The aim of our study is to investigate the natural history of scaphoid non-union.

Factors affecting the outcome of scaphoid fracture non-union surgery were assessed using data collected retrospectively from 785 cases treated at 18 centres throughout the UK. All cases had undergone this surgery before October 2014, ensuring a minimum period of 2 years between surgery and data collection from the medical records.

Smoking status significantly influenced the union rate of 94 proximal pole (non-smokers = 71%, smokers = 44%, p < 0.05) but not waist (non-smokers = 73% non-unions.

Type of bone graft did not affect the union rate of the 282 waist nonunions (iliac crest = 69%; non-vascular distal radius = 75% and vascularised distal radius = 70%: P=0.35), but did influence the union rate of the 98 proximal pole nonunions (iliac crest = 58%; non-vascular distal radius = 58% and vascularised distal radius = 82%: P=0.004).

Time interval between fracture and non-union surgery did not affect the union rate of 121 proximal pole nonunions (3–6 months = 56%; 6–12 months = 76%; 1–2 years = 69%; 2+years = 54%: P=0.5), but did influence the union rate of 303 nonunions of the waist (3–6 months = 80%; 6–12 months = 82%; 1–2 years = 65%; 2+years = 55%: P=0.02).

Non-unions of the proximal pole appear to be influenced by smoking status and bone graft type, whereas non-unions of the scaphoid are influenced by the time to non-union surgery.