DOI: 10.1111/1471-0528.13188 www.bjog.org
Vaginal progesterone in women with twin gestations complicated by short cervix: a retrospective cohort study SG Brubaker,a C Pessel,a N Zork,a C Gyamfi-Bannerman,a CV Ananthb,c a
Division of Maternal–Fetal Medicine, b Department of Obstetrics and Gynecology, c Department of Epidemiology, Columbia University, New York City, NY, USA Correspondence: Dr SG Brubaker, Division of Maternal–Fetal Medicine, Department of Obstetrics and Gynecology, Columbia University College of Physicians and Surgeons, 622 West 168th Street, New York City, NY 10032, USA. Email [email protected] Accepted 8 October 2014. Published Online 27 November 2014.
Objective To determine whether the use of vaginal progesterone
Setting Tertiary-care medical centre in New York City.
delivery prior to 35 weeks of gestation in the vaginal progesterone group, compared with the no vaginal progesterone group, was 1.8 (95% confidence interval, 95% CI 1.5–3.1) in the unadjusted analysis, 1.4 (95% CI 0.7–3.2) following multivariable adjustment for confounding factors, and 1.5 (95% CI 1.1–2.3) using propensity score methods.
Population Women with twin gestations undergoing sonographic
Conclusion Women with more risk factors for preterm delivery
cervical length screening.
were more likely to be treated with vaginal progesterone. After statistically correcting for this with propensity score methods, we found that vaginal progesterone therapy in twin pregnancies with a CL ≤2.5 cm was associated with an increased risk of preterm delivery.
in twin gestations with a cervical length (CL) of ≤2.5 cm is associated with a reduced risk of preterm delivery. Design Retrospective cohort study.
Methods Women with twin gestations with a CL of ≤2.5 cm between 16 and 32 weeks of gestation, and who delivered at our centre between 2010 and 2013, were included. We evaluated the impact of vaginal progesterone on the risk of preterm delivery using a Cox proportional hazard model, adjusted for potential confounding factors. We then performed a propensity score analysis using inverse probability of treatment weights to account for treatment selection bias and confounding. Main outcome measure Delivery prior to 35 weeks of gestation.
Keywords Propensity score, short cervix, twins, vaginal progesterone. Linked article This article is commented on by Biggio JR et al. To view this mini commentary visit http://dx.doi.org/10.1111/ 1471-0528.13187.
Results Of the 167 twin pregnancies analysed, 61 (35.7%) were treated with vaginal progesterone. The hazard ratio (HR) of Please cite this paper as: Brubaker SG, Pessel C, Zork N, Gyamfi-Bannerman C, Ananth CV. Vaginal progesterone in women with twin gestations complicated by short cervix: a retrospective cohort study. BJOG 2014; DOI: 10.1111/1471-0528.13188.
Introduction The rate of twin pregnancy in the USA has risen substantially over the last three decades, a trend that is largely explained by the increased use of assisted reproductive technology and, to a lesser extent, advanced maternal age.1 The Centers for Disease Control and Prevention estimated that one in 53 deliveries in the USA in 1980 was a twin pregnancy, compared with one in 30 in 2009.1 Multiple gestations contribute disproportionally to the burden of prematurity in the USA, primarily as a result of the increased risk of spontaneous delivery in the extremely pre-
ª 2014 Royal College of Obstetricians and Gynaecologists
term period. Infants delivered prior to 32 weeks of gestation account for almost half of all long-term neurological morbidity attributable to preterm delivery, and twins are between four and five times more likely than singletons to deliver before 32 weeks of gestation.2,3 Sonographic assessment of cervical length (CL) has long been shown to have comparable efficacy in predicting preterm delivery in twin gestations as in singletons.2,4,5 A midtrimester CL ≤2.5 cm is associated with a three-fold increase in the odds of delivery before 35 weeks of gestation;2 however, the two mainstays of therapy to reduce the risk of preterm delivery in singletons with short cervix, cervical
1
Brubaker et al.
cerclage and progesterone therapy, have yielded equivocal results in twins.6–9 To date, there is no strong evidence that either intervention reduces the risk of preterm delivery in twin gestations complicated by short cervix.10 Moreover, in the case of cerclage, there is a possibility that the intervention increases the risk of preterm birth.11 Small sample size is the primary limitation of all of the randomised trials of vaginal progesterone in twin pregnancies.12,13 Until an adequately powered randomised trial evaluating this subject is completed, information can be obtained from observational studies; however, observational studies are subject to both treatment selection and confounding biases, as treatment allocation is not random. In this study, we perform a propensity score analysis to minimise treatment selection and confounding biases. The propensity score is defined as a study participant’s probability of receiving treatment based on observed covariates.14 By accounting for the probability of receiving treatment, one can replicate some of the characteristics of a randomised trial.15 A propensity score analysis allows for a more accurate evaluation of the causal effects of a specific treatment in a setting where the treatment is not randomly allocated. We evaluated the impact of vaginal progesterone therapy on the risk of preterm delivery in twin gestations with short cervix using traditional adjustment for confounding factors followed by the propensity score analysis.
