new opportunities for research on birth outcome. Birth 1989;16:188-93. 17. Zhang J, Yu FK. What’s the relative risk? A method of correcting the odds ratio in cohort studies of common outcomes. JAMA 1998;280:1690-1. 18. Zeger SL, Liang K-Y. Longitudinal data analysis for discrete and continuous outcomes. Biometrics 1986;42:121-30. 19. Rothman KJ, Greenland S. Modern epidemiology. 2nd edition. Philadelphia: Lippincott–Williams & Wilkins; 1998. 20. Rizzo G, Arduini D, Romanini C. Cardiac and extracardiac flows in discordant twins. Am J Obstet Gynecol 1994;170:1321-7.
960 Ananth, Demissie, and Hanley
April 2003 Am J Obstet Gynecol
21. Bajoria R, Wee LY, Anwar S, Ward S. Outcome of twin pregnancies complicated by single intrauterine death in relation to vascular anatomy of the monochorionic placenta. Hum Reprod 1999;14:2124-30. 22. Wenstrom KD, Tessen JA, Zlatnik FJ, Sipes SL. Frequency, distribution, and theoretical mechanisms of hematologic and weight discordance in monochorionic twins. Obstet Gynecol 1992; 80:257-61. 23. Moore CM, McAdams AJ, Sutherland J. Intrauterine disseminated intravascular coagulation: a syndrome of multiple pregnancy with a dead twin fetus. J Pediatr 1969;74:523-8.
O
24. Blickstein I, Lancet M. The growth discordant twin. Obstet Gynecol Surv 1988;43:509-15. 25. Vintzileos AM, Rodis JF. Growth discordance in twins. In: Divon MY, editor. Abnormal fetal growth. New York: Elsevier Science; 1991. 26. Santema JG, Swaak AM, Wallenburg HCS. Expectant management of twin pregnancy with single fetal death. Br J Obstet Gynaecol 1995;102:26-30. 27. Genest DR, Williams MA, Greene MF. Estimating the time of death in stillborn fetuses, I: histologic evaluation of fetal organs; an autopsy study of 150 stillborns. Obstet Gynecol 1992;80:575-84.
N THE MOVE? Send us your new address at least six weeks ahead
Don’t miss a single issue of the journal! To ensure prompt service when you change your address, please photocopy and complete the form below. Please send your change of address notification at least six weeks before your move to ensure continued service. We regret we cannot guarantee replacement of issues missed due to late notification. JOURNAL TITLE: Fill in the title of the journal here.
OLD ADDRESS: Affix the address label from a recent issue of the journal here.
NEW ADDRESS:
Clearly print your new address here. Name Address City/State/ZIP
COPY AND MAIL THIS FORM TO: Mosby Subscription Customer Service 6277 Sea Harbor Dr Orlando, FL 32887
OR FAX TO: 407-363-9661
OR PHONE: 1-(800)654-2452 Outside the U.S., call (407)345-4000