Triple P study
Go to the Triple P website
To evaluate whether a screening program with cervical length measurement to find women at risk for a preterm delivery, is cost-effective (the Triple P screening study). Women with a short cervical length are asked to take part in a follow-up study (the Triple P treat study) that evaluates whether subsequent progesterone treatment is effective.
Multicenter cohort study (Triple P screening) with a subsequent randomised clinical trial (Triple P treat).
Women with a singleton pregnancy undergoing foetal assessment at 16-20 weeks will be screened for a short cervical length (the Triple P screening study). If at 22 weeks this finding can be reproduced, women with a cervical length shorter than 30 mm will be invited to participate in a randomised clinical trial (the Triple P treat study). Cervical length will be measured with 40.000 women, in order to include 1920 women in the randomised clinical trial.
Women participating in the Triple P treat study receive either vaginal progesterone (200 mg each night) or placebo from 24 to 34 weeks of gestation.
Primary outcome is bad neonatal outcome. Secondary outcome measures are delivery before 34 weeks, child health, growth and development at 2 years, time to delivery, preterm birth rate before 32 and 37 weeks, days of neonatal admission, maternal morbidity, and maternal admission for preterm labour and costs.
The Triple P treat study: We anticipate a bad neonatal outcome in 14% of the pregnancies in which the cervical length is < 15 mm, and in 3% of the pregnancies with a cervical length between 15 mm and 30 mm. As 1.7% of the pregnancies will show a cervical length < 15 mm, and 8.3% of the pregnancies a cervical length between 15 and 30 mm, the probability of bad neonatal outcome without intervention with progesterone in women with a cervix smaller than 30 mm will be 5.0%. This percentage is expected to be reduced to 2.5% after use of progesterone. Assuming this decrease of the incidence of bad neonatal outcome from 5% to 2.5%, using a two sided test with an alpha of 0.05 and a beta of 0.8, 1,920 women (960 per arm) are needed in this study.
The Triple P screening study: As we expect that 10% of the women will have a cervix < 30 mm, and if we assume that 50% of the eligible women will participate, we need to screen almost 40,000 women in this study.
The results of the randomised clinical trial (the Triple P treat study) will be analysed according to the intention to treat principle. The effectiveness of progesterone versus placebo will be assessed by calculating relative risks and 95% confidence intervals. Time to pregnancy will be compared using Kaplan-Meier analysis. We will evaluate whether there is an interaction between the treatment effect and cervical length, and between treatment effect and ethnicity.
The study design will enable us to compare the costs and effects of the following strategies:
I. no screening for cervical length
II. screening for cervical length, and treatment of women with a short cervical length
For each of these strategies, we will calculate the costs as well as the effects in terms of bad neonatal outcome (mortality and severe morbidity, see above). In the cost-effectiveness analysis, we will then calculate the costs per prevented case of bad neonatal outcome.
For strategy II, we will assess the screening strategy using different cut-off levels. We anticipate a stronger treatment effect in the group with a very short cervix, i.e. < 15 mm, whereas the treatment effect is likely to be less strong in the women with a cervix between 15 mm and 30 mm (number need to treat). On the other hand, the number of women with a short cervix will be smaller when a cut-off value of 15 mm is used than when the group with a cervix between 15 and 30 mm is considered to be abnormal (number needed to screen). Thus, the cost-effectiveness analysis will assess the balance between number needed to treat and number needed to screen.
We will need a run-in period of six months for the study set up, and 24 months for
inclusion. After inclusion of the last patient, 30 months (6 months pregnancy and 24 months for evaluating the child at age 2) are needed for follow-up data collection and report of results. The first report on the primary outcome is expected at three years after start of the study.
Prof. dr. B.W.J. Mol, Gynaecologist and Epidemiologist,
Department of Gynaecology and Obstetrics, AMC Amsterdam