Methods This is a retrospective cohort study of all women with twin gestations who were found to have CL ≤2.5 cm on transvaginal cervical length screening between 160/7 and 316/7 weeks of gestation at Columbia University Medical Center (CUMC), New York, between January 2010 and June 2013. This study was approved by the CUMC human subjects institutional review board (protocol AAAJ6006). The exposure of interest was treatment with vaginal progesterone and the primary outcome was time to delivery prior to 35 weeks of gestation. Exclusion criteria were: medically indicated preterm delivery at
Vaginal progesterone in women with twin gestations complicated by short cervix: a retrospective cohort study SG Brubaker,a C Pessel,a N Zork,a C Gyamfi-Bannerman,a CV Ananthb,c a
Division of Maternal–Fetal Medicine, b Department of Obstetrics and Gynecology, c Department of Epidemiology, Columbia University, New York City, NY, USA Correspondence: Dr SG Brubaker, Division of Maternal–Fetal Medicine, Department of Obstetrics and Gynecology, Columbia University College of Physicians and Surgeons, 622 West 168th Street, New York City, NY 10032, USA. Email [email protected] Accepted 8 October 2014. Published Online 27 November 2014.
Objective To determine whether the use of vaginal progesterone
Setting Tertiary-care medical centre in New York City.
delivery prior to 35 weeks of gestation in the vaginal progesterone group, compared with the no vaginal progesterone group, was 1.8 (95% confidence interval, 95% CI 1.5–3.1) in the unadjusted analysis, 1.4 (95% CI 0.7–3.2) following multivariable adjustment for confounding factors, and 1.5 (95% CI 1.1–2.3) using propensity score methods.
Population Women with twin gestations undergoing sonographic
Conclusion Women with more risk factors for preterm delivery
cervical length screening.
were more likely to be treated with vaginal progesterone. After statistically correcting for this with propensity score methods, we found that vaginal progesterone therapy in twin pregnancies with a CL ≤2.5 cm was associated with an increased risk of preterm delivery.
in twin gestations with a cervical length (CL) of ≤2.5 cm is associated with a reduced risk of preterm delivery. Design Retrospective cohort study.
Methods Women with twin gestations with a CL of ≤2.5 cm between 16 and 32 weeks of gestation, and who delivered at our centre between 2010 and 2013, were included. We evaluated the impact of vaginal progesterone on the risk of preterm delivery using a Cox proportional hazard model, adjusted for potential confounding factors. We then performed a propensity score analysis using inverse probability of treatment weights to account for treatment selection bias and confounding. Main outcome measure Delivery prior to 35 weeks of gestation.
Keywords Propensity score, short cervix, twins, vaginal progesterone. Linked article This article is commented on by Biggio JR et al. To view this mini commentary visit http://dx.doi.org/10.1111/ 1471-0528.13187.
Results Of the 167 twin pregnancies analysed, 61 (35.7%) were treated with vaginal progesterone. The hazard ratio (HR) of Please cite this paper as: Brubaker SG, Pessel C, Zork N, Gyamfi-Bannerman C, Ananth CV. Vaginal progesterone in women with twin gestations complicated by short cervix: a retrospective cohort study. BJOG 2014; DOI: 10.1111/1471-0528.13188.
Introduction The rate of twin pregnancy in the USA has risen substantially over the last three decades, a trend that is largely explained by the increased use of assisted reproductive technology and, to a lesser extent, advanced maternal age.1 The Centers for Disease Control and Prevention estimated that one in 53 deliveries in the USA in 1980 was a twin pregnancy, compared with one in 30 in 2009.1 Multiple gestations contribute disproportionally to the burden of prematurity in the USA, primarily as a result of the increased risk of spontaneous delivery in the extremely pre-
ª 2014 Royal College of Obstetricians and Gynaecologists
term period. Infants delivered prior to 32 weeks of gestation account for almost half of all long-term neurological morbidity attributable to preterm delivery, and twins are between four and five times more likely than singletons to deliver before 32 weeks of gestation.2,3 Sonographic assessment of cervical length (CL) has long been shown to have comparable efficacy in predicting preterm delivery in twin gestations as in singletons.2,4,5 A midtrimester CL ≤2.5 cm is associated with a three-fold increase in the odds of delivery before 35 weeks of gestation;2 however, the two mainstays of therapy to reduce the risk of preterm delivery in singletons with short cervix, cervical
1
Brubaker et al.
cerclage and progesterone therapy, have yielded equivocal results in twins.6–9 To date, there is no strong evidence that either intervention reduces the risk of preterm delivery in twin gestations complicated by short cervix.10 Moreover, in the case of cerclage, there is a possibility that the intervention increases the risk of preterm birth.11 Small sample size is the primary limitation of all of the randomised trials of vaginal progesterone in twin pregnancies.12,13 Until an adequately powered randomised trial evaluating this subject is completed, information can be obtained from observational studies; however, observational studies are subject to both treatment selection and confounding biases, as treatment allocation is not random. In this study, we perform a propensity score analysis to minimise treatment selection and confounding biases. The propensity score is defined as a study participant’s probability of receiving treatment based on observed covariates.14 By accounting for the probability of receiving treatment, one can replicate some of the characteristics of a randomised trial.15 A propensity score analysis allows for a more accurate evaluation of the causal effects of a specific treatment in a setting where the treatment is not randomly allocated. We evaluated the impact of vaginal progesterone therapy on the risk of preterm delivery in twin gestations with short cervix using traditional adjustment for confounding factors followed by the propensity score analysis.
Methods This is a retrospective cohort study of all women with twin gestations who were found to have CL ≤2.5 cm on transvaginal cervical length screening between 160/7 and 316/7 weeks of gestation at Columbia University Medical Center (CUMC), New York, between January 2010 and June 2013. This study was approved by the CUMC human subjects institutional review board (protocol AAAJ6006). The exposure of interest was treatment with vaginal progesterone and the primary outcome was time to delivery prior to 35 weeks of gestation. Exclusion criteria were: medically indicated preterm delivery at
