Risk factors for adverse outcomes in twin pregnancies: a narrative review

Abstract

Purpose

Globally, the incidence of twin pregnancies is rising owing to the use of assisted reproductive technologies (ART), emigration and deferment of pregnancy until advanced maternal age (AMA). While twin pregnancies have higher absolute risks of adverse outcomes, including miscarriage, stillbirth, neonatal death and preterm delivery, the impact of specific exposures and risk factors related to these outcomes may differ between twin pregnancies and singleton pregnancies. Regarding modifiable factors, data are sometimes based on evidence extrapolated from singleton or whole obstetric populations. Therefore, targeted evidence is required to provide care tailored to twin pregnancies to prevent adverse outcomes. We aimed to comprehensively review the association between different risk factors and adverse outcomes in twin pregnancies, including data on chorionicity, and to compare these to singletons.

Materials and methods

This review examines the risks associated with chorionicity, AMA, body mass index (BMI), socioeconomic and ethnic inequalities, maternal smoking, use of ART, maternal perception of fetal movement, and maternal comorbidities, including hypertensive disorders of pregnancy (HDP) and gestational diabetes mellitus (GDM). Adverse outcomes reported were preterm birth, admission to the neonatal intensive care unit (NICU), stillbirth and neonatal mortality. As such, fetal mortality and morbidity will be under-represented, as pregnancy loss before 22–24 weeks is omitted.

Results

Monochorionicity increases the risk of stillbirth, NICU admission, and preterm delivery in twin pregnancy. AMA predisposes twin pregnancies to higher risks of mortality, admission to the NICU, and preterm birth than singleton pregnancies do. Conversely, the impact of BMI, socioeconomic inequalities, smoking, ART, and HDP on adverse outcomes appears to be lower in twin pregnancies than in singleton pregnancies. This attenuation might be explained by the higher baseline risk of adverse outcomes such as preterm birth in twin pregnancies. Some exposures, such as ART use and GDM, appear to be “protective” against perinatal mortality in twin pregnancies, despite being established risk factors for adverse outcomes in singleton pregnancies, potentially related to access to specialist care. There is a paucity of evidence available to counsel mothers of twin pregnancies regarding reduced fetal movement.

Conclusions

Overall, the risk factors for adverse pregnancy outcomes differ between twin and singleton pregnancies. This highlights the need for further studies to examine the association between risk factors and adverse outcomes in twin pregnancies. The resulting data would facilitate tailored guidance for twin pregnancies, contribute to improved antenatal care, and inform wider public health strategies.

Keywords:

• Twin pregnancy
• multiple pregnancy
• preterm birth
• stillbirth
• perinatal morbidity
• risk factors

Introduction

Multiple births account for up to 3.2% of births in high-resource settings, although there is some geographical variation [1]. Globally, the incidence of twin pregnancies has increased over the last two decades, which is related to increased maternal age at conception, wider use of assisted reproductive technologies (ART), and changes in population diversity [2]. Twin pregnancies have a higher risk of miscarriage, stillbirth [3], neonatal death, and preterm delivery [4] than singleton pregnancies. Twin chorionicity and amnionicity are a major influence on outcomes [5]. Dichorionic (DC) twins have two separate placentas with fetuses in two separate amniotic sacs. They occur either because of fertilization of two ova by two separate spermatozoa (dizygotic) or the early division of an ovum fertilized by a single sperm (monozygotic). Monochorionic (MC) twins are always monozygotic and share a single placenta; 99% of them are diamniotic (MCDA). Monochorionic twins have a significantly increased risk of adverse outcomes compared to dichorionic twins, principally associated with the conjoining of fetal circulations within a single placenta, or rarely because of monoamniocity leading to congenital malformations and cord entanglement [6]. Compared to singleton pregnancies, monochorionic diamniotic (MCDA) twins have a 13 times higher stillbirth rate than singleton pregnancies, which decreases with advancing gestational age [7].

