Correlation between neonatal outcomes of twins depends on the outcome: secondary analysis of twelve randomised controlled trials

Accounting for Twins and Other Multiple Births in Perinatal Studies Conducted Using Healthcare Administration Data

The analysis of perinatal studies is complicated by twins and other multiple births even when they are not the exposure, outcome, or a confounder of interest. Common approaches to handling multiples in studies of infant outcomes include restriction to singletons, counting outcomes at the pregnancy-level (i.e., by counting if at least one twin experienced a binary outcome), or infant-level analysis including all infants and, typically, accounting for clustering of outcomes by using generalised estimating equations or mixed effects models. Several healthcare administration databases only support restriction to singletons or pregnancy-level approaches. For example, in MarketScan insurance claims data, diagnoses in twins are often assigned to a single infant identifier, thereby preventing ascertainment of infant-level outcomes among multiples. Different approaches correspond to different causal questions, produce different estimands, and often rely on different assumptions. We demonstrate the differences that can arise from these different approaches using Monte Carlo simulations, algebraic formulas, and an applied example. Furthermore, we provide guidance on the handling of multiples in perinatal studies when using healthcare administration data.

Vaginal progesterone for preventing preterm birth and adverse perinatal outcomes in twin gestations: a systematic review and meta-analysis

OBJECTIVE

To evaluate the efficacy of vaginal progesterone for the prevention of preterm birth and adverse perinatal outcomes in twin gestations.

DATA SOURCES

MEDLINE, Embase, LILACS, and CINAHL (from their inception to January 31, 2023), Cochrane databases, Google Scholar, bibliographies, and conference proceedings.

STUDY ELIGIBILITY CRITERIA

Randomized controlled trials that compared vaginal progesterone to placebo or no treatment in asymptomatic women with a twin gestation.

METHODS

The systematic review was conducted according to the Cochrane Handbook for Systematic Reviews of Interventions. The primary outcome was preterm birth <34 weeks of gestation. Secondary outcomes included adverse perinatal outcomes. Pooled relative risks with 95% confidence intervals were calculated. We assessed the risk of bias in each included study, heterogeneity, publication bias, and quality of evidence, and performed subgroup and sensitivity analyses.

RESULTS

Eleven studies (3401 women and 6802 fetuses/infants) fulfilled the inclusion criteria. Among all twin gestations, there were no significant differences between the vaginal progesterone and placebo or no treatment groups in the risk of preterm birth <34 weeks (relative risk, 0.99; 95% confidence interval, 0.84-1.17; high-quality evidence), <37 weeks (relative risk, 0.99; 95% confidence interval, 0.92-1.06; high-quality evidence), and <28 weeks (relative risk, 1.00; 95% confidence interval, 0.64-1.55; moderate-quality evidence), and spontaneous preterm birth <34 weeks of gestation (relative risk, 0.97; 95% confidence interval, 0.80-1.18; high-quality evidence). Vaginal progesterone had no significant effect on any of the perinatal outcomes evaluated. Subgroup analyses showed that there was no evidence of a different effect of vaginal progesterone on preterm birth <34 weeks of gestation related to chorionicity, type of conception, history of spontaneous preterm birth, daily dose of vaginal progesterone, and gestational age at initiation of treatment. The frequencies of preterm birth <37, <34, <32, <30, and <28 weeks of gestation and adverse perinatal outcomes did not significantly differ between the vaginal progesterone and placebo or no treatment groups in unselected twin gestations (8 studies; 3274 women and 6548 fetuses/infants). Among twin gestations with a transvaginal sonographic cervical length <30 mm (6 studies; 306 women and 612 fetuses/infants), vaginal progesterone was associated with a significant decrease in the risk of preterm birth occurring at <28 to <32 gestational weeks (relative risks, 0.48-0.65; moderate- to high-quality evidence), neonatal death (relative risk, 0.32; 95% confidence interval, 0.11-0.92; moderate-quality evidence), and birthweight <1500 g (relative risk, 0.60; 95% confidence interval, 0.39-0.88; high-quality evidence). Vaginal progesterone significantly reduced the risk of preterm birth occurring at <28 to <34 gestational weeks (relative risks, 0.41-0.68), composite neonatal morbidity and mortality (relative risk, 0.59; 95% confidence interval, 0.33-0.98), and birthweight <1500 g (relative risk, 0.55; 95% confidence interval, 0.33-0.94) in twin gestations with a transvaginal sonographic cervical length ≤25 mm (6 studies; 95 women and 190 fetuses/infants). The quality of evidence was moderate for all these outcomes.

CONCLUSION

Vaginal progesterone does not prevent preterm birth, nor does it improve perinatal outcomes in unselected twin gestations, but it appears to reduce the risk of preterm birth occurring at early gestational ages and of neonatal morbidity and mortality in twin gestations with a sonographic short cervix. However, more evidence is needed before recommending this intervention to this subset of patients.

