Are we enrolling representative cohorts of premature infants in our clinical trials?

Objective: To compare the difference in outcomes in a subset population of infants “eligible but not enrolled; ENE” vs those who were “eligible and enrolled, EE” in The Australian Placental Transfusion Study (APTS). Study design: Population-based multicentre retrospective cohort study. Results: A total of 535 (17.7%) infants were categorized as EE and 2489 (82.3%) ENE. ENE infants were significantly more premature (mean gestation 27.0 vs 28.0 weeks) but otherwise of similar anthropometric measures compared to EE infants. ENE infants had significantly higher incidences of low Apgar scores <7 at 5 min, CLD, IVH and PDA requiring treatment. Using a multivariate adjusted-analysis, ENE were at a greater risk for mortality (OR 1.86; 95% CI, 1.30–2.67, p < 0.001). Conclusion: Antenatal consenting may lead to biased population representation, which may affect trial results’ generalizability. Retrospective consent or waiver of consent may improve the generalizability of neonatal and emergency clinical trials.