Background: Bedaquiline and delamanid are the first drugs of new classes registered for tuberculosis treatment in 40 years. Each can prolong the QTc interval, with maximum effects occurring weeks after drug initiation. The cardiac safety and microbiological activity of these drugs when co-administered are not well-established. Our aim was to characterise the effects of bedaquiline, delamanid, or both on the QTc interval, longitudinally over 6 months of multidrug treatment, among patients with multidrug-resistant or rifampicin-resistant tuberculosis taking multidrug background therapy. Methods: ACTG A5343 is a phase 2, open-label, randomised, controlled trial in which adults with multidrug-resistant or rifampicin-resistant tuberculosis receiving multidrug background treatment were randomly assigned 1:1:1 by centrally, computer-generated randomisation, by means of permuted blocks to receive bedaquiline, delamanid, or both for 24 weeks. Participants were enrolled at TASK in Cape Town and the South African Tuberculosis Vaccine Initiative in Worcester, both in South Africa, and Hospital Maria Auxiliadora in Peru. Individuals with QTc greater than 450 ms were excluded. HIV-positive participants received dolutegravir-based antiretroviral therapy. Clofazimine was disallowed, and levofloxacin replaced moxifloxacin. ECG in triplicate and sputum cultures were done fortnightly. The primary endpoint was mean QTcF change from baseline (averaged over weeks 8–24); cumulative culture conversation at week 8-24 was an exploratory endpoint. Analyses included all participants who initiated study tuberculosis treatment (modified intention-to-treat population). This trial is registered with ClinicalTrials.gov, NCT02583048 and is ongoing. Findings: Between Aug 26, 2016 and July 13, 2018, of 174 screened, 84 participants (28 in each treatment group, and 31 in total with HIV) were enrolled. Two participants did not initiate study treatment (one in the delamanid group withdrew consent and one in the bedaquiline plus delamanid group) did not meet the eligibility criterion). Mean change in QTc from baseline was 12·3 ms (95% CI 7·8–16·7; bedaquiline), 8·6 ms (4·0–13·1; delamanid), and 20·7 ms (16·1–25·3) (bedaquiline plus delamanid). There were no grade 3 or 4 adverse QTc prolongation events and no deaths during study treatment. Cumulative culture conversion by week 8 was 21 (88%) of 24 (95% CI 71–97; bedaquiline), 20 (83%) of 24 (65–95; delamanid), and 19 (95%) of 20 (79–100; bedaquiline plus delamanid) and was 92% (77–99) for bedaquiline, 91% (76–99), for delamanid, and 95% (79–100) for bedaquiline plus delamanid at 24 weeks. Interpretation: Combining bedaquiline and delamanid has a modest, no more than additive, effect on the QTc interval, and initial microbiology data are encouraging. This study provides supportive evidence for use of these agents together in patients with multidrug-resistant or rifampicin-resistant tuberculosis with normal baseline QTc values. Funding: Division of AIDS, National Institutes of Health.