PROMPT: Prospective Meta-analysis for Pessary Trials Study Protocol

Abstract

Objective

Preterm birth, defined as birth before 37 weeks of gestation, is a leading cause of perinatal and infant mortality throughout the world. Preterm birth is also associated with long-term neurological disabilities and other significant health issues in children. A short cervix in the second trimester has been noted to be one of the strongest predictors of subsequent spontaneous preterm birth in both singleton and multiple pregnancies. Some studies have shown that cervical support in the form of an Arabin pessary lowers the risk of preterm birth in women with a singleton gestation and short cervical length; however, other studies have conflicting results. Our objective was to form an international collaborative of planned or ongoing randomized trials of pessary in singleton and twin gestations with a short cervix.

Study Design

In November 2014, an international group of investigators, who had initiated or were planning randomized trials of pessary for pregnant people with a short cervix and singleton or twin gestation to prevent preterm birth, formed a collaboration to plan a prospective individual patient data (IPD) meta-analysis of randomized trials (PROspective Meta-analysis of Pessary Trials [PROMPT]). The PROMPT investigators agreed on meta-analysis IPD hypotheses for singletons and twins, eligibility criteria, and a set of core baseline and outcome measures. The primary outcome is a composite of fetal death or preterm delivery before 32 weeks’ gestation. Secondary outcomes include maternal and neonatal morbidities. The PROMPT protocol may be viewed as a written agreement among the study investigators who make up the PROMPT consortium (PROSPERO ID# CRD42018067740).

Results

Results will be published in phases as the individual participating studies are concluded and published. Results of the first phase of singleton and twin pessary trials are expected to be available in late 2022. Updates are planned as participating trials are completed and published.

More than 1 in 10 of the world’s babies born in 2010 were born prematurely, resulting in an estimated 15 million preterm births before 37 weeks gestation of which more than 1 million died as a result of their prematurity.¹ Prematurity is now the second-leading cause of death in children under 5 years and the single most important cause of death in the critical first month of life.² For the babies who survive, especially early preterm birth before 32 weeks of gestation, many face a lifetime of significant disability. The incidence of preterm birth in high resource countries is 5 to 13% and, despite advances in health care, the rate of prematurity has not declined over the last four decades.^2,3 Given the economic, medical, and psychological burden of preterm birth, the efforts at identification of risk factors and investigation of preventative treatments have been considerable over the last several decades.

A short cervix during pregnancy, as measured by endo-vaginal ultrasound, was first reported to be a significant risk factor for spontaneous preterm delivery by Andersen et al in 1990.⁴ Subsequently, multiple studies have confirmed this finding in women with both singleton and twin pregnancies. In singleton pregnancies, a cervical length < 25 mm at 22 to 24 weeks was associated with a 6-fold increase in risk for preterm birth before 32 weeks (Relative risk (RR): 6.10, 95% confidence interval [CI]: 3.88–9.58, unpublished data from the Preterm Prediction Study). In twin pregnancies, after adjusting for other factors, a short cervix 25 mm was associated with a nearly 8-fold increase in the odds of a spontaneous preterm birth <32 weeks (odds ratio [OR]: 7.7, 95% CI: 1.3–11.8).⁵ Similarly, a meta-analysis demonstrated that asymptomatic women with twin gestation and cervical length 20 mm at 20 to 24 weeks had a nearly 10-fold increase in the risk of preterm birth prior to 34 weeks.⁶ Clearly, the risk of preterm birth by cervical length differs between singletons and twins, with a higher risk for preterm birth for a longer cervix in twin pregnancies.

Progesterone supplementation had shown potential as a prophylaxis to prevent recurrent preterm birth as early as 1975 and several trials of progesterone to prevent preterm delivery have been conducted since that time. In 2012, an individual participant data meta-analysis of singleton and twins trials to evaluate the efficacy and safety of vaginal progesterone for the prevention of preterm birth and neonatal morbidity and mortality showed a significant reduction in preterm birth <35 weeks (RR: 0.69, 95% CI: 0.55–0.88) and composite neonatal morbidity and mortality (RR: 0.57, 95% CI: 0.40–0.81) among other outcomes for patients who were treated with progesterone.⁷ And in 2017, an individual participant data meta-analysis of 303 women with twin gestation and short cervix showed that vaginal progesterone, compared with placebo/no treatment, was associated with a reduction in the risk of preterm birth < 33 weeks of gestation (RR: 0.69, 95% CI: 0.51–0.93).⁸ However, one unblinded trial with a considerable risk of bias provided 75% of the data and this study has since been retracted. When analyzed without the data from the study, there was no significant reduction in the preterm birth rate.⁹