Multiple imputation for handling missing outcome data in randomized trials involving a mixture of independent and paired data

Randomized trials involving independent and paired observations occur in many areas of health research, for example in paediatrics, where studies can include infants from both single and twin births. Multiple imputation (MI) is often used to address missing outcome data in randomized trials, yet its performance in trials with independent and paired observations, where design effects can be less than or greater than one, remains to be explored. Using simulated data and through application to a trial dataset, we investigated the performance of different methods of MI for a continuous or binary outcome when followed by analysis using generalized estimating equations to account for clustering due to the pairs. We found that imputing data separately for independent and paired data, with paired data imputed in wide format, was the best performing MI method, producing unbiased point and standard error estimates for the treatment effect throughout. Ignoring clustering in the imputation model performed well in settings where the design effect due to the inclusion of paired data was close to one, but otherwise led to moderately biased variance estimates. Including a random cluster effect in the imputation model led to slightly biased point estimates for binary outcome data and variance estimates that were too small in some settings. Based on these results, we recommend researchers impute independent and paired data separately where feasible to do so. The exception is if the design effect due to the inclusion of paired data is close to one, where ignoring clustering may be appropriate.

Cervical pessary to prevent preterm birth in asymptomatic high-risk women: a systematic review and meta-analysis

BACKGROUND

Randomized controlled trials that have assessed the efficacy of cervical pessary to prevent preterm birth in asymptomatic high-risk women have reported conflicting results.

OBJECTIVE

To evaluate the efficacy and safety of cervical pessary to prevent preterm birth and adverse perinatal outcomes in asymptomatic high-risk women.

DATA SOURCES

MEDLINE, EMBASE, POPLINE, CINAHL, and LILACS (from their inception to October 31, 2019), Cochrane databases, Google Scholar, bibliographies, and conference proceedings.

STUDY ELIGIBILITY CRITERIA

Randomized controlled trials that compared cervical pessary with standard care (no pessary) or alternative interventions in asymptomatic women at high risk for preterm birth.

STUDY APPRAISAL AND SYNTHESIS METHODS

The systematic review was conducted according to the Cochrane Handbook guidelines. The primary outcome was spontaneous preterm birth <34 weeks of gestation. Secondary outcomes included adverse pregnancy, maternal, and perinatal outcomes. Pooled relative risks with 95% confidence intervals were calculated. Quality of evidence was assessed using the GRADE methodology.

RESULTS

Twelve studies (4687 women and 7167 fetuses/infants) met the inclusion criteria: 8 evaluated pessary vs no pessary in women with a short cervix, 2 assessed pessary vs no pessary in unselected multiple gestations, and 2 compared pessary vs vaginal progesterone in women with a short cervix. There were no significant differences between the pessary and no pessary groups in the risk of spontaneous preterm birth <34 weeks of gestation among singleton gestations with a cervical length ≤25 mm (relative risk, 0.80; 95% confidence interval, 0.43-1.49; 6 trials, 1982 women; low-quality evidence), unselected twin gestations (relative risk, 1.05; 95% confidence interval, 0.79-1.41; 1 trial, 1177 women; moderate-quality evidence), twin gestations with a cervical length <38 mm (relative risk, 0.75; 95% confidence interval, 0.41-1.36; 3 trials, 1128 women; low-quality evidence), and twin gestations with a cervical length ≤25 mm (relative risk; 0.72, 95% confidence interval, 0.25-2.06; 2 trials, 348 women; low-quality evidence). Overall, no significant differences were observed between the pessary and no pessary groups in preterm birth <37, <32, and <28 weeks of gestation, and most adverse pregnancy, maternal, and perinatal outcomes (low- to moderate-quality evidence for most outcomes). There were no significant differences in the risk of spontaneous preterm birth <34 weeks of gestation between pessary and vaginal progesterone in singleton gestations with a cervical length ≤25 mm (relative risk, 0.99; 95% confidence interval, 0.54-1.83; 1 trial, 246 women; low-quality evidence) and twin gestations with a cervical length <38 mm (relative risk, 0.73; 95% confidence interval, 0.46-1.18; 1 trial, 297 women; very low-quality evidence). Vaginal discharge was significantly more frequent in the pessary group than in the no pessary and vaginal progesterone groups (relative risks, ∼2.20; high-quality evidence).

CONCLUSION

Current evidence does not support the use of cervical pessary to prevent preterm birth or to improve perinatal outcomes in singleton or twin gestations with a short cervix and in unselected twin gestations.

