Gestational hypertension, or pregnancy-induced hypertension, is a common (6 to 10%) pregnancy complication affecting millions of expectant mothers worldwide.1,2 It occurs 2–3 times more frequently in twin compared with singleton pregnancies.2–4 Because the elevated blood pressure in women bearing twin pregnancies may partly or merely manifest the physiologic response to the need for more blood (nutrients) supply to two fetuses, we hypothesize that gestational hypertension may have a more beneficial effect on neonatal outcomes (eg, preterm birth, fetal growth) in twin compared with singleton pregnancies. The hypothesis has important clinical implications in the management of increasing frequency of twin pregnancies complicated by gestational hypertension. For example, should obstetricians lower blood pressure to the same target level in twin compared with singleton hypertensive pregnancies? More importantly, should the decision of delivery be based on the same criteria in the management of twin compared with singleton hypertensive pregnancies? To date, there are no specific management recommendations of hypertensive disease in twin pregnancies; general management recommendations for singleton hypertensive pregnancies are commonly used.2 Nevertheless, such a practice may be questionable. If there is indeed a substantial difference in the effect of gestational hypertension on neonatal outcomes in twin pregnancies, a question may arise as to whether we need to develop twin- or multiple birth–specific management recommendations for hypertensive pregnancies. The present study was aimed to investigate the potentially differential effect of gestational hypertension on neonatal outcomes in twin compared with singleton pregnancies on top of the mere effects of being twins.
MATERIALS AND METHODS
We conducted a retrospective birth cohort-based study, using the U.S. National Center for Heath Statistics birth cohort-linked birth and infant death data sets for all live births in the years 1998–2000.5 The following exclusion criteria were applied to eliminate the potentially confounding effects of other maternal or infant medical conditions on neonatal outcomes and to avoid uncertainties in data quality (eg, recording or transcription errors in gestational age or birth weight or lack of information on important confounders such as maternal smoking or education) in such large population-based data sets: 1) triplets or higher plurality; 2) mothers with chronic hypertension before pregnancy; 3) mothers with diabetes (preexisting or gestational), heart disease, acute or chronic lung disease, genital herpes, anemia, renal disease, uterine bleeding, Rh sensitization, incompetent cervix, or any other reported maternal illness in pregnancy other than pregnancy-induced hypertension or eclampsia; 4) breech presentation at delivery; 5) births with congenital anomalies; 6) births with implausible birth weight for gestational age based on the algorithm of Alexander et al6; 7) births at very extreme gestational age: less than 20 weeks or more than 45 weeks; 8) births with missing or imputed values for any of the following variables: infant sex, plurality, gestational age, birth weight, maternal age, parity, martial status, education, smoking, alcohol use, or method of delivery; 9) births in states or counties of residence where information is unavailable in the linked birth and infant death data sets for any one of the following items: maternal education, smoking, alcohol use, maternal illness in pregnancy, method of delivery, 5-minute Apgar score, or the presence of congenital anomalies. Finally, a total of 5,626,785 nonmalformation live births (5,523,797 singletons and 102,988 twins) were eligible and included in the present study.
In the United States during the studied period (1998–2000), the diagnosis of pregnancy-induced or gestational hypertension followed the commonly accepted standards, ie, two or more occasions in blood diastolic pressure of 90 mm Hg or more or systolic pressure 140 mm Hg or more, taken 4 hours or more apart and occurring after 20 weeks of gestation without proteinuria.7 However, the U.S. linked birth and infant death data do not allow the distinction between gestational hypertension and preeclampsia (gestational hypertension plus proteinuria) and the code for pregnancy-associated hypertension in the data, including de novo hypertension with proteinuria (preeclampsia) or without proteinuria (hypertension only). Eclampsia may develop in the progression of severe gestational hypertension or preeclampsia. We examined the frequency of eclampsia as a surrogate indicator for the proportion of severe cases in hypertensive pregnancies, although it should be pointed out that eclampsia cases in the linked birth and infant death data might have included those convulsive conditions secondary to causes other than gestational hypertension or preeclampsia.