Cervical cerclage is a surgical procedure whereby a suture is placed around the endocervical canal, aiming to give mechanical support to the cervix and thereby reducing the risk of preterm birth. A 2012 Cochrane review with eight included trials demonstrated efficacy of cerclage in preventing preterm birth for women with a high-risk singleton pregnancy (210/1,196 vs. 277/1,196; RR: 0.79, 95% CI: 0.68–0.93) with no significant subgroup differences depending on the indication for cerclage.¹⁰ In a separate meta-analysis focusing only on five trials of women with prior preterm birth and a short cervix (<25 mm), Berghella et al noted a decrease in the rate of preterm birth < 35 weeks (RR: 0.70, 95% CI: 0.55–0.89). This benefit, however, has not been seen in twin pregnancies or pregnancies without a history of preterm birth. In the Berghella meta-analysis, there was actually a significant increase in preterm birth at less than 35 weeks in twin gestations (RR: 2.15, 95% CI: 1.15–4.01).¹¹ However, a reanalysis by Saccone et al in 2015 that adjusted for covariates that could affect the risk of preterm birth concluded that “the twin data seems to point mostly to no effect of cerclage, rather than to a detrimental effect” (RR: 2.19, 95% CI: 0.72–6.63).”¹² In summary, study results of cerclage use for women with short cervix and twin pregnancy are inconclusive.

Cervical pessary is another form of cervical support. The pessary has some benefits over cerclage in that the insertion is nonsurgical and, in theory, it may mechanically tilt the cervix lowering mechanical stresses, rather than a cerclage that simply encircles the cervix. In a Letter to the Editor, Cross first reported on the use of a pessary for cervical incompetence as far back as 1959.¹³ Subsequent case reports and series have suggested that the Arabin pessary, with its design that encircles the cervix and alters the cervicouterine angle, may offer benefit among women at risk for preterm birth.¹⁴ In a trial conducted in Spain (Pesario Cervical para Evitar Prematuridad [PECEP]), 385 women with a singleton pregnancy and a short cervix (<25mm) were randomized to either pessary or expectant management. The authors found a significant reduction in the rate of spontaneous preterm birth before 34 weeks among women with a short cervix (<25mm; 12 [6%] vs. 51 [27%], OR: 0·18, 95% CI: 0·08–0·37).¹⁵

The Dutch obstetric research consortium conducted a trial of pessary versus no pessary in twin pregnancies and showed that in the subgroup of women with short cervical length (<38 mm, 25th percentile), the pessary was associated with a significant reduction in the rate of preterm delivery less than 32 weeks (14 vs. 29%, RR: 0.49, 95% CI: 0.24–0.97) and in the rate of adverse neonatal outcome (10 vs. 24%, RR: 0.42, 95% CI: 0.19–0.91).¹⁶ These encouraging results for the pessary in singletons and twins with short cervix were not confirmed by other trials.^17–19

In view of all the studies performed until the formation of the consortium in 2014, three conclusions can be drawn. First, a short cervix measured in the second trimester identifies women at risk for preterm birth. Second, the cervical pessary has shown some potential as a treatment for women who are at increased risk for preterm birth due to a short cervix, but results are conflicting. Third, the factors contributing to preterm birth and the interventions for the prevention of preterm birth differ between singleton and twin gestations. These conclusions inspired several investigators and investigator groups worldwide to conduct further research in this area. The PROMPT collaboration was developed to investigate prespecified research questions with uniformly defined endpoints that cannot be addressed by any one individual trial due to recruitment, time, and funding constraints. The protocol was registered on the international database of prospective systematic reviews (PROSPERO ID CRD42018067740).

Objectives

The primary research questions to be addressed are:

• Does cervical support in the form of a pessary placed in the second trimester compared with no pessary for women with singleton gestation and short cervix less than 25 mm measured at or before 23 weeks 6 days of gestation reduce the risk of fetal death or preterm delivery before 32 weeks of gestation?