Effects of tocolysis with nifedipine or atosiban on child outcome: follow-up of the APOSTEL III trial

OBJECTIVE

To compare the long-term effects of tocolysis with nifedipine or atosiban on child outcome at age 2.5-5.5 years.

DESIGN

The APOSTEL III trial was a multicentre randomised controlled trial that compared tocolysis with nifedipine or atosiban in 503 women with threatened preterm birth. Neonatal outcomes did not differ between both treatment arms, except for a higher incidence of intubation in the atosiban group.

METHODS

Parents were asked to complete four questionnaires regarding neurodevelopment, executive function, behaviour problems and general health.

MAIN OUTCOME MEASURES

The main long-term outcome measure was a composite of abnormal development at the age of 2.5-5.5 years.

RESULTS

Of the 426 women eligible for follow-up, 196 (46%) parents returned the questionnaires for 115 children in the nifedipine group and 110 children in the atosiban group. Abnormal development occurred in 32 children (30%) in the nifedipine group and in 38 children (38%) in the atosiban group (OR 0.74, 95% CI 0.41-1.34). The separate outcomes for neurodevelopment, executive function, behaviour, and general health showed no significant differences between the groups. Sensitivity analysis for all children of the APOSTEL III trial, including a comparison of deceased children, resulted in a higher rate of healthy survival in the nifedipine group (64 versus 54%), but there was no significant difference in the overall mortality rate (5.4 versus 2.7%). There were no significant subgroup effects.

CONCLUSION

Outcomes on broad child neurodevelopment, executive function, behaviour and general health were comparable in both groups. Neither nifedipine nor atosiban can be considered as the preferred treatment for women with threatened preterm birth.

TWEETABLE ABSTRACT

Nifedipine- and atosiban-exposed children had comparable long-term outcomes, including neurodevelopment, executive function and behaviour.

SMFM Special Statement: State of the science on multifetal gestations: unique considerations and importance

We sought to review the state of the science for research on multiple gestations. A literature search was performed with the use of PubMed for studies to quantify the representation of multiple gestations for a sample period (2012-2016) that were limited to phase III and IV randomized controlled trials, that were written in English, and that addressed at least 1 of 4 major pregnancy complications: fetal growth restriction or small-for-gestational-age fetus, gestational diabetes mellitus, preeclampsia, and preterm delivery. Of the 226 studies that are included in the analysis, multiple pregnancies were most represented in studies of preterm delivery: 17% of trials recruited both singleton and multiple pregnancies; another 18% of trials recruited only multiple pregnancies. For trials that studied preeclampsia, fetal growth restriction, and gestational diabetes mellitus, 17%, 8%, and 2%, respectively, recruited both singleton and multiple gestations. None of the trials on these 3 topics were limited to women with a multiple pregnancy. Women with a multiple pregnancy are at risk for complications similar to those of women with singleton pregnancies, but their risk is usually higher. Also, the pathophysiologic condition for some complications differs in multiple gestations from those that occur in singleton gestations. Conditions that are unique to multiple pregnancies include excess placenta, placental crowding or inability of the uteroplacental unit to support the normal growth of multiple fetuses, or suboptimal placental implantation sites with an increased risk of abnormal placental location. Other adverse outcomes in multiple gestations are also influenced by twin-specific risk factors, most notably chorionicity. Although twins have been well represented in many studies of preterm birth, these studies have failed to identify adequate predictive tests (short cervical length established over 2 decades ago remains the single best predictor), to establish effective interventions, and to differentiate the underlying pathophysiologic condition of twin preterm birth. Questions about fetal growth also remain. Twin growth deviates from that of singleton gestations starting at approximately 32 weeks of gestation; however, research with long-term follow-up is needed to better distinguish pathologic and physiologic growth deviations, which include growth discordance among pairs (or more). There are virtually no clinical trials that are specific to twins for gestational diabetes mellitus or preeclampsia, and subgroups for multiple pregnancies in existing trials are not large enough to allow definite conclusions. Another important area is the determination of appropriate maternal nutrition or micronutrient supplementation to optimize pregnancy and child health. There are also unique aspects to consider for research design in multiple gestations, such as designation and tracking of the correct fetus prenatally and through delivery. The correct statistical methods must be used to account for correlated data because multiple fetuses share the same mother and intrauterine environment. In summary, multiple gestations often are excluded from research studies, despite a disproportionate contribution to national rates of perinatal morbidity, mortality, and health-care costs. It is important to consider the enrollment of multifetal pregnancies in studies that target mainly women with singleton gestations, even when sample size is inadequate, so that insights that are specific to multiple gestations can be obtained when results of smaller studies are pooled together. The care of pregnant women with multiple gestations presents unique challenges; unfortunately, evidence-based clinical management that includes the diagnosis and treatment of common obstetrics problems are not well-defined for this population.