Hernandez, Jennifer S. MD; Twickler, Diane M. MD; McIntire, Donald D. PhD; Dashe, Jodi S. MD
Hydramnios, an excessive amount of amniotic fluid, is identified in approximately 1% of singleton gestations and has been reported in as many as 7–14% of twin gestations.1–5 In twins, as in singletons, hydramnios may occur in the setting of maternal diabetes mellitus and may be associated with fetal anomalies and aneuploidy.6–8 Pregnancies with hydramnios generally are considered to be at increased risk for perinatal morbidity and mortality.9–11
Information is limited about the prevalence of hydramnios in dichorionic and monochorionic twins, the severity of hydramnios in twins, and how chorionicity and severity of hydramnios may affect pregnancy outcomes. Although hydramnios is a defining feature of twin-twin transfusion syndrome in monochorionic gestations, the adverse outcomes in such pregnancies are multifactorial and considered secondary to hemodynamic imbalance between the twins rather than amniotic fluid abnormalities.11 Our interest was the effect of hydramnios in the absence of complications such as twin–twin transfusion syndrome or twin-reversed arterial perfusion sequence. We reasoned that such information might be useful for counseling families and guiding pregnancy management. The objective of this study was to estimate the effect of varying degrees of hydramnios on pregnancy outcomes in both dichorionic and monochorionic twin gestations.
MATERIALS AND METHODS
This was a retrospective cohort study of all women with twin pregnancies who underwent prenatal ultrasonography between August 1997 and December 2010 and delivered liveborn neonates or stillborn fetuses at 24 weeks of gestation or more at a single institution. At our hospital, twin gestations are followed-up with serial ultrasound examinations every 4 to 6 weeks in the absence of complications. Chorionicity is assessed at the initial ultrasound examination, based on the presence or absence of a “twin peak” sign,12 number of placentas, thickness of the dividing membrane, and fetal sex. Hydramnios was diagnosed if the single deepest pocket of amniotic fluid was at least 8 cm at any point in gestation. Hydramnios was further categorized as mild if the deepest pocket of fluid was 8–9.9 cm, as moderate if the deepest pocket of fluid was 10–11.9 cm, and as severe if the deepest pocket of fluid was 12 cm or more; these are the definitions used in our sonography unit. The largest measurement, representing the greatest degree of hydramnios, was used for analysis. All examinations were performed at our sonography unit by registered diagnostic medical sonographers and over-read by faculty who specialize in obstetric sonography, in accordance with American Institute of Ultrasound in Medicine Practice Accreditation Standards. Whenever hydramnios was diagnosed, targeted sonography was performed by faculty who specialized in obstetric sonography. Our ultrasound unit was accredited by the American Institute of Ultrasound in Medicine throughout the study period.
Pregnancies were identified using a computerized ultrasound database. Maternal demographic characteristics and pregnancy outcome information were obtained using an obstetrical operations database that contains selected pregnancy and neonatal outcome information for all women who delivered at our hospital. Body mass index (calculated as weight (kg)/[height (m)]2) was assessed at the initial prenatal visit. Our definition of diabetes included pregnancies with either pregestational or gestational diabetes, regardless of insulin requirement. Birth weight percentiles were based on a published nomogram.13 At our hospital, evaluation for major congenital anomalies is routinely performed in the immediate neonatal period by attending pediatricians, and this information is recorded by trained perinatal research nurses. All structural abnormalities that are potentially life-threatening or those requiring surgery are included, as are chromosomal abnormalities. Information about stillborn fetuses (birth weight more than 500 g) is also recorded.
Twin pregnancies affected by hydramnios were compared with pregnancies in which the single deepest pocket of amniotic fluid was less than 8 cm. Monoamniotic pregnancies were excluded, as were pregnancies affected by twin–twin transfusion syndrome or twin-reversed arterial perfusion sequence. Twin–twin transfusion syndrome was defined as the presence of a single deepest pocket of at least 8 cm in one twin, along with a single deepest pocket of 2 cm or less in the other twin, at any point in gestation.14 The association between hydramnios and perinatal outcomes was evaluated separately in monochorionic and dichorionic gestations. Pregnancies with anomalous neonates and those resulting in a stillbirth of one or both twins were reviewed individually and analyzed separately. For consistency, a pregnancy was considered to have an outcome of interest if the outcome affected one or both twins. No power calculations were made, because this retrospective report includes our entire twin population from 1997 through 2010. Statistical tests included χ2, Mantel-Haenszel χ2 for trend,15 and analysis of variance. Statistical analyses were performed using SAS 9. P<.5 was considered significant. The study was approved by the Institutional Review Board of the University of Texas Southwestern Medical Center.