Although preliminary evidence suggests that causes of adverse outcomes may differ between twin pregnancies compared to singleton pregnancies [8], guidelines for optimizing outcomes in twin pregnancies are often based on evidence extrapolated from singletons [9]. To prevent adverse outcomes, targeted evidence is needed to provide tailored care exclusively for twin pregnancies [10]. Here, we aimed to comprehensively review the association between different exposures and adverse outcomes in twins and compare this to singleton pregnancies to better identify modifiable risk factors for adverse outcomes in twin pregnancies.

Fetal and maternal characteristics

Direct effects of chorionicity and amnionicity

Monochorionicity carries a higher risk of stillbirth (aOR 1.81, 95%CI 1.13–2.82), admission to the neonatal intensive care unit (NICU) (aOR 1.41, 95%CI 1.12–1.78), and preterm delivery <34 weeks (aOR 1.47, 95%CI 1.17–1.86) when compared to DC twins, but these differences diminish after accounting for TTTS, fetal anomalies (discordant and concordant) and monoamniocity [11]. Twin-to-twin transfusion syndrome (TTTS) arises due to intertwin vascular anastomoses in MC twins; it occurs in 10–15% of MC twin pregnancies [12] and more commonly in MCDA than in MCMA twins [13]. Without intervention, there is a high risk of stillbirth (25.2%, 95%CI 13.4–39.3%), neonatal death (12.2%, 95%CI 4.8–22.4%), and preterm birth (50.5%, 95%CI 19.9–80.9%) [14]. It is possible that MC mortality is underestimated further due to the omission of early pregnancy losses (<22 weeks) from these figures, representing a “hidden mortality” associated with the most complicated MC twin pregnancies [5]. Other MC-specific complications include twin reversed arterial perfusion sequence, twin anemia polycythemia sequence (TAPS) and selective fetal growth restriction (sGR). There was no difference in the risk of neonatal death (aOR 0.43, 95%CI 0.13–1.47) preeclampsia, gestational diabetes (GDM), placental abruption, or placenta praevia [11]. With respect to birth weight, there was no significant difference in birth weight discordance between MCDA (10.4%) and DCDA (11.3%) twins [15].

Chorionicity has implications for optimal timing of delivery to reduce the risk of adverse outcomes. In DC twin pregnancies, the rate of stillbirth increases from 1.2 (95% CI 0.7–1.8) to 10.6 per 1,000 births (95%CI 7.1–15.3) between 34 and 38 weeks. Conversely, the rate of neonatal death falls over the same period from 6.7 (95%CI 3.3–13.5) to 1.5 per 1,000 births (95%CI 0.7–3.3) [16]. This pattern was replicated in MC twins. In DC twin pregnancies, the risk of perinatal death outweighs the risk of stillbirth beyond 37 weeks, whereas this balance of risks occurs earlier, around 36 weeks, for MC twin pregnancies [16]. Consequently, current national clinical guidelines in the UK recommend birth at 37 weeks for DC twins and 36 weeks of gestation for MC twins, the effect of which may contribute to the attenuation of associations between exposures and adverse outcomes.

Maternal age

Women of advanced maternal age (AMA) are more likely to become pregnant with twins, partly related to the more frequent use of ART [17]. A meta-analysis of twin pregnancies found that women with AMA (≥40 years) had an increased risk of preterm birth (Relative Risk (RR) 2.27, 95%CI 1.09–5.80), admission to the NICU (RR 1.78, 95%CI 1.21–2.64) and perinatal or neonatal mortality (RR 5.76, 95%CI 1.11–29.70) [18]. In comparison, a meta-analysis examining singleton pregnancies showed that AMA was associated with a relatively smaller increase in the risk of preterm birth (OR 1.20, 95%CI 1.07–1.33), NICU admission (OR 1.38, 95%CI 1.31–1.47), and neonatal death (1.41, 95%CI 1.14–1.75) [19]. AMA was associated with an increased risk of stillbirth by 1.75 (95%CI 1.62–1.89) times [19]. Wang et al. find that among women of AMA, the absolute risk of preterm birth is higher in twins conceived with IVF compared to twins conceived spontaneously and singletons born to women of AMA [20]. Singletons conceived with IVF also had a higher absolute risk of preterm birth compared to singletons conceived spontaneously. Conversely, absolute risk of stillbirth was higher in twins conceived with spontaneously to women of AMA. While the absolute risk of stillbirth was lower among singletons, it was again higher among spontaneously conceived singletons compared to IVF conceived. This difference could be due to higher antenatal surveillance of IVF pregnancies, resulting in fewer stillbirth but higher rates of iatrogenic preterm deliveries.