• Does cervical support in the form of a pessary placed in the second trimester compared with no pessary for women with twin gestation and short cervix less than 30 mm measured at or before 23 weeks 6 days of gestation reduce the risk of fetal death or preterm delivery before 32 weeks of gestation?

Hypotheses

Singleton:

Compared with no pessary, cervical support in the form of a pessary placed in the second trimester compared with no pessary is associated with a one-third reduction from 12.0 to 8.0% in fetal death or preterm delivery before 32 weeks of gestation.

Twins:

Compared with no pessary, cervical support in the form of a pessary placed in the second trimester compared with no pessary is associated with a one-third reduction from 20.0 to 13.3% in fetal death or preterm delivery before 32 weeks of gestation.

Sample Size

Singleton:

A total sample size of 972 women (Table 1) would have 88% power to detect a one-third reduction in the pessary arm compared with standard of care from 12 to 8%, with a two-sided type 1 error of 5%.

Table 1.

Characteristics of randomized trials participating in the PROMPT collaboration

Trial acronym, registry number, and PMID number (if published)	Country of recruitment	Inclusion criteria	Gestational age at randomization	Intervention	Comparison group (s)	Primary outcome	Planned sample size per group
P5 RBR-3t8prz PMID: 34856583	Brazil	Singleton or twin CL ≤ 30 mm	180/7–226/7	Ingamed pessary plus vaginal progesterone (200 mg daily)	Vaginal progesterone (200 mg daily)	Composite of stillbirth, neonatal death, PVL, severe RDS, BPD, IVH grade IIB or worse, NEC, proven sepsis; before discharge	475a
NCT 02716909 PMID: 29260226	Italy	Singleton CL ≤ 25 mm	180/7–236/7	Arabin pessary	Standard of care	Spontaneous PTD < 34 wk	150a
Quad P NTR 4414	The Netherlands	Singleton CL ≤ 35 mm	180/7–226/7	Arabin pessary	Vaginal progesterone (200 mg daily)	Composite of PVL > grade I, severe RDS, BPD, IVH grade III or IV, NEC > stage 1, proven sepsis, (intrapartum) stillbirth, and death; before discharge	314
PC NTR 4415	The Netherlands	Singleton (1) prior PTD < 34 wk and CL < 25 mm (2) multiple PTDs and eligible for cerclage < 16 wk	160/7–236/7	Arabin pessary	Cerclage	PTD < 32 wk	222
RECAP ISRCTN 11186205	United Kingdom	Singleton CL < 3rd centile + prior SPTD or pPROM < 34 wk or history of significant cervical surgery	160/7–246/7	Arabin pessary	• Vaginal progesterone (200 mg daily) • Cerclage	Feasibility	8b
TO PS NCT 02901626	United States	Singleton CL ≤ 20 mm	180/7–236/7	Arabin pessary	Standard of care	PTD or fetal death < 37 wk	425
PoPPS NCT 02056652 PMID: 28940481	United States	Singleton CL ≤ 25 mm	180/7–236/7	Bioteque cup pessary	Standard of care	PTD < 37 wk	61*
SuPPoRT ISRCTN 13364447	United Kingdom	Singleton < 25 mmand Prior pPROM or PTD/ second trimester loss or cervical procedure	140/7–236/7	Arabin pessary	• Vaginal progesterone (200 mg daily) • Cerclage	PTD < 37 wk	180
STOPPIT-2 ISRCTN 98835694 PMID: 33780463	United Kingdom	Twin ≤30 mm	180/7–206/7	Arabin	Standard of care	• Spontaneous PTD < 34 wk • Composite of stillbirth or neonatal death, PVL, early respiratory morbidity, IVH, NEC, proven sepsis; up to 4 wk after EDD	250a
PROSPECT NCT 02518594	United States	Twin <30 mm	160/7–236/7	Arabin pessary	• Vaginal progesterone (200 mg daily) • Vaginal placebo	Delivery or fetal death of either twin < 35 wk	210
PESSAR’ ONE NCT 02328989 PMID: 35123930	France	Twin <35 mm	160/7–240/7	Arabin pessary	Standard of care	Composite of fetal or neonatal death, BPD, IVH grade III or IV, PVL, NEC grade II or higher, proven sepsis, ROP requiring treatment	157a
PoPPT NCT 02056639 PMID: 28170117	USA	Twin ≤30 mm	180/7–276/7	Bioteque Cup	Standard of care	PTD < 34 wk	23a
AP NCT 02623881 PMID: 30741812	Vietnam	Twin <38 mm	160/7–220/7	Arabin pessary	Vaginal progesterone (400 mg daily)	PTD < 34 wk	150a
Quad P Twins NTR 4414	The Netherlands	Twin ≤38 mm	160/7–206/7	Arabin pessary	Vaginal progesterone (200 mg daily)	Composite of PVL > grade I, severe RDS, BPD, IVH grade III and IV, NEC > stage 1, proven sepsis, (Intrapartum) stillbirth, and death before discharge	166

