Towards a new taxonomy of preterm birth

Abstract

Disease categories traditionally reflect a historical clustering of clinical phenotypes based on biologic and nonbiologic features. Multiomics approaches have striven to identify signatures to develop individualized categorizations through tests and/or therapies for 'personalized' medicine. Precision health classifies clinical syndromes into endotype clusters based on novel technological advancements, which can reveal insights into the etiologies of phenotypical syndromes. A new taxonomy of preterm birth should be considered in this context, as not all preterm infants of similar gestational ages are the same because most have different biologic vulnerabilities and hence different health trajectories. Even the choice of interventions may affect observed clinical conditions. Thus, a new taxonomy of prematurity would help to advance the field of neonatology, but also obstetrics and perinatology by adopting anticipatory and more targeted approaches to the care of preterm infants with the intent of preventing and treating some of the most common newborn pathologic conditions.

Fig. 1. A new data-driven taxonomy of prematurity expressed throughout pregnancy and early life.

Preterm infants of similar gestational ages and birthweight are not all the same – some are more mature than others despite shared chronology and all have different biological vulnerabilities that can affect multiple organ systems. Thus, a more complete taxonomy of prematurity is needed that could track infant prognosis throughout pregnancy and in the peripartum period through a data-driven lens with full consideration of the maternal-fetal dyad throughout gestation and after delivery. The current scale of biological assay technologies to interrogate the immune system, genome, proteome, and metabolome has coincided with more robust data infrastructure for EHRs in order to enable integrative AI approaches in the obstetric and neonatal domains. Furthermore, wearable technologies have also entered into the domain as a way to understand the health status on a second-by-second basis. This data-driven approach is also readily extendable to the neonatal intensive care setting, as the real-time read-outs provided by data monitoring systems may provide additional insights needed for timely healthcare decisions regarding changes in ventilator settings, oxygen administration, the use of cardiotonic medications, and other time-sensitive clinical maneuvers.