Age extremes increase the risk of preterm birth in twins [21]. One study showed a stepwise decrease in risk up to the age of 40 years, after which 58.1% of twins were born preterm. Conversely, the risk of stillbirth and neonatal death was lower in women in the older age groups than in the rest of the twin population. Other studies have suggested that AMA does not confer an increased risk of adverse perinatal outcomes in twin pregnancies [22].

Body mass index (BMI)

The association between BMI and adverse perinatal outcomes appears weaker in twin pregnancies than in singletons [23–25]. This might relate to more frequent antenatal scanning in twin pregnancies, leading to earlier identification of problems with fetal growth. Alternatively, the attenuation of this established risk factor on outcomes in twin pregnancy might be the effect of the inherently higher risks associated with twin pregnancies and chorionicity masking the effects of individual risk factors.

Evidence is conflicting regarding whether raised BMI increases the risk of preterm birth in twin pregnancies; some studies found no significant association with very preterm birth (<32 weeks) [23] or preterm birth (<37 weeks) [24,25], while others found an increased risk of total (aOR 1.43, 95%CI 1.01–2.03), spontaneous (aOR 1.58, 95%CI 1.05–2.36) [26], and very preterm birth (OR 1.65, 95%CI 1.10–2.48) [25,27]. In singleton pregnancies, increased maternal BMI increases the risk of preterm birth (OR 1.54, 95%CI 1.06–2.24) and very preterm birth (OR 2.41, 95%CI 1.21–4.77) [25], often in a dose-response pattern [24,28]. This is mainly driven by the increased risk of provider-initiated delivery [23]. One study in singletons found that underweight women had the highest risk of preterm birth (aOR 1.45 95%CI 1.26–1.67) [24,29]. Underweight BMI also increases the risk of preterm birth in twin pregnancy (aRR 1.67, 95%CI 1.17–2.37), mediated by an increased risk of spontaneous preterm birth [23].

In singleton pregnancies, maternal obesity was associated with a stepwise increase in stillbirth risk from BMI 25-29.9 (aOR 1.51, 95%CI 1.40–1.62) to ≥35 (aOR 2.85, 95%CI 2.52–3.22) [28,30]. In twin pregnancies, there was a 1.31 (95%CI 1.02–1.68) times higher risk of stillbirth amongst obese mothers, but this association was not related to the degree of obesity [30]. Similarly, there was a stepwise increase in the risk of NICU admission among singleton pregnancies from maternal BMI 25-29.9 (aRR 1.16, 95%CI 1.14–1.18) to ≥40 (aRR 1.58, 95%CI 1.51–1.64) [23], but NICU admission was not more common among twin pregnancies with increased maternal BMI [23,27].

Socioeconomic and ethnic inequalities

Socioeconomic inequalities are structural risk factors for adverse maternal and perinatal outcomes [31]. Studies examining ethnic disparities in the risk of adverse perinatal outcomes suggest an attenuated risk in twin pregnancies compared to singleton pregnancies. Black women had a 1.3-times higher risk of preterm birth than White women in twin pregnancies, compared to a 2.3-times higher risk in singleton pregnancies [32]. This might again relate to more frequent antenatal contact or the masking effect of inherently higher risk twin pregnancies on individual risk factors. Soffer et al. found that African-American, Asian, and Hispanic women did not have a higher risk of very preterm twin birth than White women [33]. Another study found no relationship between ethnicity and the rate of unplanned twin Cesarean births or any other adverse perinatal outcomes [34].

