BACKGROUND: Pneumococcus remains an important cause of morbidity in pregnant women with HIV and their infants. We compared the safety and immunogenicity of PCV-10 and PPV-23 with placebo administered in pregnancy. METHODS: This double-blind, multicentre, randomised controlled trial was done at eight outpatient clinics in Brazil. Eligible participants were adult women with HIV who were pregnant at a gestational age between 14 weeks and less than 34 weeks and who were taking antiretroviral therapy at study entry. Participants were randomly assigned (1:1:1) to receive either PCV-10, PPV-23, or placebo. Participants and study teams were unaware of treatment allocation. Antibodies against seven vaccine serotypes in PCV-10 and PPV-23 were measured by ELISA. The primary outcomes were maternal and infant safety assessed by the frequency of adverse events of grade 3 or higher; maternal seroresponse (defined as ≥2-fold increase in antibodies from baseline to 28 days after immunisation) against five or more serotypes; and infant seroprotection (defined as anti-pneumococcus antibody concentration of ≥0·35 µg/mL) against five or more serotypes at 8 weeks of life. The study was powered to detect differences of 20% or higher in the primary immunological outcomes between treatment groups. This trial is registered with ClinicalTrials.gov, NCT02717494. FINDINGS: Between April 1, 2016, and Nov 30, 2017, we enrolled 347 pregnant women with HIV, of whom 116 were randomly assigned to the PCV-10 group, 115 to the PPV-23 group, and 116 to the placebo group. One participant in the PCV-10 group did not receive the vaccine and was excluded from subsequent analyses. The frequency of adverse events of grade 3 or higher during the first 4 weeks was similar in the vaccine and placebo groups (3% [90% CI 1-7] for the PCV-10 group, 2% [0-5] for the PPV-23 group, and 3% [1-8] for the placebo group). However, injection site and systemic grade 2 adverse reactions were reported more frequently during the first 4 weeks in the vaccine groups than in the placebo group (14% [9-20] for the PCV-10 group, 7% [4-12] for the PPV-23 group, and 3% [1-7] for the placebo group). The frequency of grade 3 or higher adverse effects was similar across maternal treatment groups (20% [14-27] for the PCV-10 group, 21% [14-28] for the PPV-23 group, and 20% [14-27] for the placebo group). Seroresponses against five or more serotypes were present in 74 (65%) of 114 women in the PCV-10 group, 72 (65%) of 110 women in the PPV-23 group, and none of the 113 women in the placebo group at 4 weeks post vaccination (p<0·0001 for PPV-23 group vs placebo and PCV-10 group vs placebo). Seroresponse differences of 20% or higher in vaccine compared with placebo recipients persisted up to 24 weeks post partum. At birth, 76 (67%) of 113 infants in the PCV-10 group, 62 (57%) of 109 infants in the PPV-23 group, and 19 (17%) of 115 infants in the placebo group had seroprotection against five or more serotypes (p<0·0001 for PPV-23 vs placebo and PCV-10 vs placebo). At 8 weeks, the outcome was met by 20 (19%) of 108 infants in the PCV-10 group, 24 (23%) of 104 infants in the PPV-23 group, and one (1%) of 109 infants in the placebo group (p<0·0001). Although a difference of 20% or higher compared with placebo was observed only in the infants who received PPV-23 at 8 weeks of life, the difference between the two vaccine groups was not appreciable. INTERPRETATION: PCV-10 and PPV-23 were equally safe and immunogenic in pregnant women with HIV and conferred similar levels of seroprotection to their infants. In areas in which childhood PCV administration decreased the circulation of PCV serotypes, PPV-23 administration to pregnant women with HIV might be more advantageous than PCV by virtue of including a broader range of serotypes. FUNDING: Eunice Kennedy Shriver National Institute of Child Health and Human Development. TRANSLATION: For the Portuguese translation of the abstract see Supplementary Materials section.