During the study period, 1,951 twin gestations underwent ultrasound evaluation and liveborn neonates or stillborn fetuses were delivered at our hospital. Of these, 1,311 (67%) were dichorionic and 640 (33%) were monochorionic. Hydramnios was identified in 348 pregnancies and was equally prevalent in dichorionic and monochorionic pregnancies, 234 (18%, 95% confidence interval 16–20%) and 114 (18%, 95% confidence interval 15–21%), respectively. Hydramnios was mild in 73% of affected cases, moderate in 20% of affected cases, and severe in only 7% of cases, such that the overall prevalence of severe hydramnios among twin gestations was 1% (26 of 1,951). Hydramnios was further categorized according to chorionicity, as shown in Figure 1.
Maternal demographic and pregnancy characteristics are presented in Table 1. In neither dichorionic nor monochorionic pregnancies was the presence or degree of hydramnios associated with maternal age, ethnicity, nulliparity, body mass index at initial prenatal visit, diagnosis of gestational or pregestational diabetes, or mode of delivery (all P>.05). Additionally, there was no significant difference in the mean gestational age at the time the greatest degree of hydramnios was identified, regardless of severity of hydramnios, in either dichorionic or monochorionic twins (both P>.05). Furthermore, the time from identification of the greatest degree of hydramnios to delivery also was not significantly different between categories of hydramnios or by chorionicity (both P>.05).
Pregnancies with anomalous neonates were analyzed separately. This group included 56 liveborn neonates and three stillborn anomalous fetuses, with one anomalous neonate per pregnancy in each case. The overall prevalence of neonates with major anomalies or aneuploidy in our series was 3%. As shown in Table 2, hydramnios was significantly associated with birth of an anomalous neonate in both dichorionic and monochorionic gestations. Further, the anomaly rate increased with the degree of hydramnios in both dichorionic and monochorionic gestations (both P for trend <.001). Pregnancies with severe hydramnios were at particular risk for an anomalous neonate (20% for dichorionic twins and 18% for monochorionic twins). The specific anomalies and aneuploidy are shown in Table 3 according to degree of hydramnios and chorionicity.
Pregnancies with nonanomalous stillbirths were analyzed separately and also are presented in Table 2. For dichorionic pregnancies, the prevalence of stillbirth was 0.9% (95% confidence interval 0.1–3.1%), with no stillborn fetuses among pregnancies with either moderate or severe hydramnios. For monochorionic pregnancies, the prevalence of stillbirth was 3%, and although relatively few pregnancies were complicated by severe hydramnios, stillbirth occurred in 3 of 11 pregnancies (27%, P<.001). On review of these three cases, none met criteria for twin–twin transfusion syndrome at the last ultrasound examination performed before stillbirth occurred; therefore, they were not excluded. However, in retrospect, it is not possible to definitively exclude twin–twin transfusion syndrome or twin-anemia-polycythemia sequence. In the first case, mild hydramnios was first identified at 27 weeks, with normal fluid surrounding the co-twin, and when stillbirth was noted at 31 weeks, the surviving twin had severe hydramnios and the stillborn fetus had anhydramnios. In the second case, severe hydramnios was identified at 25 weeks, which had resolved at follow-up imaging, and when stillbirth was noted at 31 weeks, the surviving twin had normal amniotic fluid and the stillborn fetus had anhydramnios. After delivery 1 month later, autopsy was performed but was noncontributory. In the third case, both twins were found to have severe hydramnios at 35 weeks, with 8% weight discordance, and 3 days later both were found to be stillborn. Autopsy described organ congestion, cardiomegaly, and ascites in the suspected recipient twin, and pale organs concerning for anemia were found in the suspected donor twin, suggesting twin-anemia-polycythemia sequence. However, both fetuses also had evidence of Klebsiella bacteremia.
Pregnancy and neonatal outcomes of nonanomaous liveborn neonates are shown in Table 4. For dichorionic twins, the mean birth weight increased significantly with degree of hydramnios (P<.001), such that dichorionic neonates from pregnancies with moderate or severe hydramnios were more than 200 g heavier than neonates from pregnancies without hydramnios. In addition, dichorionic twins with hydramnios were more likely to be large for gestational age (higher than the 90th percentile; P<.001). There was no association between birth weight and hydramnios in monochorionic twins. For both dichorionic and monochorionic twins, the presence of hydramnios did not affect the proportion of pregnancies, resulting in a preterm birth at or before 36 weeks, 34 weeks, or 32 weeks (Table 4). Similarly, hydramnios was not associated with birth weight discordance of 25% or more, small-for-gestational-age neonates (less than the 10th percentile), need for admission to the neonatal intensive care unit, or neonatal death in either dichorionic or monochorionic pregnancies (all P>.05).