In the UK, socioeconomic inequalities are responsible for a large proportion of stillbirths (24%), preterm births (19%), and pregnancies affected by FGR (31%) [35]. Furthermore, socioeconomic disparities in birth outcomes exist across all ethnic groups [36]. Twins were excluded from these studies. Challenges in addressing structural factors have created a ceiling for progress in reducing stillbirths among singletons [31], with social gradients persisting despite declining stillbirth rates [37]. In London, Black (aOR 2.15, 95%CI 1.56–2.97) and South Asian (aOR 2.33, 95%CI 1.42–3.48) women had twice the risk of stillbirth compared to White women [38].

Women in the most deprived quintile had a 3.16 times (95%CI 1.47–6.77) higher risk of late stillbirth than those in the least deprived quintile [39]. Although increased risk is still present, the risk of stillbirth associated with deprivation is lesser for twins than for singletons, with a 1.54 (95%CI 1.28–1.85) and 1.94 (95%CI 1.84–2.05) times higher rate of stillbirth between the most and least deprived decile in twins and singletons respectively [40].

Behavioral and social factors

Smoking

Cigarette smoking is a well-established risk factor for low birth weight, preterm birth, NICU admission and stillbirth [41]. There was no difference in the prevalence of smoking among women with singleton and twin pregnancies [42].

The relationship between maternal smoking and adverse perinatal outcomes in twin pregnancies is unclear. Pollack et al. demonstrated that a higher baseline prevalence of adverse outcomes in twins partially masks the risks associated with smoking [42]. The RRs for low birth weight in singletons and twins were 2.21 (95%CI 2.20–3.33) and 1.33 (95%CI 1.30–1.37) respectively, and for infant mortality were 1.73 (95%CI 1.46–2.13) and 1.25 (95%CI 0.97–1.60) respectively. The population-attributable risk for smoking and the incidence of low birth weight in singleton births was 11.1%, compared to 3.5% among twins. However, incremental risk calculations showed that the impact of maternal smoking is greater among twins for certain outcomes, and every 123 per 1,000 cases of low birth weight among twins was attributable to maternal smoking, compared to 55 per 1,000 cases among singletons. Maternal smoking contributes to an increased risk of low birth weight in twins independent of the effect of gestation [43–45]. Cooperstock et al. found that cigarette smoking increased the risk of fetal growth discordance by 1.23-times, and that it was an independent risk factor for preterm birth and small for gestational age babies [46].

A population-based study suggested that maternal smoking had little effect on the length of gestation among twins compared to the significant reduction seen in singleton pregnancies [43]. This is corroborated by several studies on twin pregnancies, which did not find an increased risk of preterm birth in smokers [44,47]. One study showed a dose-response relationship between the number of cigarettes smoked and shortened length of gestation but failed to demonstrate statistical significance for an association between smoking and preterm birth [48].

Use of ART

The introduction of ART has contributed to an increasing incidence of twin pregnancies [49]. ART encompasses ovulation induction, intrauterine insemination, and IVF. Systematic reviews have shown that ART increases the risk of preterm birth and low birth weight compared with spontaneously conceived singletons [50]. Mothers who use ART tend to be of higher maternal age and have higher rates of obstetric complications such as GDM and pregnancy-induced hypertension [51]. Furthermore, they are an inherently high-risk group owing to preexisting risk factors associated with infertility [52]. On the other hand, women who use ART are more likely to have protective exposures, such as a higher socioeconomic status [51] and earlier prenatal diagnosis of twin pregnancy.

A systematic review [53] found that twins conceived with ART had an increased risk of preterm birth (RR 1.07, 95%CI 1.02–1.13) and admission to the NICU (RR 1.05, 95%CI 1.01–1.09) compared to spontaneously conceived twins. However, the risk of perinatal mortality was lower for ART conceived twins (RR 0.58, 95%CI 0.44–0.77) than for those conceived spontaneously. In comparison, singletons conceived with ART had an increased risk of preterm birth (RR 2.04, 95%CI 1.80–2.32), NICU admission (RR 1.27, 95%CI 1.16–1.40), and perinatal mortality (RR 1.68, 95%CI 1.11–2.55) compared to spontaneously conceived singletons. These results were synthesized from studies which matched for important confounders, which almost always included maternal age.