Crude rates and relative risks (RRs) were calculated to examine the crude effects of gestational hypertension on neonatal outcomes in singleton and twin pregnancies separately. The neonatal outcomes examined included preterm birth (less than 37 completed weeks of gestation), very preterm birth (less than 32 weeks), intrauterine growth restriction (IUGR, less than the third percentile in birth weight for gestational age, based on internal fetal growth standards for singleton nonmalformation live births to mothers without any reported maternal illnesses in pregnancy using PROC UNIVARIATE in Statistical Analysis System (SAS Institute Inc., Cary, NC) to obtain the sex- and gestational age–specific crude third percentile cutoff values—reference values available upon request—similar findings in IUGR risks were observed when alternative standards6 were used), low birth weight (LBW, less than 2,500 g), very low birth weight (less than 1,500 g), low 5-minute Apgar score (less than 4),8 and neonatal death (0–27 days of age). Multivariable logistic regression was used to obtain the adjusted odds ratios (aOR) for examining the risk patterns after controlling for maternal and pregnancy characteristics: maternal race (white, African American, others), age (younger than 20, 20–34, 35 or more years), education (less than high school, high school, some postsecondary, college or university), marital status (not married; married), parity (primiparous, multiparous), smoking (yes, no), alcohol use (yes, no), prenatal care inadequacy (yes, no), infant sex (boy, girl), and mode of delivery (cesarean, vaginal). Because twin pair-specific linked perinatal data were unavailable for years 1998–2000, we conducted an exploratory assessment of the robustness of the findings on the effect of gestational hypertension on neonatal outcomes for twins using the 1995–1997 multiple birth specific database,9 allowing intra-cluster (twin) correlations based on generalized estimated equations; similar reassuring effect sizes in twins were observed (results available upon request). Such identical effect size estimates for maternal risk factors with or without considering intraclass correlations in twins for large sample size data were also demonstrated in a recent study.10 All data management and analyses were carried out using Statistical Analysis System 8 (SAS Institute Inc.).
Large differences in maternal and pregnancy characteristics were observed between twin and singleton pregnancies (Table 1). Mothers bearing twin pregnancies were roughly 50% less likely to be primiparous or less than 20 years of age, 30% less likely to be an African American, have education less than high school or inadequate prenatal care, or be an alcohol drinker in pregnancy, 20% less likely to be a smoker or unmarried, but 50% more likely to be 35 years of age or older, 40% more likely to have a history of preterm or small for gestational age birth, and 30% more likely to have attended college or university education. Gestational hypertension occurred in 7.9% of twin pregnancies, 2.2 times more frequently than in 3.6% of singleton pregnancies. Eclampsia was 3.3 times more frequent in twin (8.8 per 1,000) compared with singleton (2.7 per 1,000) pregnancies. Gestational hypertension increased the risk of eclampsia (see MATERIALS AND METHODS, including convulsive conditions secondary to causes other than gestational hypertension) by nine-fold (from 2.1 per 1,000 to 18.8 per 1,000) in singleton pregnancies, compared with 4.4-fold (from 7.0 per 1,000 to 30.7 per 1,000) in twin pregnancies. Twins were 2.7 times more likely to be delivered by caesarean.
As expected, the absolute rates of all adverse neonatal outcomes (preterm birth, low birth weight, IUGR, low 5-minute Apgar score, and neonatal death) were much higher in twin compared with singleton pregnancies (Table 1). However, comparing gestational hypertensive with nonevent healthy pregnancies, the presence of gestational hypertension did not increase the risks of adverse neonatal outcomes in the same pattern for twins compared with singletons; a much more favorable effect was observed for twins (Table 2). The risk difference or attributable risk for gestational hypertension was always positive for any adverse neonatal outcomes among singleton pregnancies, but was negative for very preterm, very LBW, low 5-minute Apgar, or neonatal death among twin pregnancies. The RRs for singletons comparing gestational hypertensive with healthy nonevent pregnancies were 2.23 for preterm birth, 2.78 for very preterm birth, 2.49 for IUGR, 3.27 for LBW, 1.33 for low 5-minute Apgar score, and 1.07 for neonatal death, respectively. The corresponding RRs for twins were much lower or showed reversed associations: 1.21, 0.60, 1.12, 1.04, 0.32, and 0.21, respectively. The aORs showed a similar risk pattern in twin and singleton pregnancies after controlling for maternal race, age, marital status, parity, education, smoking, alcohol use, prenatal care use, and mode of delivery (Table 3 and Fig. 1).
Gestational hypertension did not significantly increase the risk of gestational age–specific risk of low 5-min Apgar score in twin pregnancies (Table 4). In contrast, the risk of low 5-minute Apgar score was increased by over 50% for singleton hypertensive pregnancies at term or postterm. Gestational hypertension did not significantly increase the gestational age–specific risk of neonatal death for both twin and singleton pregnancies and even showed a protective effect against neonatal death for preterm twin or singleton births. The lowest neonatal death rate for gestational hypertensive pregnancies was observed at term (37–41 weeks) for singletons, and at mild preterm (32–36 weeks) for twins.
The present study demonstrated a much more favorable effect of gestational hypertension on a number of important neonatal outcomes in twin compared with singleton pregnancies uncomplicated by other maternal illnesses in pregnancy. Not surprisingly, the absolute rates of any adverse neonatal outcomes were much higher in twin compared with singleton pregnancies regardless of the presence or absence of gestational hypertension. However, such elevated risks are merely attributable to the effects of being twins. Gestational hypertension itself confers a substantial increased risk of adverse neonatal outcomes in singleton pregnancies not observed in twin pregnancies.
