Background: Pregnant women and infants <6 months old have a high baseline risk for pneumococcal disease compared to the general population, particularly among Indigenous populations living in poverty and low‐resource settings. Efficacy trials of pneumococcal vaccination in pregnancy examining adverse birth outcomes are lacking. Aims: We report adverse birth events as secondary outcomes from the ‘PneuMum’ randomised controlled trial of 23‐valent pneumococcal polysaccharide vaccination (23vPPV) in pregnancy (August 2006–January 2011). Materials and methods: Australian Aboriginal women aged 17–39 years with singleton uncomplicated pregnancies were randomised (1:2 ratio) to receive 23vPPV or no 23vPPV in pregnancy at 30–36 weeks gestation. We compared risks of stillbirth, preterm birth, low birthweight (LBW), and small for gestational age (SGA) between vaccinated and unvaccinated pregnant women. Cox proportional hazard ratios (HRs) were calculated on an intention‐to‐treat basis. Results: Among 227 enrolled participants, 75 (33%) received 23vPPV in pregnancy. Risk differences in adverse birth outcomes between 23vPPV vaccinated and unvaccinated pregnant women were; preterm birth 9% vs 4% (HR 2.79; 95% CI 0.94–8.32) P = 0.07; LBW 9% vs 5% (HR 2.09; 95% CI 0.76–5.78) P = 0.15; and SGA 15% vs 17% (HR 1.02; 95% CI 0.50–2.06) P = 0.96. There were no stillbirths. Conclusions: We found a numerically higher rate of preterm births among women who received 23vPPV in pregnancy compared to unvaccinated pregnant women. Although further investigation with larger participant numbers is needed to better evaluate this safety signal, the contribution of safety results from smaller studies using appropriate data analysis methodologies is critical, particularly as more clinical trials in pneumococcal vaccination in pregnancy are progressing.