Background: Preterm birth is a major societal and economic problem accounting for 80 to 90% of neonatal morbidity and mortality worldwide. It is recognized as a complex multifactorial condition comprising several distinct clinical phenotypes with different underlying etiologies. As animal models are expensive and fail to mimic the biology of spontaneous preterm birth in humans, understanding the pathophysiology requires detailed clinical studies. Meta-analyses and clinical translation of data, however, are limited by heterogeneity of study design and size, publication and reporting biases, definition of patient groups, and a lack of standard universal definitions. This article provides a harmonized open-source template for designing clinical studies addressing preterm birth.
Methods: Recommendations are made for clinical definitions, choice and assignment to preterm birth phenotypes, selection of enriched populations and control pregnancies, and potential confounding factors. In addition, recommendations are made for study design, sample size and power calculations, the minimal data sets needed for any study of preterm birth, and the optimal data set of an ideal study.
Results: Recommended patient phenotypes are infection, uterine overdistension, hemorrhage, stress (either maternal or fetal), and idiopathic. Confounding factors include medical conditions, obesity, antenatal glucocorticoids, multifetal pregnancies, and fetal sex. Guidelines regarding study design, sample size, and clinical data acquisition are provided to serve as a universal template for preterm birth studies.
Conclusions: Adoption of a harmonized template will allow generation of protocols and studies with a basic degree of compatibility and will allow data to be compared, and samples and data sets to be combined for meaningful meta-analyses.