Several studies have demonstrated elevated risks of gestational hypertension, preeclampsia, and adverse birth outcomes in twin compared with singleton pregnancies.3,4,11,12 However, it is important to know the effects of gestational hypertension on neonatal outcomes within twin compared with singleton pregnancies, separating the mere effects of being twins. We contend that only comparisons within twins and singletons themselves could make a clear case eliminating the effects solely attributable to being twins. The much more favorable effect of gestational hypertension on neonatal outcomes in twin compared with singleton pregnancies were clearly observed even after accounting for large differences in maternal and pregnancy characteristics. The observed reduced risk of neonatal death associated with hypertensive twin pregnancies is consistent with the reported lower perinatal death rates among such pregnancies based on populations in tertiary care centers with much higher incidence rates.4,12 In contrast to the reported decreased risk of preterm birth (less than 37 or 35 weeks) in gestational hypertensive compared with nonhypertensive twin pregnancies,4,12 we found that gestational hypertension was associated with an increased risk of preterm birth (less than 37 weeks) but a substantially (40%) decreased risk of very preterm birth (less than 32 weeks) in twin pregnancies.
Some evidence may support our findings of the less unfavorable effect of pregnancy-induced hypertension on neonatal outcomes in twins. Twin pregnancies are associated with a roughly 20% increase in maternal cardiac output than singleton pregnancies; the increase in stroke volume is primarily due to an increase in cardiac preload by a disproportionate increase in circulating blood volume (average increase of 1,570 mL in a singleton pregnancy compared with 1,960 mL in a twin pregnancy).13 Gestational hypertension was reported to be a risk factor for placenta abruption in singleton births (RR 2.3) but not in twin births in a study based on the U.S. linked perinatal data sets 1995–1996.14 Also, the recurrent risk of preeclampsia was 7.3% in twin pregnancies, only roughly one half that of 14.1% in singleton pregnancies.15 These findings suggest that gestational hypertension in twin pregnancies may represent a somewhat different entity in perhaps its placental pathophysiology and the role of constitutional propensity to the disease. We speculate that gestational hypertension may have more pathologic placental changes affecting blood supply to the fetus in singleton pregnancies, but fewer such changes in twin pregnancies in which the elevated blood pressure may partly or merely reflect the physiologic responses to more blood supply (nutrients) demand from two fetuses. Should placental changes be less likely “pathologic” in twin hypertensive pregnancies, the effect on neonatal outcomes would likely be different from singleton pregnancies. This is clearly shown by the observed 2.5-fold increased risk of IUGR associated with gestational hypertension in singleton pregnancies, but virtually an absence of any increased risk in twin pregnancies. Better oxygen and nutrient supply to two fetuses associated with elevated blood pressure may be more beneficial and consequently explain the seemly paradoxical substantial protective effects of gestational hypertension against very preterm birth, low 5-minute Apgar score, or neonatal death in twin pregnancies.
Our findings may have important clinical implications. The current practice of following the same protocol in the management of hypertensive disorders in pregnancy regardless of plurality may be questionable, considering the generally much more favorable effect of gestational hypertension on neonatal outcomes in twin pregnancies. For example, it is perhaps questionable to target the same lowered blood pressure levels in the management of hypertensive disorders in twin compared with singleton pregnancies or use the same criteria on the decision of delivery. It has already been shown that the greater the magnitude of blood pressure reduction in the treatment of hypertension in pregnancy, the greater the likelihood that IUGR will occur.16 Our findings further suggest that such adverse effects associated with antihypertensive treatment could be much worse for twin pregnancies. There seems to be a need for further interventional studies to assess whether we should consider more conservative twin- or multiple birth–specific management recommendations for hypertensive disorders in pregnancy, particularly for mild or moderate hypertension, for which the likelihood of any maternal benefits of antihypertensive treatment also seems very small.16,17
