Abstract Background Symptomatic thumb-base osteoarthritis (OA) affects 20% of people aged >55 years. Patients with thumb-base OA may experience pain, work disability, hand dysfunction and reduced quality of life (QoL). Although hand splints are a recommended intervention, this is supported by only low level evidence. Furthermore, the impact of placebo hand splints in OA remains unclear. This is the first trial powered to determine the efficacy of active versus placebo thumb splints, when added to a programme of self-management, for patients with symptomatic thumb-base OA. Methods A pragmatic, multi-centre, superiority randomised controlled clinical effectiveness and efficacy trial was conducted. Patients with symptomatic thumb-base OA and at least moderate thumb-base pain referred to UK out-patient occupational and physiotherapists were included. Patients with recent fractures, intra-articular joint injections and medical red flags were excluded. Eligible participants were randomised equally to either: (A) 8 weeks of self-management; (B) 8 weeks of self-management plus a verum splint; or (C) 8 weeks of self-management plus a placebo splint. The self-management programme was designed and delivered by therapists, based on recommended evidence-based practice. Participants in groups B and C were blinded to the type of splint received; clinical outcome assessors were also blinded. Participants were followed-up for 12 weeks from randomisation. The primary outcome was AUSCAN hand pain at 8 weeks. Secondary outcomes included work productivity, QoL, hand stiffness, hand function, disability and satisfaction. The trial was registered (ISRCTN. No 54744256). Results 17 UK NHS therapy departments assessed 751 patients for eligibility. 349 patients were randomised: 116, 116 and 117 to groups A, B and C, respectively. Treatment groups were well balanced for baseline characteristics. At baseline, no participants in group A received a splint; 99% and 97% of participants in groups B and C received the correct allocated splint. 80% of all participants attended their 4 week follow-up appointment. 8 week primary endpoint data were available for 84% of randomised participants; missing data were balanced across the groups. There was no evidence of clinically or statistically significant differences between treatment arms (B vs. C: -0.4 (95% CI -1.4, 0.6 p = 0.41), B vs. A: -0.5 (95% CI -1.4, 0.4 p = 0.26), C vs. A: -0.1 (95% CI -1.0, 0.8 p = 0.78). The per-protocol analysis, sensitivity analysis for missing data and analyses of secondary endpoints supported these findings. Ten participants reported adverse reactions: 3, 5 and 2 in groups A, B and C, respectively. Conclusion The addition of splinting to the recommended self-management package delivered by therapists did not confer additional benefit. Patients randomised to receive verum splints did not report better outcomes than those randomised to placebo splints. Disclosures J. Adams None. P. Barratt None. N.K. Arden None. S. Barbosa Bouças None. S. Bradley None. M. Doherty None. S. Dutton None. K. Dziedzic None. R. Gooberman-Hill None. K. Hislop None. C. Hutt-Greenyer None. V. Jansen None. R. Lunego None. I. Rombach None. M. Wiliams None.