Abbreviations: BPD, bronchopulmonary dysplasia; CL, cervical length; EDD, estimated delivery date; IVH, intraventricular hemorrhage; NEC, necrotizing enterocolitis; pPROM, preterm premature rupture of membranes; PTD, preterm delivery; PVL, periventricular leukomalacia; RDS, respiratory distress syndrome; ROP, retinopathy of prematurity; US, ultrasound.

^aActual sample size in pessary arm at end of recruitment.

^bTotal enrolled.

Twins:

A total sample size of 753 women (Table 1) would have 88% power to detect a one-third reduction in the pessary arm compared with standard of care from 20 to 13.3% with a two-sided type 1 error of 5%.

Materials and Methods

The Principal Investigators of all eligible trials were approached and agreed to participate in the PROMPT Collaboration and provide the relevant individual patient data (IPD) upon completion of their respective trials. A set of core variables and common definitions were agreed upon by the members, and a detailed analysis plan will be developed prior to the analysis of the combined data.

Eligibility Criteria for Studies to Be Included

• Randomized controlled trial of pessary for preterm birth prevention. The trial must be registered with a clinical trials registry.

• A full protocol must be available and shared. The protocols and consent form must be approved by the local ethics board before trial start-up.

• There must be no impediment to sharing data with the PROMPT data coordinating center.

• Adequate randomization method.

• Central randomization assignment with adequate randomization concealment (i.e., protection of the allocation sequence before and until assignment).

• By protocol, a cervical length measurement must be performed on each participant before randomization. There must be a study-wide method for standardizing cervical length measurement.

• There must be a study-wide method for training providers in pessary placement.

• The main inclusion criterion for the trial cannot be placenta previa. Inclusion criteria may be short cervical length, cervical incompetence, or prior preterm birth.

• The protocol and case report forms must have been available to the PROMPT Executive Committee by December 31, 2016.

Participants

For data from a participant to be included, all of the following criteria must be satisfied at the time of that participant’s randomization.

• Cervical length <40 mm by cervical length measurement prior to randomization. The upper limit of 40 mm should allow for population differences in cervical length distributions.

• Asymptomatic at randomization, i.e., at randomization the participant must have intact membranes and not be in labor. The indication for her inclusion into the randomized trial can be for short cervix or cervical incompetence, but there cannot be a plan for a history-indicated cerclage indicated by her obstetrical history. Participants with cerclage prior to randomization will be excluded.

• Gestational age at cervical length measurement 16 weeks 0 days or later.

• Gestational age at randomization less than or equal to 23 weeks 6 days.

• The participant must have had a dating ultrasound prior to randomization for a spontaneously conceived pregnancy. Gestational age for spontaneous conception will be estimated by a combination of the dating ultrasound results and first date of last menstrual period, according to the local standards.

Interventions

The intervention is a pessary that encircles the cervix (such as the Arabin, Bioteque, or Ingamed pessaries). Open-label vaginal progesterone or cerclage may be included as part of the intervention or as a randomized intervention. Cerclage may be included as an additional randomization arm. The pessary should be placed shortly after randomization and is intended to remain in place until at least 32 weeks of gestation.

The comparison groups include:

• Standard of care, which is defined as either (1) no pessary if the pessary arm does not specify open-label vaginal progesterone, (2) identical-appearing placebo to vaginal progesterone, if progesterone is included as an active comparator, or (3) vaginal progesterone or cerclage if the same treatment is also used in the intervention arm (Fig. 1).

• Vaginal progesterone or cerclage may also be active comparators. In this case the comparison is between a pessary only arm and a progesterone only arm or a cerclage only arm.

Fig. 1.

Primary comparison of pessary versus control.

Outcomes

See Table 2

Table 2.