Underreporting or misclassification of the exposure or outcomes is a common potential problem in such large population-based perinatal databases. However, such misclassifications are likely nondifferential and would more likely lean toward null association. The U.S. linked perinatal data could not distinguish gestational hypertension from preeclampsia. Nevertheless, it is known that the proportion of preeclampsia (to gestational hypertension) is higher and early onset of hypertension more frequent in twin pregnancies,2,3 which was partially confirmed by the observed even greater risk of eclampsia in twin compared with singleton pregnancies, suggesting that the more beneficial effect on neonatal outcomes could not be explained by a less clinically pronounced disease in twin pregnancies. We did not have information on birth order, twin discordance, chorionicity, or type of service providers, which may affect neonatal outcomes within twin pairs. Differential effect of gestational hypertension on birth weight discordance may be a potential causal pathway that partly explains the differential effect on neonatal outcomes in twin compared with singleton pregnancies. Our findings are relevant to the overall effect of gestational hypertension on neonatal outcomes for both twins without knowing the effect on a particular twin. We speculate that type of service providers might be generally better for twins compared with singletons because women bearing twin or high-order pregnancies are considered high-risk pregnancies and more likely referred to and delivered in tertiary care centers and therefore might partly account for the less unfavorable effect of gestational hypertension on the risks of adverse neonatal outcomes in twin pregnancies. Also, findings from the present study reflect the outcomes under the current practice of uniform management protocols for hypertensive disease in pregnancy. We have no information on how gestational hypertensive patients were managed or the intrapartum management of the fetus in numerous hospitals. However, a claim cannot be made as to the sufficiency or benefits of the current practice in the management of hypertensive disease in twin pregnancies, because it is unknown how neonatal outcomes might be if more conservative management recommendations were applied in hypertensive pregnancies with twin or multiple fetuses. It should be further cautioned that we did not have data on the profile of severe maternal morbidities or deaths that were linked to gestational hypertension and therefore could not adequately address the effect of gestational hypertension on severe adverse maternal outcomes in twin compared with singleton pregnancies, although our results indicated that gestational hypertension might increase the risk of “eclampsia” to also a lesser extent in twin (4.4-fold) compared with singleton (nine-fold) pregnancies.
1. Solomon CG, Seely EW. Hypertension in pregnancy. Endocrinol Metab Clin North Am 2006;35:157–71.
2. Krotz S, Fajardo J, Ghandi S, Patel A, Keith LG. Hypertensive disease in twin pregnancies: a review. Twin Res 2002;5:8–14.
3. Santema JG, Koppelaar I, Wallenburg HC. Hypertensive disorders in twin pregnancy. Eur J Obstet Gynecol Reprod Biol 1995;58:9–13.
4. Sibai BM, Hauth J, Caritis S, Lindheimer MD, MacPherson C, Klebanoff M, et al. Hypertensive disorders in twin versus singleton gestations. National Institute of Child Health and Human Development Network of Maternal-Fetal Medicine Units. Am J Obstet Gynecol 2000;182:938–42.
5. Centers for Disease Control and Prevention (CDC), National Center for Heath Statistics (NCHS). Birth cohort linked birth/infant death data sets (1998, 1999, 2000). NCHS CD-ROM series 20.
6. Alexander GR, Himes JH, Kaufman RB, Mor J, Kogan M. A United States national reference for fetal growth. Obstet Gynecol 1996;87:163–8.
7. National High Blood Pressure Education Program Working Group report on high blood pressure in pregnancy. Am J Obstet Gynecol 1990;163:1691–712.
8. Casey BM, McIntire DD, Leveno KJ. The continuing value of the Apgar score for the assessment of newborn infants. N Engl J Med 2001;344:467–71.
9. Martin JA, Curtin SA, Saulnier ML, Mousavi J. The matched multiple birth file, 1995-1997. Hyattsville (MD): National Center for Health Statistics; 2001.
10. Ananth CV, Platt RW, Savitz DA. Regression models for clustered binary responses: implications of ignoring the intracluster correlation in an analysis of perinatal mortality in twin gestations. Ann Epidemiol 2005;15:293–301.
11. Ros HS, Cnattingius S, Lipworth L. Comparison of risk factors for preeclampsia and gestational hypertension in a population-based cohort study. Am J Epidemiol 1998;147:1062–70.
12. Campbell DM, MacGillivray I. Preeclampsia in twin pregnancies: incidence and outcome. Hypertens Pregnancy 1999;18:197–207.
13. Norwitz ER, Edusa V, Park JS. Maternal physiology and complications of multiple pregnancy. Semin Perinatol 2005;29:338–48.
14. Ananth CV, Smulian JC, Demissie K, Vintzileos AM, Knuppel RA. Placental abruption among singleton and twin births in the United States: risk factor profiles. Am J Epidemiol 2001;153:771–8.
15. Trogstad L, Skrondal A, Stoltenberg C, Magnus P, Nesheim BI, Eskild A. Recurrence risk of preeclampsia in twin and singleton pregnancies. Am J Med Genet A 2004;126:41–5.
16. von Dadelszen P, Ornstein MP, Bull SB, Logan AG, Koren G, Magee LA. Fall in mean arterial pressure and fetal growth restriction in pregnancy hypertension: a meta-analysis. Lancet 2000;355:87–92.
17. Sibai BH. Treatment of hypertension in pregnant women. N Engl J Med 1996;335:257–65.