Background. Our aim was to describe the steps in planning, implementing, and running a multicentre cohort study of maternal and perinatal health using a high-quality biobank comprised of maternal serum, plasma, and hair samples collected from five sites in Brazil. The Preterm SAMBA study, conducted by the Brazilian Network for Studies on Reproductive and Perinatal Health, was an innovative approach used to identify women at higher risk for preterm birth. It is also of great importance in the study of other maternal and perinatal complications in the context of Brazil, which is a middle-income country. Methods. We described phases of planning, implementing, and running the Preterm SAMBA study, a multicentre Brazilian cohort study of low-risk nulliparous pregnant women, to validate a set of metabolite biomarkers for preterm birth identified in an external cohort. Procedures and strategies used to plan, implement, and maintain this multicentre preterm birth study are described in detail. Barriers and experience cited in the current narrative are not usually discussed in the scientific literature or published study protocols. Results. Several barriers and strategies were identified in different phases of the Preterm SAMBA study at different levels of the study framework (steering committee; coordinating and local centres). Strategies implemented and resources used in the study are a legacy of the Brazilian Network, aimed at training collaborators in such complex settings. Conclusion. The Brazilian Network for Studies on Reproductive and Perinatal Health has gained some experience in conducting a multicentre cohort study using a resourceful biobank which may be helpful to other research groups and maternal/perinatal health networks that plan on employing a similar approach to a similar background.