BACKGROUND: Many patients with psoriatic arthritis have an inadequate response to tumor necrosis factor (TNF) inhibitors. Guselkumab, a specific inhibitor of interleukin-23 (IL-23) via IL-23 p19 subunit binding, significantly improved psoriatic arthritis signs and symptoms with an acceptable safety profile in a phase 2 trial. METHODS: This multicentre, double-blind, randomised, placebo-controlled, phase 3 trial was done at 86 sites in 13 countries across Asia, Australasia, Europe, and North America and enrolled adults with active psoriatic arthritis (at least three swollen and three tender joints; and C-reactive protein ≥0·3 mg/dL) despite standard therapies. Eligibility criteria included inadequate response to or intolerance of standard treatment, including at least 4 months of apremilast, at least 3 months of non-biologic disease-modifying antirheumatic drugs (DMARDs), or at least 4 weeks of non-steroidal anti-inflammatory drugs for psoriatic arthritis. About 30% of study participants could have previously received one or two TNF inhibitors. Patients were randomly assigned (1:1:1, computer-generated permuted blocks; stratified by baseline DMARD and previous TNF inhibitor use) to subcutaneous guselkumab 100 mg every 4 weeks; guselkumab 100 mg at weeks 0, 4, then every 8 weeks; or matching placebo. The primary endpoint was American College of Rheumatology 20% improvement (ACR20) at week 24 in all patients per assigned treatment group using non-responder imputation. Safety was assessed in all patients per treatment received. This trial is registered at ClinicalTrials.gov, NCT03162796 (active, not recruiting). FINDINGS: From Aug 28, 2017, to Aug 17, 2018, we screened 624 patients, of whom 381 were randomly assigned and treated with guselkumab every 4 weeks (n=128), guselkumab every 8 weeks (n=127), or placebo (n=126). 362 patients continued study treatment up to week 24. The primary endpoint was met: ACR20 at week 24 was achieved by significantly greater proportions of patients in the guselkumab every 4 weeks group (76 [59%] of 128 [95% CI 50-68]) and every 8 weeks group (66 [52%] of 127 [43-61]) than in the placebo group (28 [22%] of 126 [15-30]), with percentage differences versus placebo of 37% (95% CI 26-48) for the every 4 weeks group and 30% (19-41) for the every 8 weeks group (both p<0·0001). Serious adverse events up to week 24 occurred in no patients receiving guselkumab every 4 weeks, four (3%) patients receiving guselkumab every 8 weeks, and five (4%) patients receiving placebo. Up to week 24, one patient in the placebo group died from cardiac failure and two had serious infections; no guselkumab-treated patient died or had serious infections. INTERPRETATION: Guselkumab demonstrated a favourable benefit-risk profile and might be an effective treatment option for patients with active psoriatic arthritis. FUNDING: Janssen Research and Development.