A population-based study found that ART twins had a lower stillbirth rate than non-ART twins (1.4% vs. 2.3%, p < 0.01) [51], in contrast to studies that found no significant difference [54] or a higher perinatal mortality rate in ART twins [55,56]. When a composite outcome was formed, ART twins had a 1.28 (95%CI 1.09–1.50) times higher risk of adverse neonatal outcome compared to non-ART twins, which might be explained by higher baseline risks in women with subfertility [57].

However, other studies found no difference in adverse outcomes between ART and non-ART twins, suggesting that ART does not confer additional risks in twin pregnancies [58–60]. Fitzsimmons et al. found a significantly higher perinatal mortality rate in spontaneously conceived twins than in those conceived with ART [60]. This could be explained by the higher number of MC twins in the spontaneously conceived cohort. One study found that preterm birth (<32 weeks) was more common in non-ART twins [58]. Although ART is associated with significant risks in singleton pregnancies, there are potentially protective effects associated with ART for some outcomes in twin pregnancies. This could be due to the higher absolute risk of adverse outcomes in twin pregnancies, or the influence of opposing protective factors in women who use ART.

Maternal perception of fetal movements

Reduced fetal movements (RFM) are associated with FGR and stillbirth in singleton pregnancies [61]. RFM was reported as the most common presentation (57.4%) prior to stillbirth [62] and was associated with a 5.5-fold (95%CI 2.81–10.85) increase in the risk of stillbirth [63]. Extrapolation from studies of singleton pregnancies is potentially problematic since studies suggest that twins exhibit different patterns of fetal movement [64,65]. Nevertheless, the MBRRACE Confidential Inquiry into Twins highlighted RFM as a “red flag” symptom for recognition of a compromised fetus or TTTS [8].

To our knowledge, Levy et al. are the only authors to publish data on the association between RFM and adverse perinatal outcomes in twin pregnancies [66]. The rate of adverse composite neonatal outcomes in the RFM group was 29.5% compared to 19.2% in the non-RFM group (p = 0.01). Two cases of fetal demise and three cases of neonatal death were recorded, all of which occurred in the RFM group. However, there are no data available specifically for association between double or single fetal demise in twin pregnancy with RFM. NICU admissions were significantly higher in twins with RFM than in those without RFM (32.5% vs. 19.2%, p = 0.001). Comparatively, studies in singletons found no association with NICU admission [67], or found that neonates with RFM were less likely to be admitted to the NICU (RR 0.57, 95%CI 0.46–0.70). This could be due to the higher rate of preterm births in singletons without RFM (5.2% vs. 3.8%, p < 0.01) [63].

Medical disorders of pregnancy

Gestational diabetes

Twin pregnancy is a risk factor for the development of GDM [68], likely due to higher levels of diabetogenic hormones in larger placental masses. The risk of GDM between MC and DC pregnancies is the same [11]. Hiersch et al. found that GDM is associated with fewer adverse outcomes in twin pregnancies than in singleton pregnancies [69]. GDM was associated with a higher risk of preterm birth in twins (adjusted RR (aRR) 1.21, 95%CI 1.08–1.37) and singletons (aRR 1.48, 95%CI 1.39–1.57), and a higher risk of large for gestational age twins (aRR 2.53, 95%CI 1.52–4.23) and singletons (aRR 1.18, 95%CI 1.13–1.23). GDM was associated with NICU admission in singleton pregnancies (aRR 1.44, 95%CI 1.38–1.50), but not in twin pregnancies (aRR 1.12, 95%CI 1.00–1.25) [69]. However, a meta-analysis found that GDM in twin pregnancy carried a 1.49 (95%CI 1.10–2.02) times higher risk of NICU admission [70]. Although diabetes in pregnancy has been associated with an increased risk of stillbirth, GDM appeared to decrease the risk of stillbirth in twins (OR 0.39, 95%CI 0.16–0.98) and singletons (OR 0.63, 95%CI 0.49–0.80) [71], potentially related to access to specialist care and additional antenatal screening due to recognition of high risk status.