There are four main findings from our series of nearly 2,000 twin pregnancies. First, hydramnios is common, identified in one in six twin gestations, independent of chorionicity, and most often diagnosed at its most severe in the late second or early third trimester. Second, the prevalence of anomalies increased significantly with greater degrees of hydramnios. Third, although stillbirths were fortunately infrequent, they were significantly more common in monochorionic pregnancies complicated by severe hydramnios. And, finally, pregnancies with hydramnios–in the absence of a fetal anomaly–were not more likely to be delivered preterm, were not complicated by small-for-gestational-age status or birth weight discordance, and did not include neonatal morbidity or mortality.
It is difficult to compare our prevalence of hydramnios with that reported in other series. As in singletons, a largest vertical pocket of amniotic fluid of 8 cm has been used to define hydramnios in monochorionic twins for the purpose of diagnosing twin–twin transfusion syndrome.14 However, there are no established criteria for degrees of hydramnios in twins. Our definitions are based on amniotic fluid index criteria that we and others have used to categorize hydramnios in singletons,8,16 along with the general principle that the single deepest pocket of amniotic fluid is approximately one-third the total amniotic fluid index.17 We did not adjust for gestational age. Magann et al9 performed serial sonography in nearly 300 diamniotic twin pairs between 17 and 37 weeks and found that the single deepest pocket of amniotic fluid did not vary with gestational age. Chau et al18 also reported that the single deepest pocket of amniotic fluid was stable across gestation in twin pregnancies. A potential limitation of our series is that pregnancies were considered to have hydramnios based on the largest measurement of amniotic fluid at any point in gestation. We reasoned that this type of analysis might be more useful for counseling. Not knowing if the hydramnios resolved later in the pregnancy, however, could certainly affect our overall prevalence rate.
The finding that selected pregnancy outcomes were more common in the setting of severe hydramnios would support use of our categorization. Specifically, severe hydramnios, defined as a single deepest pocket of fluid of at least 12 cm, was associated with a nearly 20% risk of major anomalies or aneuploidy, regardless of chorionicity. Our group and others have previously reported that the severity of hydramnios correlates with anomaly prevalence of singletons.8,16 We also found that monochorionic gestations with severe hydramnios were at increased risk for stillbirth. Although none of these cases met criteria for twin–twin transfusion syndrome before the occurrence of stillbirth, we realized on investigating them that it would be impossible to definitively exclude acute twin–twin transfusion syndrome or twin-anemia-polycythemia sequence as an etiology, even after autopsy was performed. This illustrates that monochorionic twins have a unique risk compared with dichorionic twins—a risk that should be acknowledged in the setting of severe hydramnios. We suggest that frequent antepartum surveillance of monochorionic gestations may be warranted in the setting of severe hydramnios.
An unexpected finding was that hydramnios was not associated with preterm birth, growth impairment, or adverse neonatal outcome. This was particularly true for dichorionic pregnancies, in which the rates of these outcomes were virtually identical in pregnancies with and without hydramnios, regardless of degree. This is in contrast to a study by Orhan et al,3 in which twins with hydramnios were at increased risk for preterm birth, small for gestational age status, and neonatal death. In that series, hydramnios was defined as an amniotic fluid index of more than 25 cm and occurred in only 1.5% of twins, and pregnancies with complications such as twin–twin transfusion syndrome were not excluded, such that their cohort may have been enriched with at-risk pregnancies.3 A limitation of our series is that even though we included nearly 2,000 sets of twins, monochorionic pregnancies with severe hydramnios were infrequent, such that with greater numbers we may have identified an increase in preterm birth with severe hydramnios.
There are several strengths of this study. First, at our institution, prenatal care is not modified by the diagnosis of hydramnios. Therefore, our finding that adverse pregnancy outcomes are not more common in the setting of hydramnios represents the true natural history of this finding, Second, because we estimated the effect of hydramnios separately in dichorionic and monochorionic pregnancies, providers may modify their counseling based on the specific type of twin gestation when faced with this common finding in practice. Further, we stratified the findings by severity of hydramnios. This is important in practice as well, because pregnancies complicated by mild hydramnios do not appear to be at the same risk as those complicated by severe hydramnios. Additional limitations of our study also should be mentioned. Relatively few pregnancies were complicated by severe hydramnios, and this is a population in which risks may be increased, particularly in monochorionic twins. In addition, because we evaluated outcomes according to the greatest degree of hydramnios present during gestation, thereby including each patient only once, we were not able to evaluate the course of hydramnios across gestation. It is conceivable that risks may be related to whether hydramnios progresses or resolves, and prospective study of the course of hydramnios in twins may clarify this issue.