Primary and secondary outcomes

Primary	Composite of fetal death or preterm delivery before 32 wk gestation
Secondary fetal/neonatal	• Fetal death (at any time after randomization) • Neonatal death up to 28 d after estimated delivery date • Serious neonatal morbidity up to 28 d after estimated delivery date. This is defined by intraventricular hemorrhage grades 3 or 4, periventricular leukomalacia grades III or IV, chronic lung disease, retinopathy of prematurity requiring treatment, necrotizing enterocolitis (Bell Stage II or above), proven early sepsis <72 h, patent ductus arteriosus requiring treatment • Major fetal anomaly • Seizures requiring treatment • Gestational age at fetal death or delivery • Small for gestational age by World Health Organization standards (<5th and 10th percentiles) • Admission to special care nursery • Length of stay in hospital (total including other institutions if transferred)
Secondary maternal	• Reasons for removal of the pessary • Vaginal discharge • Vaginal bleeding • Type of labor (no labor, spontaneous, induced) • Maternal complications including genital infection, urinary infection, postpartum infection • Preterm premature rupture of membranes (pPROM) • Type of membrane rupture • Type of delivery, use of tocolysis, use of magnesium sulfate, use of corticosteroids • Maternal death up to 28 d after estimated delivery date • Maternal length of stay in hospital

Notes: The primary and secondary outcomes were chosen to be consistent with the CROWN initiative (Core Outcomes in Women’s Health), in particular with the subset chosen to evaluate interventions to prevent preterm birth, as agreed by PROMPT investigators at the time of protocol development.^20,21

• Planned subgroup analyses

Subgroup effects will be investigated using an interaction term between the subgroup of interest and treatment in the regression model. If an interaction is found to be statistically significant (p< 0.05), a stratified analysis will be performed to investigate the effect of the pessary in different strata of the subgroups. The following subgroup analyses are planned:

• Type of pessary (Arabin, other).

• Cervical length (< 10, 10–≤20,>20;< 25,≥25);< 30,≥30 mm.

• Obstetrical history (nulliparous, previous preterm delivery, previous term delivery only).

• Gestational age at randomization (< 20, ≥ 20 wk).

• Chorionicity (twins only).

• Low- or middle- versus high-resource country.

• Chronology (first half of enrollment vs. second half).

• The interaction between treatment group and gestational age at randomization or cervical length as continuous variables will also be considered.

Ethics and Management Issues

Search Methods for Identification of Studies

Searches for trials that may be eligible for participation in this prospective meta-analysis include searches of ClinicalTrials.gov and the International Clinical Trials Registry Platform (ICTRP). Searches continued through 2019. Searches have also been conducted for published protocols on online databases such as MEDLINE and Embase as well as internet searches for nonpeer reviewed articles and other publications using Google. The following terms are used: (1) for ICTRP; an initial search was made on July 1, 2015 using search term “preterm” and (2) for all other search methods the following terms are used “pessary” AND “cervical length”; “pessary” AND “preterm.” Trials will only be allowed to join PROMPT if they have started recruitment before the results of any of the other participating trials are made available and if the planned completion of enrollment is before an end-date to be determined. However, there will be an effort to add later trials to update the initial individual participant meta-analysis in later phases.

Project Organization

Membership of the PROMPT Collaboration includes representative(s) from each of the trials contributing data to the project with an accompanying project coordination and data management structure. The Steering Committee includes one or two investigators representing each trial as well as consultants in IPD meta-analysis, epidemiology, biostatistics, and clinical trials. An Executive Committee will provide oversight and monitor progress. Data management and analysis will be performed by the Data Coordinating Center of the Eunice Kennedy Shriver Institute of Child Health and Human Development (NICHD) Maternal Fetal Medicine Units (MFMU) Network, located at the George Washington University Biostatistics Center, United States.

Publication Policy

Each trial has the right to publish their study before the PROMPT Collaboration results are published. When publishing individual results, participating trials are asked to include reference to the PROMPT Collaboration.

Participating Trials

Eight singleton trials and seven twins trials were identified as meeting study eligibility and are included in PROMPT. A summary of singleton and twins trials is provided in Table 1.

Key Points.

• Short cervical length predicts preterm birth.

• Results of prior cervical pessary trials are mixed.

• Meta-analysis of pessary trials protocol.