There is conflicting evidence regarding the effect of GDM on other adverse outcomes in twin pregnancies. Some studies suggest that GDM is associated with a greater risk of macrosomia, prematurity [72], NICU admission [73] and respiratory distress [74] in twins, while others find no difference or attenuation of these risks when comparing twin pregnancies affected and unaffected by GDM [75–78]. The lack of difference in the risk of macrosomia, an established adverse outcome associated with GDM in singleton pregnancies, might result from the baseline lower birth weights observed in twins.

Hypertensive disorders of pregnancy

Hypertensive disorders of pregnancy (HDP) include preeclampsia and gestational hypertension. Uncomplicated twin pregnancies have a higher cardiovascular demand than singleton pregnancies and demonstrate equivalent maternal cardiac remodeling to singleton pregnancies with HDP [79]. Furthermore, twin pregnancy is a risk factor for developing HDP [80,81], with one study finding an attributable risk of 8.4% for the development of preeclampsia [82]. There is no difference in risk of preeclampsia between MC and DC pregnancies [11].

A population study found that twin pregnancies had a 4.07 (95%CI 3.65–4.54) times higher risk of preeclampsia than singleton pregnancies and a prevalence of HDP of 11.8% in twin pregnancies compared to 3.4% in singletons [82]. Furthermore, preeclampsia develops earlier in twin pregnancies [83], and is more likely to be severe at the time of diagnosis [84]. However, the frequency of eclampsia appears to be similar between twins (0.6%) and singletons (0.7%) [83]. There is conflicting evidence regarding the risk of gestational hypertension, the prevalence of which did not differ significantly in a population study (2.2% vs. 1.7% among twin and singleton pregnancies, respectively) [82], whereas earlier evidence suggests a higher risk of gestational hypertension in twin pregnancies [85].

Aviram et al. found an increased risk of preterm birth in twins affected by HDP compared to normotensive twins (aRR 1.30, 95%CI 1.24–1.35) [81], which was smaller than that reported by Hayes-Ryan et al. (aOR 2.5, 95%CI 1.6–3.7) [86]. There was also an increased risk of admission to the NICU (aRR 1.21, 95%CI 1.18–1.25) [81], which was corroborated by Hehir et al. [87]. Sibai et al. found an increased risk of perinatal mortality in normotensive twins compared to twins with HDP but lost statistical significance after adjusting for gestational age [80]. Comparatively, the increased risk for singletons with HDP compared to normotensive singletons of preterm birth was aRR 2.42 (95%CI 2.36–2.49) and of admission to NICU was aRR 1.81 (95%CI 1.78–1.85) [81]. There was no increased risk of neonatal death for singletons with HDP (aRR 1.14, 95%CI 0.87–1.51).

Conclusion

This review highlights that the risk factors for adverse pregnancy outcomes differ between twin pregnancies and singleton pregnancies (summarized in Figure 1). A recurrent theme was the attenuation of the impact of risk factors established in singleton twin pregnancies. This may result from the inherently higher risk of twin pregnancy and the effects of chorionicity, masking the effects of individual factors. Despite dichorionic and monochorionic twin pregnancies carrying specific risks [88], few epidemiological studies have stratified the findings based on chorionicity. Our findings highlight the need for further studies in twin pregnancies to identify modifiable risk factors for adverse outcomes in twin pregnancies. Such studies must be adjusted for chorionicity to achieve meaningful outcomes. The resulting data would facilitate more tailored guidance for twin pregnancies, contribute to improved antenatal care, and inform wider public health strategies.

Figure 1. Summary table of effect sizes of adverse outcomes for twins vs singletons.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data sharing is not applicable to this article as no new data were created or analyzed in this study.

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.