In conclusion, this study helps us better-evaluate the antenatal and neonatal risks associated with the common finding of hydramnios in twin gestations. Our results suggest that when severe hydramnios is identified in either dichorionic or monochorionic twin gestations, specialized sonography is warranted because of the association between this finding and fetal abnormalities. Furthermore, fetal surveillance should be considered for monochorionic gestations with severe hydramnios after viability because of the increased risk for stillbirth in this cohort. In contrast, common adverse outcomes such as preterm birth and growth restriction did not appear to be not more frequent in the setting of hydramnios in either dichorionic or monochorionic twins.
1. Biggio JR Jr, Wenstrom KD, Dubard MB, Cliver SP. Hydramnios prediction of adverse perinatal outcome. Obstet Gynecol 1999;94:773–7.
2. Hill LM, Breckle R, Thomas ML, Fries JK. Polyhydramnios: Ultrasonically detected prevalence and neonatal outcome. Obstet Gynecol 1987;69:21–5.
3. Orhan A, Kurzel RB, Istwan NB, Rhea D, Burgess E, Stanziano G. The impact of hydramnios on pregnancy outcome in twin gestations. J Perinatol 2005;25:8–10.
4. Cunningham FG, Leveno KJ, Bloom SL, Hauth JC, Rouse DJ, Spong CY, editors. Williams obstetrics. 23rd ed. New York (NY): McGraw Hill; 2010. pp. 491–2.
5. Kramer WB, Van den Veyver IB, Kirshon B. Treatment of polyhydramnios with indomethacin. Clin Perinatol 1994;21:615–30.
6. Pregestational diabetes mellitus. Practice Bulletin No. 60. American Congress of Obstetricians and Gynecologists. Obstet Gynecol 2005;105:675–85.
7. Vink JY, Poggi SH, Ghidini A, Spong CY. Amniotic fluid index and birth weight: Is there a relationship in diabetics with poor glycemic control? Am J Obstet Gynecol 2006;195:848–50.
8. Dashe JS, McIntire DD, Ramus RM, Santos-Ramos R, Twickler DM. Hydramnios: Anomaly prevalence and sonographic detection. Obstet Gynecol 2002;100:134–9.
9. Magann EF, Doherty DA, Ennen CS, Chauhan SP, Shield D, Gjesdal SM, et al.. The ultrasound estimation of amniotic fluid volume in diamniotic twin pregnancies and prediction of peripartum outcomes. Am J Obstet Gynecol 2007:196:570.e1–8.
10. Varma TR, Bateman S, Patel RH, Chamberlain GVP, Pillai U. The relationship of increased amniotic fluid volume to perinatal outcome. Int J Gynecol Obstet 1988;27:327–33.
11. Habli M, Lim FY, Crombleholme T. Twin-to-twin transfusion syndrome: A comprehensive review. Clin Perinatol 2009;36:391–416, x.
12. Finberg HJ. The “twin peak” sign: reliable evidence of dichorionic twinning. J Ultrasound Med 1992;11:571–7.
13. Ananth CV, Vintzileos AM, Shen-Schwarz S, Smulian JC, Lai YL. Standards of birth weight in twin gestations stratified by placental chorionicity. Obstet Gynecol 1998;91:917–24.
14. Quintero RA, Morales WJ, Allen MH, Bornick PW, Johnson PK, Kruger M. Staging of twin-twin transfusion syndrome. J Perinatol 1999;19:550–5.
15. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959;22:719–48.
16. Lazebnik N, Many A. The severity of polyhydramnios, estimated fetal weight and preterm delivery are independent risk factors for the presence of congenital malformations. Gynecol Obstet Invest 1999;48:28–32.
17. Nyberg DA, McGahan JP, Pretorius DH, Pilu G. Abnormalities of amniotic fluid. In: Diagnostic imaging of fetal anomalies. Philadelphia (PA): Lippincott Williams and Wilkins; 2003.
18. Chau AC, Kjos SL, Kovacs BW. Ultrasonographic measurement of amniotic fluid volume in normal diamniotic twin pregnancies. Am J Obstet Gynecol 1996;174:1003–7.