Three randomised trials have also i

Three randomised trials have also included twin gestations [61–63]. Rust et al. [63] included 14 twin
pregnancies among the 113 patients in a study from the USA, with seven allocated to the study group
having a McDonald cerclage and seven in the control group having no cerclage. Multiple gestation
correlated with decreased gestational age at delivery, but not with increased neonatal morbidity [63].
In a Dutch study [61], 35 women were also randomised to a McDonald cerclage or no surgical treatment.
In this smaller trial, cerclage reduced the rate of preterm delivery before 34 weeks as well as
compound perinatal morbidity. Unfortunately, no details are given on multiple gestations, but they
constituted a limited part of the study population [59,61]. In the last trial, 61 women were randomly
assigned to McDonald cerclage or bed rest [62]. This trial included four twin gestations, three assigned
to the cerclage group and one assigned to the control group. All three twin gestations in the cerclage
group had early preterm births at 20, 21 and 22 weeks of gestation, respectively, whereas the twin
gestation randomised to the control group delivered at 34 weeks’ gestation [62]. Thus, although the
data from randomised trials are sparse, data from these three trials have subsequently been reviewed
in a meta-analysis, including a total of 49 twin pregnancies [64]. The rate of preterm delivery before 35
weeks was 75.0% (18 out of 24) in the cerclage group compared with 36.0% (9 out of 25) in the control
group, giving a relative risk of 2.2 (95% CI 1.2 to 4.0), whereas the perinatal mortality rate was 22.9% in
the cerclage group compared with 6.0% in the control group, resulting in a non-significant relative risk
of 2.7 (95% CI 0.8 to 8.5).
Although the effect of cerclage in twin pregnancies has not been properly assessed in adequately
sized randomised trials, it must be concluded that cervical cerclage in twin pregnancies actually seems
to increase the rate of preterm delivery. It cannot be excluded that twin gestations with shortened
cervical length may benefit from insertion of an emergency cerclage, as suggested by a small retrospective
analysis of 14 cases [65].
Several pathophysiologic processes may lead to cervical shortening, and it is unclear why insertion
of cerclage seems to improve perinatal outcome in high-risk singleton pregnancies, whereas the
opposite is true in twin pregnancies. Increased uterine activity, subclinical infection, or placental
abruption, may be more common in twin pregnancy, and may be exacerbated by the presence of a
foreign body in the cervical stroma [63].