Can history repeat itself? Improved assessment of the best biomarker is the first step

Prophylactic therapies aimed at unselected cohorts of women with twin pregnancies have been unsuccessful in reducing the frequency of preterm birth, and have on occasion proven harmful. Supplemental progestogens, the cervical pessary, and bedrest have all failed to reproducibly demonstrate substantial benefit in large trials enrolling unselected populations. Hence, a detailed description of potentially useful biomarkers providing prognostication for testing an intervention, or avoiding an intervention, is the next logical step for developing effective therapeutics in twin pregnancy. In the mid-nineties, the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) Preterm Prediction Study enrolled approximately 2900 women, ultimately comparing a variety of biomarkers and risk factors to predict preterm birth (Iams et al. N Engl J Med 1996;334:567–72). In 2002, Iams et al. later described the superiority of transvaginal cervical length measurement over a select group of biomarkers for predicting preterm birth in singletons (Iams et al. N Engl J Med 2002;346:250–55). Despite these observations, investigators have been slow in defining the utility of biomarkers in a very large data set of twin pregnancies. In this edition of BJOG, Kindinger et al. begin to correct that deficiency. These investigators performed an individual patient or participant data (IPD) meta-analysis that provides essential information for the prediction of preterm birth utilizing cervical length. The conclusions are consistent with an analysis by Conde-Agudelo et al., but provide greater detail (Conde-Agudelo et al. Am J Obstet Gynecol 2010;203:128.e1–12). Kindinger et al. characterize novel information in twin pregnancies, drawing upon important past lessons derived from singleton gestations. As similarly demonstrated by Berghella et al., in order to optimise the prediction of gestational age at delivery in twin pregnancies, the gestational age at screening must be considered (Berghella et al. Obstet Gynecol 2007;110:311–7). The value of cervical length measurement in twin pregnancies at gestational ages <20 weeks of gestation is also described. With this detailed information investigators can now better design trials for women undergoing cervical surveillance, and attempt to alter end points which are readily attainable in this challenging population. Like trials in singletons, the choice of primary outcome is critical for demonstrating efficacy. Any trial in twin gestations testing an intervention indicated for a sonographic short cervix early in the second trimester should strongly consider a primary end point of preterm birth <28 weeks of gestation. From this data, later identification of a short cervix suggests the need to choose a different primary endpoint such as a composite of neonatal morbidity and mortality. The importance of a 100% negative predictive value for birth <28 weeks of gestation provided by this data also has utility, especially for our patients. Kindinger et al. noted that a cervical length of greater than 43 mm at 22–24 weeks of gestation is associated with an exceptionally low risk for second-trimester birth. These data provide valuable reassurance and should limit interventions, such as fetal fibronectin screening. Hopefully cervical length data can eventually be augmented by other biomarkers, when appropriate, to provide even more robust predictions describing more specific pathophysiologies. If history is to repeat itself and the sequence of detailed analysis of cervical length data in a select population biomarker is followed by successful clinical trials, the final step is for phase-three trials to repeatedly demonstrate efficacy. Cervical length is now a better-characterised biomarker for the prediction of preterm birth in twin pregnancies. Let's randomise. Full disclosure of interests available to view online as supporting information. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.