BACKGROUND: We assessed maternal 23-valent pneumococcal polysaccharide (23vPPV) vaccine efficacy (VE) against middle ear disease and pneumococcal carriage amongst Australian Indigenous infants. METHODS: In an open label, allocation concealed, outcome-assessor blinded, community stratified, randomised controlled trial, healthy pregnant Indigenous women aged 17-39 years in the Northern Territory of Australia received the 23vPPV (1:1:1) at: 30-36 weeks gestation, birth, or were unvaccinated (ClinicalTrials.gov NCT00714064). Co-primary outcomes were the point prevalences of infant middle ear disease and 23vPPV-type carriage at age 7 months. RESULTS: The consent rate was 50% (313/632). Among 227 eligible participants randomised, retention rates were 86% (66/77) controls; 89% (67/75) pregnancy vaccinees; 88% (66/75) birth vaccinees. At infant age 7 months, ear disease prevalence was: 71% (47/66) controls, 63% (42/67) pregnancy vaccinees, 76% (50/66) birth vaccinees; and 23vPPV-type carriage was: 26% (17/66) controls, 18% (12/67) pregnancy vaccinees, 18% (12/66) birth vaccinees. For pregnancy vaccinees, VE was 12% (95% CI -12% to 31%) against infant ear disease and 30% (95% CI -34% to 64%) against 23vPPV-type carriage. In a post-hoc analysis, VE against infant ear disease concurrent with carriage of 23vPPV or related types was 51% (95% CI -2% to 76%). There were no serious adverse effects following receipt of the 23vPPV in pregnancy or at birth. CONCLUSIONS: In a high risk population, our study was unable to demonstrate efficacy of 23vPPV in pregnancy against the co-primary outcomes of either all-cause infant ear disease or 23vPPV-type nasopharyngeal carriage at age 7 months. Efficacy against ear disease concurrent with carriage of vaccine-related serotypes (a more specific outcome) suggests 23vPPV in pregnancy may complement childhood pneumococcal vaccination programs.