Duchenne muscular dystrophy, the most common childhood muscular dystrophy, is caused by dystrophin deficiency. Preclinical and phase 2 study data have suggested that givinostat, a histone deacetylase inhibitor, might help to counteract the effects of this deficiency. We aimed to evaluate the safety and efficacy of givinostat in the treatment of Duchenne muscular dystrophy. This multicentre, randomised, double-blind, placebo-controlled, phase 3 trial was done at 41 tertiary care sites in 11 countries. Eligible participants were ambulant, male, and aged at least 6 years, had a genetically confirmed diagnosis of Duchenne muscular dystrophy, completed two four-stair climb assessments with a mean of 8 s or less (≤1 s variance), had a time-to-rise of at least 3 s but less than 10 s, and had received systemic corticosteroids for at least 6 months. Participating boys were randomly assigned (2:1, allocated according to a list generated by the interactive response technology provider) to receive either oral givinostat or matching placebo twice a day for 72 weeks, stratified by concomitant steroid use. Boys, investigators, and site and sponsor staff were masked to treatment assignment. The dose was flexible, based on weight, and was reduced if not tolerated. Boys were divided into two groups on the basis of their baseline vastus lateralis fat fraction (VLFF; measured by magnetic resonance spectroscopy): group A comprised boys with a VLFF of more than 5% but no more than 30%, whereas group B comprised boys with a VLFF of 5% or less, or more than 30%. The primary endpoint compared the effects of givinostat and placebo on the change in results of the four-stair climb assessment between baseline and 72 weeks, in the intention-to-treat, group A population. Safety was assessed in all randomly assigned boys who received at least one dose of study drug. When the first 50 boys in group A completed 12 months of treatment, an interim futility assessment was conducted, after which the sample size was adapted using masked data from the four-stair climb assessments. Furthermore, the starting dose of givinostat was reduced following a protocol amendment. This trial is registered with ClinicalTrials.gov, NCT02851797, and is complete. Between June 6, 2017, and Feb 22, 2022, 359 boys were assessed for eligibility. Of these, 179 were enrolled into the study (median age 9·8 years [IQR 8·1-11·0]), all of whom were randomly assigned (118 to receive givinostat and 61 to receive placebo); 170 (95%) boys completed the study. Of the 179 boys enrolled, 120 (67%) were in group A (81 givinostat and 39 placebo); of these, 114 (95%) completed the study. For participants in group A, comparing the results of the four-stair climb assessment at 72 weeks and baseline, the geometric least squares mean ratio was 1·27 (95% CI 1·17-1·37) for boys receiving givinostat and 1·48 (1·32-1·66) for those receiving placebo (ratio 0·86, 95% CI 0·745-0·989; p=0·035). The most common adverse events in the givinostat group were diarrhoea (43 [36%] of 118 boys vs 11 [18%] of 61 receiving placebo) and vomiting (34 [29%] vs 8 [13%]); no treatment-related deaths occurred. Among ambulant boys with Duchenne muscular dystrophy, results of the four-stair climb assessment worsened in both groups over the study period; however, the decline was significantly smaller with givinostat than with placebo. The dose of givinostat was reduced after an interim safety analysis, but no new safety signals were reported. An ongoing extension study is evaluating the long-term safety and efficacy of givinostat in patients with Duchenne muscular dystrophy. Italfarmaco.